The Middle Stone Age of Nigeria in its West African Context [1 ed.] 1789691389, 9781789691382

This book provides a full up to date account of the evidence relating to the Middle Stone Age in Nigeria and the other c

312 65 28MB

English Pages 260 Year 2020

Report DMCA / Copyright

DOWNLOAD PDF FILE

Table of contents :
Cover
Contents
Copyright Page
Bookmark 3
Title Page
Contents
List of Figures
Chapter 1
Geography
Geology and Geomorphology
The Middle Stone Age in West Africa: Introduction
Vegetation and Climate
Climatic and Environmental History
Archaeological perspectives
Institutional Framework
References
Climatic and Environmental History
Geography
Geology and Geomorphology
Vegetation and Climate
Archaeological perspectives
Institutional Framework
Chapter 2
The Middle Stone Age of Nigeria
Northern Nigeria
Zenabi
Introduction
Site situation and stratigraphy
(1) Earlier investigations
(2) Results of work done in 1976-1978
(3) 1976-78 and Jos Museum collections
(4) Nature of the industry
Conclusion
Mai Lumba
Yada Gungume
Tibchi
Yelwa
Ningi Hills
Banke
Saminaka
Nok
Pingell
Rop
Summary
Sangoan
Southern Nigeria
The Bodija Formation
Ajibode
Olude-Araromi
Asejire
References
Northern Nigeria
Southern Nigeria
Summary
Chapter 3
Cameroun
Northern Cameroun
Southern Cameroun
The Middle Stone Age in West Africa
Niger
Adrar Bous
Bilma
Seggédim
Mékrou Valley
Ghana
Asokrochona, the Ghana Nautical College, and Tema
Other sites in Ghana
Ivory Coast
Attinguié and Anyama
South-west Ivory Coast
Sierra Leone
Dated deposits in the mining areas
The alluvial sequence in the mining areas
Yengema cave
Artefacts recovered from the alluvial deposits
Mali
Summary
Senegal
West and South
North and East
Guinea
Summary
Conclusion
References
Cameroun
Ghana
Niger
Ivory Coast
Mali
Senegal
Sierra Leone
Guinea
Chapter 4
West Africa: regional summary
References
Chapter 5
A wider perspective
Climate
Archaeology
Human Evolution
Conclusion
References
Archaeology
Climate
Human Evolution
Envoi
Figure 1. West Africa: Political boundaries.
Figure 2. West Africa: Geographical features.
Figure 3. West Africa: Geology.
Figure 4. Vegetation zones of Nigeria (after Keay, 1965).
Figure 5. Vegetation zones of West Africa.
Table 1. The pluvial and inter-pluvial system as assumed to exist in Africa (after Clark, 1959, Table 2).
Figure 6. Oxygen isotope stages and the palaeomagnetic time scale (after Smart and Frances, 1991, Fig. 9.1).
Figure 7. Palynological and stratigraphic record at lake Bosumtwi (after Miller and Gosling, 2014, Fig. 2).
Figure 8. Preferential and recurrent Levallois techniques (after Boëda, 1994, Fig. 176).
Figure 9. Nubian cores (1 and 2), ‘classic’ and Halfan cores (3 and 4) (after Van Peer, 1991, Fig. 3).
Figure 10. The study area north of the Jos Plateau.
Figure 11. The Younger Granite Ring Complexes of Nigeria (after Turner 1976, Fig. 1).
Figure 12. The Younger Granite Ring Complexes of Niger (after Ngako et al., 2006, Fig. 3).
Figure 13. The area north of the Jos Plateau: Physical Geography.
Figure 14. The area north of the Jos Plateau: Geology.
Figure 15. Alluvial tin-mining at Nok (after Fagg 1977, Frontispiece).
Figure 16. Geological sketch map of Zenabi (after Bond, 1956, 201).
Figure 17. Carbonised tree trunk in Zenabi No.1 Paddock (after Fagg, 1956, Fig. 6).
Figure 18. Zenabi sections A, B, and C (after Anozie, 1975, Fig. 16).
Figure 19. Zenabi Fulata, 23 March 1976.
Figure 20. Zenabi Fulata, 23 March 1976, in situ find.
Figure 21. Zenabi Fulata, 23 March 1976, in situ find.
Figure 22. Zenabi Falls, March 1976.
Figure 23. Zenabi Falls, July 1976.
Figure 24. Zenabi contour map showing mining leases 3741, 9282, and 5174.
Figure 25. Zenabi plan as surveyed in 1976-1978 with position of sections 1, 2, and 3.
Figure 26. Zenabi sections 1-3, 1976 and 1978.
Table 2. Zenabi Section 1 stratigraphy 1976-1978.
Figure 27. Zenabi Fulata, work in progress, July 1976.
Figure 28. Zenabi Fulata, section1.
Figure 29. Zenabi Fulata, section 2.
Figure 30. Zenabi Tudo, March 1976.
Table 3. Zenabi artefact inventory.
Figure 31. Zenabi Fulata. 1, limace. 2, angle burin. 3, double sidescraper on Levallois flake/blade. [PAJ 1980 Fig. 1. ABC].
Figure 32. Zenabi Fulata. 3, 1-platform flake/blade core. 5, endscraper. Zenabi Karara. 1, transverse convex sidescraper. 2, Levallois point. 4, disc core. [PAJ 1980 Fig. 2. ABCDE].
Figure 33. Zenabi Fulata 1976. 1 and 4, transverse and straight sidescrapers. Zenabi Fulata 1977, finds derived from 1976 section, upper tinwash. 2, limace. 3, flake.
Figure 34. Zenabi Fulata 1978. Lower tinwash, upper series. 2, plain platform flake. 3, waste flake. Base of deep sounding. 1, core-like chunk. 4, flake/blade. 5, waste flake.
Figure 35. Zenabi Fulata 1977, finds derived from 1976 section, upper tinwash. Initial/preparatory core.
Figure 36. Zenabi, Jos Museum. MP56/33. Initial Levallois core.
Figure 37. Zenabi, Jos Museum. MP56/33. Levallois flake/blade core.
Figure 38. Zenabi, Jos Museum. C. Levallois flake core.
Figure 39. Zenabi, Jos Museum. MP56/33. 1, Levallois core. Karara. 2, Levallois flake core.
Figure 40. Zenabi, Jos Museum. A47/10. 1, Levallois flake core. ZNP 1962. 2, Levallois point core. Z ‘west of spur’ 1948. 3, 1-platform flake/blade core. C47/10. 4, Levallois flake core.
Figure 41. Zenabi, Jos Museum. A. 1-platform flake/blade core. C. 2, 1-platform blade core.
Figure 42. Zenabi, Jos Museum. B47/10. 1 and 2, 1-platform flake/blade cores. Karara 1948. 3, 1-platform ‘boat-shape’ core. A ‘in situ’. 4, 2-platform flake/blade core.
Figure 43. Zenabi, Jos Museum. Karara ‘1/2 mile north of bridge’. Disc core.
Figure 44. Zenabi, Jos Museum. MP 1962 ‘L.G.’. Disc core.
Figure 45. Zenabi, Jos Museum. MP 1964. 1, disc core. MP56/34 ‘B.G.’. 2, disc core. Karara 1948. 3, disc core.
Figure 46. Zenabi, Jos Museum. Fulata. 1, Levallois flake. C47/10. 2, Levallois flake. Karara. 3, Levallois flake/point. A47/10. 4, Levallois blade.
Figure 47. Zenabi, Jos Museum. ZR ‘gravel’ 1962. 1, Levallois flake. Karara 1948. 2, Levallois flake. MP56/33. 3, Levallois blade. D. 4, retouched Levallois flake. ZNP 1962. 5, Levallois flake/blade.
Figure 48. Zenabi, Jos Museum. A47/10. 1, overshot blade. C. 2, overshot flake. ZMP 56/33. 3, pseudo-levallois point. Karara. 4, pseudo-levallois point. 5, blade.
Figure 49. Zenabi, Jos Museum. A. 1, limace. Karara 1948. 2, limace. 3, flake.
Figure 50. Zenabi, Jos Museum. ZNP 1962. 1, limace. ZR 1962. 2, endscraper. Karara. 3, endscraper. A. 4, endscraper.
Figure 51. Zenabi, Jos Museum. MP56/33. 1, transverse convex sidescraper. Karara. 2, multiple canted sidescraper. A. 3, multiple dihedral burin. 4, burin on retouched truncation.
Figure 52. Zenabi, Jos Museum. ‘No.1 middle gravels I/S’. 1, convergent sidescraper with alternate retouch. ZR 1962 ‘I/S gravels’. 2, notch. ZMP. 3, canted sidescraper. Karara. 4, double sidescraper.
Figure 53. Zenabi, Jos Museum. MP56/33. 1, trihedral. Karara. 2, double sidescraper.
Figure 54. Zenabi, Jos Museum. ZMP. 1, trihedral. B. 2, canted sidescraper on Levallois flake.
Figure 55. Zenabi Fulata 1976. 1, canted sidescraper. 2, transverse convex sidescraper. 3, sidescraper on ventral surface. Zenabi Tudo 1968. 4, Levallois point.
Figure 56. Zenabi Tudo. 1, disc core. 2, pseudo-levallois point. 3, Levallois flake. 4, Levallois blade.
Table 4. Zenabi artefact measurements.
Figure 57. Tibchi and Mai Lumba and their position in relation to the Tibchi and Yeli hills.
Figure 58. Mai Lumba lease 988, 24 March 1976.
Figure 59. Mai Lumba lease 988, 24 March 1976.
Figure 60. Mai Lumba lease 963, ‘kafanchan’ working, June 1976.
Table 5. Mai Lumba lease 988. Stratigraphy, sections 1-4.
Figure 61. Mai Lumba lease 988, 21 April 1978.
Figure 62. Mai Lumba lease 988, 21 April 1978.
Figure 63. Mai Lumba 1976-1978 lease 988, plan and sections 1-4.
Figure 64. Mai Lumba lease 988, pump in operation, 15 April 1978.
Figure 65. Mai Lumba lease 988, section 1, 23 April 1978.
Figure 66. Mai Lumba lease 988, section 1, 23 April 1978.
Figure 67. Mai Lumba lease 988, section 2, 24 April 1978.
Figure 68. Mai Lumba lease 988, section 2, 25 April 1978.
Figure 69. Mai Lumba lease 988, section 3, 29 April 1978.
Figure 70. Mai Lumba lease 988, section 3, 29 April 1978.
Figure 71. Mai Lumba. 1 and 2, Levallois blade and flake/blade. 3, Levallois blade core. 4, convergent sidescraper. [PAJ 1986 Fig. 9.2].
Figure 72. Mai Lumba. 1, Levallois flake core [PAJ 1986 Fig. 9.4.1]. 2, flake/blade. 3 and 4, flakes. 5, disc core. 2-5, excavated 1978, 2-4 upper tinwash, 5 lower tinwash.
Table 6. Mai Lumba and Yada Gungume. Artefact inventory.
Figure 73. Mai Lumba. Initial core.
Figure 74. Mai Lumba. Levallois cores. 1, flake/point core. 2-3, flake cores. 4, 2-platform flake/blade core.
Figure 75. Mai Lumba. 1 and 2, 2-platform flake/blade cores, 2 struck from opposing directions.
Figure 76. Mai Lumba. 1 and 4, 2-platform flake/blade cores. 2 and 3, 1-platform flake/blade cores.
Figure 77. Mai Lumba. 1 and 2, disc cores. 3 and 4, 1-platform flake/blade cores.
Figure 78. Mai Lumba. 1-4, disc cores.
Figure 79. Mai Lumba. 1 and 4, Levallois points. 2-3 and 5, Levallois flake/blades.
Figure 80. Mai Lumba. 1-6, Levallois flakes and flake/blades.
Figure 81. Mai Lumba. 1-5, Levallois flakes.
Figure 82. Mai Lumba. 1, flake/blade. 2-4, flakes.
Figure 83. Mai Lumba. 2, blade. 1, 3-4, flakes.
Figure 84. Mai Lumba. 1 and 2, convex and double sidescrapers. 3, pseudo-levallois point. 4-5, overshot flake/blades.
Figure 85. Mai Lumba. 1 and 5, sidescrapers on ventral surface. 2 and 3, straight and convex sidescrapers. 4, Levallois flake/blade.
Figure 86. Mai Lumba. 1 and 3, broken ground stone axes. 4, Mai Lumba Tudo, double/convergent sidescraper.
Figure 87. Mai Lumba. 1, transverse convex sidescraper. 2, convergent sidescraper. 3, broken ground stone axe.
Figure 88. Mai Lumba. Biface.
Figure 89. Mai Lumba lease 988, ground stone axe.
Figure 90. Mai Lumba lease 988, ground stone axes.
Figure 91. Mai Lumba, disc and 2-platform Levallois flake/blade core.
Figure 92. Mai Lumba, disc and Levallois flake core.
Figure 93. Mai Lumba, 2-platform flake/blade cores.
Figure 94. Mai Lumba, microsyenite initial core.
Figure 95. Mai Lumba, rare raw materials at the site.
Figure 96. Tibchi 1976 paddock, plan and sections.
Figure 97. Tibchi, general view of paddock, 2 July 1976.
Figure 98. Tibchi, section 1, 2 July 1976.
Figure 99. Tibchi, tin-miners removing lower tin-wash, 2 July 1976.
Figure 100. Tibchi, section 2, 2 July 1976.
Figure 101. Yada Gungume. 1, biface. Tibchi. 2, 2-platform core. 3, sidescraper on ventral surface.
Table 7. Tibchi 1976 paddock. Stratigraphy sections 1 and 2.
Figure 102. Tibchi. 1, disc core. 2, Levallois/disc core. 3 and 4, disc cores.
Table 8. Yelwa and Tibchi. Artefact inventory.
Figure 103. Tibchi. 1, Levallois flake core. 2, Levallois/disc core. 3, 1-platform core/chopper. 4, 1-platform core.
Figure 104. Tibchi. 1, 1-platform core. 2, Levallois blade. 3, Levallois flake/blade. 4, overshot Levallois flake/blade.
Figure 105. Tibchi. 1, blade. 2 and 5, flake/blades. 3 and 7, Levallois flakes. 4, convex sidescraper. 6, Levallois blade. 8, pseudo-levallois point.
Figure 106. Tibchi. 1, pseudo-levallois point. 2, 5 and 6, flakes. 3, canted sidescraper. 4, transverse convex sidescraper. 7, blade.
Figure 107. Tibchi. 1, convergent sidescraper/limace. 2, transverse convex sidescraper. 3, sidescraper on ventral surface. 4, straight sidescraper with thinned back.
Figure 108. Tibchi. 1, double sidescraper/limace. 2, canted sidescraper/core. 3, limace.
Figure 109. Tibchi. 1, limace. 2, concave sidescraper. 3, biface.
Figure 110. Tibchi. Canted sidescraper/limace.
Figure 111. Tibchi. 1, limace. 2 and 3, transverse convex sidescrapers.
Figure 112. Tibchi, limaces (L and R), dorsal view.
Figure 113. Tibchi, limace (L), side view.
Figure 115. Yelwa 1978 lease 4311, plan and section.
Figure 116. Yelwa village, March 1977.
Table 9. Yelwa lease 4311. Stratigraphy.
Figure 117. Yelwa lease 4311, general view, 10 April 1978.
Figure 118. Yelwa lease 4311, mining work in progress, 10 April 1978.
Figure 119. Yelwa, Jos Museum. 1, limace. 3 and 4, transverse convex sidescrapers. 2, YMP 1962, blade.
Figure 120. Yelwa, Jos Museum. YMP LG 1962. 1, initial core. 2, 1-platform flake/blade core. Yelwa 56/31. 3, disc core. 4, convergent sidescraper.
Figure 121. Yelwa 1978. 1, 2-platform core. 2, disc core. 3, 1-platform flake/blade core.
Figure 122. Yelwa 1978. 1, 1-platform flake/blade core. 2, flake.
Figure 123. Yelwa 1978. Initial/preparatory core, in situ in tinwash.
Figure 124. Ningi Hills: Physical Geography.
Figure 125. Ningi Hills: Geology.
Figure 126. Koluki, 2 May 1978.
Table 10. Ningi Hills. Artefact inventory.
Figure 127. Koluki, 2 May 1978.
Figure 128. Old Ningi, 9 May 1978.
Figure 129. Jigawa, 8 May 1978.
Figure 130. Jigawa. 1, Levallois flake/blade core. 2 and 5, disc cores. 3 and 6, 1-platform flake/blade cores. 4, flake/blade.
Figure 131. Jigawa. 1 and 2, disc cores.
Figure 132. Jigawa. 1, disc core. 2 and 4, Levallois flakes. 3, Levallois blade. 5, flake. 6 and 7, blades.
Figure 133. Koluki. 1, 1-platform flake/blade core. Tsofon Ningi. 2, Levallois flake/blade core. Jigawa. 3 and 6, flakes. 4, straight sidescraper on Levallois flake/blade. 5, bifacial sidescraper.
Figure 134. Tabela. 1 and 4, 1-platform flake/blade cores. 2, disc core. 3, Levallois flake/blade core. 5, Levallois blade.
Figure 135. Tabela. 1, convex sidescraper. 2, disc core. 3, flake, in situ upper tinwash. 4, endscraper. 5, Levallois flake.
Figure 136. Jigawa, disc cores.
Figure 137. Banke, general view looking north, 27 March 1977.
Figure 138. Banke, map showing position of leases on river Marwan Karini.
Figure 139. Banke, plan of paddocks in lease 13468 over the period 1976-1978, with sections 1-3 and location of samples for radiocarbon dating.
Figure 140. Banke, NW and SE paddocks, 29 December 1976.
Figure 141. Banke, SW extension paddock, 8 April 1978.
Figure 142. Banke, quartz handaxe (July 1976) and perforated stone artefact (June 1977).
Figure 143. Banke, SE paddock 29 December 1976, clay horizon above the upper tin-wash.
Figure 144. Banke, SE paddock 27 March 1977, water pump in action.
Figure 145. Banke, SE paddock 27 March 1977, clay horizon and lower tin-wash.
Figure 146. Banke, SE paddock 27 March 1977, tree trunk fallen from clay horizon above the upper tin-wash.
Figure 147. Banke, sections 1 and 2 (SE paddock) and 3 (SW extension paddock).
Figure 148. Banke, section 3 (SW extension paddock) with tree trunk above the upper tin-wash.
Table 11. Banke lease 13468. Stratigraphy, sections 1-3.
Figure 149. Banke, two cleavers found in April 1978.
Figure 150. Banke. 1 and 4, 1-platform flake/blade cores. 2 and 3, disc cores. 5, blade with marginal retouch/sidescraper.
Figure 151. Banke. 1, disc core. 2, endscraper. 3, ground stone axe in course of manufacture. 4, broken half of ground stone axe in course of manufacture. 5, Levallois blade. 6, notch/denticulate. 7, transverse sidescraper.
Figure 152. Banke. 1, apparent roughout for ground stone axe. 2, 1-platform initial core. 3, perforated stone artefact.
Figure 153. Banke. Cleaver.
Table 12. Banke lease 13468. Artefact inventory.
Figure 154. Saminaka, April 1982.
Figure 155. Saminaka map (after Opadeji 2001, Fig. 2).
Figure 156. Saminaka sketch of cross-section (after Opadeji 2001, Fig. 13).
Table 13. Saminaka Stone Age (SSA) Artefact Inventory.
Figure 157. Sketch map of the Nok Valley (after Fagg 1956, Fig. 9).
Figure 158. Hypothetical section of the Nok Valley (after Fagg 1956, Fig. 8).
Figure 159. Nok valley, 27 March 1976.
Figure 160. Nok abandoned paddock, 27 March 1976.
Figure 161. Pingell, April 1976.
Figure 162. Pingell, April 1976.
Figure 163. Distribution of Sangoan sites in Northern Nigeria (adapted from Soper, 1965, 184-185).
Table 14. Nigerian Sangoan assemblages(after Soper, 1965, Table III).
Figure 164. Sangoan tools (descriptions after Soper, 1965, Figs. E, G, and J). 1, Keffi, quartz push-plane. 2, Yumu, quartz handaxe. 3, Nassarawa, quartz handaxe. 4, Jebba, large quartzite pick.
Figure 165. Jebba. Locality 4, December 1983.
Figure 166. Jebba. Locality 3, December 1983.
Figure 167. Jebba. Locality 3, artefact horizon indicated by Arthur Harding.
Figure 168. Jebba. Locality 3, artefact horizon.
Figure 169. Jebba. 1-platform quartzite core.
Figure 170. Jebba. sidescraper, pseudo-levallois point, two flakes.
Figure 171. Jebba. two flake/blade cores, one round scraper.
Figure 172. Jebba. quartzite biface pick, quartz uniface pick.
Figure 173. Jebba. quartzite pebbles/choppers.
Figure 174. Geological Map of Ibadan (after Burke and Durotoye, and Kiladejo, 1980, Fig. 2).
Figure 175. Sites in the Ibadan area (after Bagodo, 2012, Fig. 15).
Figure 176. Bodija Railway Cutting, 13 December 1979.
Figure 177. Bodija Formation type section in the Bodija Railway Cutting, Ibadan (after Burke and Durotoye, 1971, Fig. 6).
Table 15. Bodija Formation. Stratigraphic column (after Durotoye, 1976, Table 1).
Figure 178. Bodija Railway Cutting, particle size distribution for Orita and Agodi members and weathered in situ gneiss (after Durotoye 1972, and Kiladejo, 1980, Fig. 11).
Figure 179. Sites in the vicinity of Ajibode (after Momin, 1995, Fig. 1, and Bagodo, 2012, Fig. 13).
Table 16. Ajibode UMF site. Stratigraphy of Trench D.
Table 17. Ajibode UMF site. Dimensions of flakes, cores, points, and picks.
Figure 180. Olude-Araromi, general view, sections left of the road, 13 December 1979.
Table 18. Olude-Araromi.Stratigraphy with major units identified.
Figure 181. Bodija Formation sections at Olude-Araromi (after Kiladejo, 1980, Fig. 8).
Figure 182. Palaeolithic sites in Northern Cameroun (modified after Marliac 2006, Carte 9 bis).
Figure 183. Northern Cameroun. Traverse number 4 (after Marliac 2006, Figure 2).
Figure 184. Mindif 80. 1, 2-platform flake/blade core. 2 and 4, Levallois points. 3, Levallois point core. 5, flake. 6 and 7, Levallois blades. [PAJ 1986 Fig. 9.6: 4-7].
Figure 185. Mindif 80. 1, Levallois flake. 2, Levallois blade. 3, flake from disc core. 4, core edge removal flake. 5, flake. 6, flake from disc core. 7, convex sidescraper on Levallois point. 8, chopper. [PAJ 1986 Fig. 9.6: 1-3].
Figure 186. Dent de Mindif, 9 February 1980.
Figure 187. Dent de Mindif, 9 February 1980.
Figure 188. Dent de Mindif, ferruginous nodule horizon.
Figure 189. Dent de Mindif, ferruginous nodule horizon and artefacts.
Figure 190. Katchel, 9 February 1980.
Figure 191. Katchel, 9 February 1980.
Figure 192. Northern Cameroun. Traverse number 1 (after Marliac 2006, Figure 3).
Figure 193. Stratigraphy of Mayo Louti site (after Digara 1988: 80).
Figure 194. Mayo Louti, general view, 11 February 1980.
Table 19. Mayo Louti and Nassarao. Artefact inventory.
Figure 195. Mayo Louti, ‘basse terrasse’, 11 February 1980.
Figure 196. Mayo Louti, ‘moyennes terrasses’, 11 February 1980.
Figure 197. Mayo Louti, Douroumian ‘glacis’, 11 February 1980.
Figure 198. Mayo Louti. Levallois flake/blade, pseudo-levallois point, disc cores.
Figure 199. Mayo Louti. 1, initial core. 2 and 4, 1-platform flake/blade cores. 3, 1-platform core/chopper.
Figure 200. Mayo Louti. 1, 1-platform flake/blade core. 2, Levallois flake core. 3, Levallois point core. 4, disc core. 5, Levallois blade.
Figure 201. Mayo Louti. 1, Levallois blade core. 2, Levallois blade. 3, Levallois point. 4, Levallois flake/point. 5 and 6, Levallois point cores. 7, Levallois flake/point.
Figure 202. Mayo Louti. 1, Levallois flake/point core. 2, Levallois blade. 3, Levallois flake core. 4, Levallois blade. 5 and 6, Levallois flake/points.
Figure 203. Mayo Louti. 1, flake. 2, concave sidescraper on Levallois flake. 3, straight sidescraper. 4, pseudo-levallois point. 5, convex sidescraper.
Figure 204. Mayo Louti. 1 and 6, concave sidescrapers. 2, concave-convex sidescraper. 3 and 4, pseudo-levallois points. 5, convex sidescraper. 7, endscraper.
Figure 205. Mayo Louti. 1, retouched Levallois point/Mousterian point. 3, convex sidescraper. Douroum. 2, flake. 4, Levallois blade. 5, Levallois flake.
Figure 206. Stratigraphy of Nassarao site (after Digara 1988: 186).
Figure 207. Douroum, 8 February 1980.
Figure 208. Douroum, 8 February 1980.
Figure 209. Douroum. 1 and 3, disc cores. Sanguéré. 2, straight sidescraper.
Figure 210. Kossi, 6 February 1980.
Figure 211. Kossi, 6 February 1980.
Figure 212. Kossi, above the Mayo Kébi, 6 February 1980.
Figure 213. Tongo. 1, 1-platform flake core. 2, Levallois flake/point core. Kossi. 3, flake. 4, convex sidescraper. 5, concave sidescraper.
Figure 214. Sanguéré, 5 February 1980.
Figure 215. Sanguéré, 5 February 1980.
Figure 216. Sanguéré. 1, 2-platform Levallois blade core. 2, disc/Levallois core.
Figure 217. Sanguéré. 1, Levallois flake core. 2, disc core.
Figure 218. Sanguéré. 1, Levallois flake core. 2, Levallois blade. 3, Levallois blade core.
Figure 219. Sanguéré. 1, Levallois point core. 2, flake. 3, concave sidescraper.
Figure 220. Tongo. 1, disc core. 2, Levallois point. 3, flake. Sanguéré. 4, transverse convex sidescraper. 5, Levallois flake. 6, Levallois blade.
Figure 221. Sanguéré, disc and Levallois cores.
Figure 222. Sanguéré, Levallois cores.
Figure 223. Tongo, 5 February 1980.
Figure 224. Tongo, 5 February 1980.
Figure 225. Southern Cameroun. 1, Geological section at Locality 156. 2, Generalized geological section at Locality 114. 3, Locality 156, unifacial pick. (after Omi 1977, Figs. 24, 25.2, and 32).
Figure 226. Southern Cameroun. Locality 114, pick.(after Omi 1977, Fig. 38).
Figure 227. Southern Cameroun. Locality 114. 1 and 2, picks. 3, unifacial scraper. (after Omi 1977, Fig. 39).
Figure 228. Adrar Bous stratigraphic sections (after Clark et al., 2008, Fig. 4.1).
Figure 229. Bilma stratigraphic section (after Maley et al., 1971, Fig. 5). 3, Holocene layers. 2, Calcareous diatomites. 3, Silémi terrace, including the following sequence from the top: (e) sand with calcareous nodules (d) lacustrine limestone with
Table 20. Seggédim. Artefact inventory.
Figure 230. Map of Ghana, with principal sites indicated.
Figure 231. Asokrochona railway cutting, 28 December 1980.
Figure 232. Asokrochona railway cutting (after Davies 1967, Fig.1).
Figure 233. Asokrochona generalised section (after Nygaard and Talbot 1976, Fig.2).
Figure 234. Ghana Nautical College, 28 December 1980.
Table 21. Asokrochona artefact totals
Table 22. Asokrochona tool classes
Figure 235. Asokrochona [Oliver Davies]. 1, 2 and 4, disc cores. 3, Levallois flake/blade.
Figure 236. Asokrochona [Oliver Davies]. 1, core-axe. 2, blade. 3, convex sidescraper. [QCG Fig. 67.18, 21, 3].
Figure 237. Asokrochona [Oliver Davies]. 1, handaxe. [CQG Fig. 66.1]. 2, Levallois flake/point.
Figure 238. Asokrochona. 1, initial core. 2, disc core. 3, 2-platform flake/blade core. 4, 1-platform flake/blade core. [1, test pits I-IV, 2-4 Area II].
Figure 239. Asokrochona. 1, 3, 5 and 6, disc cores. 2, notch. 4, sidescraper on Levallois flake. [1 and 3-6, Area II, 4 Area I].
Figure 240. Asokrochona. 1 and 5, bifaces in course of manufacture. 2, point/convergent sidescraper. 3, disc core. 4, convex sidescraper. 6, convergent sidescraper. 7, awl. [Area II].
Figure 241. Asokrochona. 1, convergent sidescraper. 2, denticulate. 3 and 5, chopping tools. 4, chopper. [Area II].
Figure 242. Asokrochona. 1, pick. 2, transverse convex sidescraper. [Area II].
Figure 243. Asokrochona. 1, handaxe. 2, point. [Area II].
Figure 244. Asokrochona. 1, point. 2, concave sidescraper/notch. 3, awl. [Area II]. Tema [Oliver Davies]. 4, Levallois flake. 5, blade. 6, disc core. [QCG Fig. 57.7, 56.9, 56.7]. 7, disc/bipolar core [Tema West I 1976 Nygaard].
Figure 245. Tema [Oliver Davies]. 1, handaxe/pick. 2, Levallois flake. 5, pseudo-levallois point. [QCG Fig. 56.16 and 17]. 3, convex sidescraper. 4, chopper. 6, chopping tool. [Tema West II 1976 Nygaard].
Figure 246. Asokrochona quartz bifaces.
Figure 247. Hohoe handaxe/pick.
Figure 248. Chawenu [Oliver Davies]. 1, Levallois flake/blade core. 2, Levallois flake. 3, biface in course of manufacture. 4, pseudo-levallois point. 5, endscraper. 6, chopping tool.
Figure 249. Chawenu [Oliver Davies]. 1 and 2, disc cores. Hohoe [Oliver Davies]. 3, handaxe/pick.
Figure 250. Ghana Nautical College. 1, disc core. 2, double sidescraper. [Nygaard 1976]. Narago [Oliver Davies]. 3, pseudo-levallois point. 4, retouched Levallois point. 5, Levallois flake. 6, retouched Levallois flake. [QCG Fig. 62.6, 62.8].
Figure 251. Birimi location and topographic map (after Quickert et al. 2003, Fig. 1).
Figure 252. Birimi MSA profile east wall of main gully (after Quickert et al. 2003, Fig. 2).
Figure 253. Abidjan area, with Attinguié and Anyama.
Figure 254. South-west Ivory Coast, with Gouabuo and Sablière.
Figure 255. Plan of Anyama (after Guédé and Tastet, 1986, Fig. 1c).
Figure 256. Anyama, view to quarry looking west from roadside, 7 November 1983.
Figure 257. Anyama, west face of quarry with excavated area at top, 7 November 1983.
Figure 258. Anyama, excavated area surface view to north, 7 November 1983.
Figure 259. Stratigraphy of Anyama (after Liubin and Guédé, 2000, Fig. 6.I).
Table 23. Anyama I-IV layer D and Gouabuo I and II
Table 24. Anyama I-III layer C and Sablière
Figure 260. Sierra Leone map.
Figure 261. Yengema mining area.
Figure 262. Yengema NDMC golf course, 16 October 1986.
Figure 263. Yengema rest house and cave, 17 October 1986.
Figure 264. Cross section of a typical valley (after H.J.E. Haggard).
Figure 265. Gbogbora river east of Fotingaya, 17 October 1986.
Figure 266. Gbogbora river ‘cooperative mining site’, 17 October 1986.
Table 25. Yengema. Dimensions (means + SDs) of bifacial artefacts.
Figure 267. Yengema broken bifaces.
Figure 268. Yengema broken bifaces.
Figure 269. Yengema complete bifaces.
Figure 270. Ground stone axe found by Mr A. Nabby in Yengema area.
Figure 271. Quartz handaxe found by Mr H.H. Jackson in Koidu area.
Figure 272. Map of Mali with Dogon plateau and sites.
Figure 273. Ounjougou location and geomorphology (after Soriano et al. 2010a, Fig. 1).
Figure 274. Yamé confluence of four rivers, 12 February 2001.
Figure 275. Yamé view upstream.
Figure 276. Oumounaama 4.
Figure 277. Oumounaama 5.
Figure 278. Kokolo 3.
Figure 279. Kokolo 3.
Figure 280. Kokolo 2.
Figure 281. Yamé campement, 14 February 2001.
Figure 282. Ounjougou archaeological levels (after Chevrier et al. 2018, Fig. 10).
Figure 283. Bandiagara escarpment and Songona (after Rasse et al. 2012, Photo 1).
Figure 284. Map of Senegal, showing position of more detailed charts of Senegal and Falémé river valleys and Cape Verde.
Figure 285. Cape Verde peninsula.
Figure 286. Falémé river valley, recently excavated sites (after Lebrun et al. 2017, Fig. 1).
Figure 287. Senegal river valley, recently excavated sites (after Scerri et al. 2015, Fig. 5).
Table 26. Four artefact inventories from Senegal
Figure 288. Tiémassas excavation section (after Descamps 1979, Fig. 29).
Figure 289. Tiémassas ravine first stop, 20 October 1988.
Figure 290. Tiémassas ravine first stop.
Figure 291. Tiémassas marigot, 20 October 1988
Figure 292. Tiémassas marigot.
Figure 293. Tiémassas bridge section, sables/gravillons.
Figure 294. Tiémassas bridge section, detail.
Figure 295. Tiémassas disc cores found at site.
Figure 296. Sébikotane ravine entrance, 15 October 1988.
Figure 297. Sébikotane ravine entrance.
Figure 298. Sébikotane ravine east side.
Figure 299. Sébikotane biface on eastern slope.
Figure 300. Sébikotane ravine west side and Wm Ponty school.
Figure 301. Fann biface findspots (after Descamps 1979, Fig. 29).
Figure 302. Biface Faculté de Droit 1959 (after Descamps 1979, Fig. 8).
Figure 303. Bafoulabé. 1 and 4, disc/Levallois cores. 2, pseudo-levallois point. 3, disc core. 5, Levallois point. 6 and 7, Levallois flake/points.
Figure 304. Bafoulabé. 1, canted sidescraper. 2 - 4, Levallois points. 5, Levallois flake/point. 6, Levallois flake/blade. 8, flake/blade. Sébikotane. 7, 2-platform core.
Figure 305. Bafoulabé. 1 and 2, flake/blades. 3 and 4, flakes. Sébikotane. 5 and 6, 2-platform cores. 7 and 8, Levallois points. 9, transverse convex sidescraper.
Figure 306. Sébikotane. 1, double sidescraper. 2, convex sidescraper. 3, bifacial point. 4, canted sidescraper. 5, 8 and 9, convergent sidescrapers. 6, tanged point base. 7, bifacial foliate point. 10, transverse convex sidescraper. 11, notch.
Figure 307. Sébikotane. 1, chopper/axe. 2, endscraper. 3, retouched Levallois point.
Figure 308. Previously known sites along Senegal and Falémé river valleys (after Camara and Duboscq 1987, Fig. 1).
Figure 309. Diagrammatic section through Falémé river deposits (after Camara and Duboscq 1984, Fig. 4).
Figure 310. Takoutala. 1, Levallois core. 2, straight/concave sidescraper. 3, Levallois flake. 4, Levallois flake/blade. 5 and 6, disc cores.
Figure 311. Takoutala. 1, convergent sidescraper. 2, canted sidescraper. 3, Levallois point. Badoye. 4, convex sidescraper.
Figure 312. Map of Guinea.
Table 27. Fouta Djalon 14 sites summary statistics.
Figure 313. Fouta Djalon. Plan and section of site A1.(after Boriskovsky and Soloviev 1978, Fig. 10).
Figure 314. Fouta Djalon. Artefacts from sites A1 and A7. (after Boriskovsky and Soloviev 1978, Fig. 9).
Figure 315. African evolutionary and paleoclimatic changes (after de Menocal, 2011).
Figure 316. The roots of modern human behaviour (after Gibbons, 2018).
Figure 317. Middle Palaeolithic-Middle Stone Age regional variants in Africa (after Clark, 1992, Fig. 1).
Back Cover
Recommend Papers

The Middle Stone Age of Nigeria in its West African Context [1 ed.]
 1789691389, 9781789691382

  • 0 0 0
  • Like this paper and download? You can publish your own PDF file online for free in a few minutes! Sign Up
File loading please wait...
Citation preview

The Middle Stone Age of Nigeria in its West African Context Philip Allsworth-Jones

The Middle Stone Age of Nigeria in its West African Context

Philip Allsworth-Jones

Archaeopress Archaeology

Archaeopress Publishing Ltd Summertown Pavilion 18-24 Middle Way Summertown Oxford OX2 7LG www.archaeopress.com

ISBN 978-1-78969-138-2 ISBN 978-1-78969-139-9 (e-Pdf)

© P Allsworth-Jones and Archaeopress 2019 Cover illustration: Umaru Gol, on the road to Zenabi, 10 April 1978

All rights reserved. No part of this book may be reproduced, or transmitted, in any form or by any means, electronic, mechanical, photocopying or otherwise, without the prior written permission of the copyright owners.

This book is available direct from Archaeopress or from our website www.archaeopress.com

For my companions of those years Radovan and Emilia Subhash and Vibha George and Marjorie

There is no aim, there is no existence of aim There is the road, the road, the road Ghalib (1797-1869)

And through the wide world I went, wonders to hear William Langland (1332-1400) Piers Plowman

Contents Contents ��������������������������������������������������������������������������������������������������������������������������������������������� i List of Figures����������������������������������������������������������������������������������������������������������������������������������� iii Preface ������������������������������������������������������������������������������������������������������������������������������������������� xiii Chapter 1 The Middle Stone Age in West Africa: Introduction ���������������������������������������������������������1 Geography ............................................................................................................................................................. 1 Geology and Geomorphology ............................................................................................................................. 1 Vegetation and Climate ....................................................................................................................................... 5 Climatic and Environmental History ................................................................................................................ 7 Archaeological perspectives ............................................................................................................................. 12 Institutional Framework ................................................................................................................................... 15 References ........................................................................................................................................................... 16 Geography ...................................................................................................................................................... 16 Geology and Geomorphology ..................................................................................................................... 16 Vegetation and Climate ............................................................................................................................... 16 Climatic and Environmental History ......................................................................................................... 16 Archaeological perspectives ....................................................................................................................... 17 Institutional Framework ............................................................................................................................. 18 Chapter 2 The Middle Stone Age of Nigeria ������������������������������������������������������������������������������������19 Northern Nigeria ................................................................................................................................................ 20 Zenabi .................................................................................................................................................................. 26 Introduction .................................................................................................................................................. 26 Site situation and stratigraphy................................................................................................................... 26 (1) Earlier investigations........................................................................................................................ 26 (2) Results of work done in 1976-1978 ................................................................................................. 29 (3) 1976-78 and Jos Museum collections.............................................................................................. 36 (4) Nature of the industry ...................................................................................................................... 40 Conclusion ........................................................................................................................................................... 46 Mai Lumba ..................................................................................................................................................... 46 Yada Gungume ............................................................................................................................................. 56 Tibchi ............................................................................................................................................................. 65 Yelwa .............................................................................................................................................................. 71 Ningi Hills ...................................................................................................................................................... 76 Banke ............................................................................................................................................................. 82 Saminaka ........................................................................................................................................................ 93 Nok .................................................................................................................................................................. 96 Pingell ........................................................................................................................................................... 101 Rop ................................................................................................................................................................ 102 Summary............................................................................................................................................................ 103 Sangoan ....................................................................................................................................................... 103 Southern Nigeria .............................................................................................................................................. 107 The Bodija Formation................................................................................................................................. 108 Ajibode.......................................................................................................................................................... 116 Olude-Araromi ............................................................................................................................................ 117 Asejire ........................................................................................................................................................... 119 Summary............................................................................................................................................................ 120 References ......................................................................................................................................................... 120 Northern Nigeria ........................................................................................................................................ 120 Southern Nigeria......................................................................................................................................... 121

i

Chapter 3 The Middle Stone Age in West Africa���������������������������������������������������������������������������� 123 Cameroun .......................................................................................................................................................... 123 Northern Cameroun ................................................................................................................................... 123 Southern Cameroun ................................................................................................................................... 145 Niger ................................................................................................................................................................... 147 Adrar Bous ................................................................................................................................................... 147 Bilma ............................................................................................................................................................. 150 Seggédim ...................................................................................................................................................... 150 Mékrou Valley ............................................................................................................................................. 151 Ghana.................................................................................................................................................................. 151 Asokrochona, the Ghana Nautical College, and Tema .......................................................................... 154 Other sites in Ghana ................................................................................................................................... 163 Ivory Coast ......................................................................................................................................................... 167 Attinguié and Anyama ............................................................................................................................... 169 South-west Ivory Coast .............................................................................................................................. 175 Sierra Leone ...................................................................................................................................................... 176 Dated deposits in the mining areas ......................................................................................................... 177 Yengema cave.............................................................................................................................................. 178 The alluvial sequence in the mining areas ............................................................................................. 178 Artefacts recovered from the alluvial deposits ................................................................................... 180 Summary............................................................................................................................................................ 184 Mali ..................................................................................................................................................................... 184 Senegal ............................................................................................................................................................... 192 West and South ........................................................................................................................................... 196 North and East............................................................................................................................................. 208 Summary............................................................................................................................................................ 212 Guinea ................................................................................................................................................................ 212 Conclusion ......................................................................................................................................................... 216 References ......................................................................................................................................................... 216 Cameroun ..................................................................................................................................................... 216 Niger ............................................................................................................................................................. 217 Ghana ............................................................................................................................................................ 217 Ivory Coast ................................................................................................................................................... 218 Sierra Leone ................................................................................................................................................. 219 Mali ............................................................................................................................................................... 219 Senegal ......................................................................................................................................................... 219 Guinea ........................................................................................................................................................... 220 Chapter 4 West Africa: regional summary ������������������������������������������������������������������������������������ 221 References ......................................................................................................................................................... 225 Chapter 5 A wider perspective�������������������������������������������������������������������������������������������������������227 Climate ............................................................................................................................................................... 227 Archaeology ...................................................................................................................................................... 228 Human Evolution.............................................................................................................................................. 231 Conclusion ......................................................................................................................................................... 232 References ......................................................................................................................................................... 234 Climate.......................................................................................................................................................... 234 Archaeology ................................................................................................................................................. 234 Human Evolution ........................................................................................................................................ 235 Envoi ................................................................................................................................................................... 237

ii

List of Figures Figure 1. West Africa: Political boundaries. .............................................................................................................................2 Figure 2. West Africa: Geographical features...........................................................................................................................3 Figure 3. West Africa: Geology. ...................................................................................................................................................4 Figure 4. Vegetation zones of Nigeria (after Keay, 1965). ......................................................................................................5 Figure 5. Vegetation zones of West Africa................................................................................................................................6 Figure 6. Oxygen isotope stages and the palaeomagnetic time scale (after Smart and Frances, 1991, Fig. 9.1)...........9 Figure 7. Palynological and stratigraphic record at lake Bosumtwi (after Miller and Gosling, 2014, Fig. 2)...............11 Figure 8. Preferential and recurrent Levallois techniques (after Boëda, 1994, Fig. 176). ..............................................13 Figure 9. Nubian cores (1 and 2), ‘classic’ and Halfan cores (3 and 4) (after Van Peer, 1991, Fig. 3). ............................14 Figure 10. The study area north of the Jos Plateau. ..............................................................................................................19 Figure 11. The Younger Granite Ring Complexes of Nigeria (after Turner 1976, Fig. 1).................................................21 Figure 12. The Younger Granite Ring Complexes of Niger (after Ngako et al., 2006, Fig. 3). ..........................................22 Figure 13. The area north of the Jos Plateau: Physical Geography. ....................................................................................23 Figure 14. The area north of the Jos Plateau: Geology..........................................................................................................24 Figure 15. Alluvial tin-mining at Nok (after Fagg 1977, Frontispiece)...............................................................................25 Figure 16. Geological sketch map of Zenabi (after Bond, 1956, 201). .................................................................................27 Figure 17. Carbonised tree trunk in Zenabi No.1 Paddock (after Fagg, 1956, Fig. 6). ......................................................28 Figure 18. Zenabi sections A, B, and C (after Anozie, 1975, Fig. 16). ..................................................................................28 Figure 19. Zenabi Fulata, 23 March 1976. ...............................................................................................................................29 Figure 20. Zenabi Fulata, 23 March 1976, in situ find. ..........................................................................................................30 Figure 21. Zenabi Fulata, 23 March 1976, in situ find. ...........................................................................................................30 Figure 22. Zenabi Falls, March 1976.........................................................................................................................................30 Figure 23. Zenabi Falls, July 1976. ............................................................................................................................................31 Figure 24. Zenabi contour map showing mining leases 3741, 9282, and 5174..................................................................32 Figure 25. Zenabi plan as surveyed in 1976-1978 with position of sections 1, 2, and 3...................................................33 Figure 26. Zenabi sections 1-3, 1976 and 1978. ......................................................................................................................34 Figure 27. Zenabi Fulata, work in progress, July 1976. .........................................................................................................35 Figure 28. Zenabi Fulata, section1. ..........................................................................................................................................35 Figure 29. Zenabi Fulata, section 2. .........................................................................................................................................35 Figure 30. Zenabi Tudo, March 1976. ......................................................................................................................................36 Figure 31. Zenabi Fulata. 1, limace. 2, angle burin. 3, double sidescraper on Levallois flake/blade. [PAJ 1980 Fig. 1. ABC].....................................................................................................................................................................................38 Figure 32. Zenabi Fulata. 3, 1-platform flake/blade core. 5, endscraper. Zenabi Karara. 1, transverse convex sidescraper. 2, Levallois point. 4, disc core. [PAJ 1980 Fig. 2. ABCDE]. .......................................................................39 Figure 33. Zenabi Fulata 1976. 1 and 4, transverse and straight sidescrapers. Zenabi Fulata 1977, finds derived from 1976 section, upper tinwash. 2, limace. 3, flake....................................................................................................39 Figure 34. Zenabi Fulata 1978. Lower tinwash, upper series. 2, plain platform flake. 3, waste flake. Base of deep sounding. 1, core-like chunk. 4, flake/blade. 5, waste flake. .......................................................................................39 Figure 35. Zenabi Fulata 1977, finds derived from 1976 section, upper tinwash. Initial/preparatory core. ..............39 Figure 36. Zenabi, Jos Museum. MP56/33. Initial Levallois core......................................................................................39 Figure 37. Zenabi, Jos Museum. MP56/33. Levallois flake/blade core. ...........................................................................40 Figure 38. Zenabi, Jos Museum. C. Levallois flake core. ....................................................................................................40 Figure 39. Zenabi, Jos Museum. MP56/33. 1, Levallois core. Karara. 2, Levallois flake core.......................................40 Figure 40. Zenabi, Jos Museum. A47/10. 1, Levallois flake core. ZNP 1962. 2, Levallois point core. Z ‘west of spur’ 1948. 3, 1-platform flake/blade core. C47/10. 4, Levallois flake core. .......................................................................40 Figure 41. Zenabi, Jos Museum. A. 1-platform flake/blade core. C. 2, 1-platform blade core. ....................................41 Figure 42. Zenabi, Jos Museum. B47/10. 1 and 2, 1-platform flake/blade cores. Karara 1948. 3, 1-platform ‘boatshape’ core. A ‘in situ’. 4, 2-platform flake/blade core. ................................................................................................41 iii

Figure 43. Zenabi, Jos Museum. Karara ‘1/2 mile north of bridge’. Disc core................................................................41 Figure 44. Zenabi, Jos Museum. MP 1962 ‘L.G.’. Disc core. .................................................................................................41 Figure 45. Zenabi, Jos Museum. MP 1964. 1, disc core. MP56/34 ‘B.G.’. 2, disc core. Karara 1948. 3, disc core......42 Figure 46. Zenabi, Jos Museum. Fulata. 1, Levallois flake. C47/10. 2, Levallois flake. Karara. 3, Levallois flake/ point. A47/10. 4, Levallois blade. ......................................................................................................................................42 Figure 47. Zenabi, Jos Museum. ZR ‘gravel’ 1962. 1, Levallois flake. Karara 1948. 2, Levallois flake. MP56/33. 3, Levallois blade. D. 4, retouched Levallois flake. ZNP 1962. 5, Levallois flake/blade. ...............................................42 Figure 48. Zenabi, Jos Museum. A47/10. 1, overshot blade. C. 2, overshot flake. ZMP 56/33. 3, pseudo-levallois point. Karara. 4, pseudo-levallois point. 5, blade. .........................................................................................................42 Figure 49. Zenabi, Jos Museum. A. 1, limace. Karara 1948. 2, limace. 3, flake. .............................................................43 Figure 50. Zenabi, Jos Museum. ZNP 1962. 1, limace. ZR 1962. 2, endscraper. Karara. 3, endscraper. A. 4, endscraper..............................................................................................................................................................................43 Figure 51. Zenabi, Jos Museum. MP56/33. 1, transverse convex sidescraper. Karara. 2, multiple canted sidescraper. A. 3, multiple dihedral burin. 4, burin on retouched truncation. ........................................................43 Figure 52. Zenabi, Jos Museum. ‘No.1 middle gravels I/S’. 1, convergent sidescraper with alternate retouch. ZR 1962 ‘I/S gravels’. 2, notch. ZMP. 3, canted sidescraper. Karara. 4, double sidescraper. .......................................43 Figure 53. Zenabi, Jos Museum. MP56/33. 1, trihedral. Karara. 2, double sidescraper................................................44 Figure 54. Zenabi, Jos Museum. ZMP. 1, trihedral. B. 2, canted sidescraper on Levallois flake..................................44 Figure 55. Zenabi Fulata 1976. 1, canted sidescraper. 2, transverse convex sidescraper. 3, sidescraper on ventral surface. Zenabi Tudo 1968. 4, Levallois point. ................................................................................................................44 Figure 56. Zenabi Tudo. 1, disc core. 2, pseudo-levallois point. 3, Levallois flake. 4, Levallois blade. ......................44 Figure 57. Tibchi and Mai Lumba and their position in relation to the Tibchi and Yeli hills. .......................................47 Figure 58. Mai Lumba lease 988, 24 March 1976. ...................................................................................................................48 Figure 59. Mai Lumba lease 988, 24 March 1976. ...................................................................................................................48 Figure 60. Mai Lumba lease 963, ‘kafanchan’ working, June 1976. .....................................................................................49 Figure 61. Mai Lumba lease 988, 21 April 1978. .....................................................................................................................51 Figure 62. Mai Lumba lease 988, 21 April 1978. .....................................................................................................................51 Figure 63. Mai Lumba 1976-1978 lease 988, plan and sections 1-4. ....................................................................................52 Figure 64. Mai Lumba lease 988, pump in operation, 15 April 1978. ..................................................................................53 Figure 65. Mai Lumba lease 988, section 1, 23 April 1978. ...................................................................................................53 Figure 66. Mai Lumba lease 988, section 1, 23 April 1978. ...................................................................................................54 Figure 67. Mai Lumba lease 988, section 2, 24 April 1978. ...................................................................................................54 Figure 68. Mai Lumba lease 988, section 2, 25 April 1978. ...................................................................................................55 Figure 69. Mai Lumba lease 988, section 3, 29 April 1978. ...................................................................................................55 Figure 70. Mai Lumba lease 988, section 3, 29 April 1978. ...................................................................................................56 Figure 71. Mai Lumba. 1 and 2, Levallois blade and flake/blade. 3, Levallois blade core. 4, convergent sidescraper. [PAJ 1986 Fig. 9.2]. .................................................................................................................................................................57 Figure 72. Mai Lumba. 1, Levallois flake core [PAJ 1986 Fig. 9.4.1]. 2, flake/blade. 3 and 4, flakes. 5, disc core. 2-5, excavated 1978, 2-4 upper tinwash, 5 lower tinwash. .............................................................................................57 Figure 73. Mai Lumba. Initial core. .........................................................................................................................................58 Figure 74. Mai Lumba. Levallois cores. 1, flake/point core. 2-3, flake cores. 4, 2-platform flake/blade core. .........58 Figure 75. Mai Lumba. 1 and 2, 2-platform flake/blade cores, 2 struck from opposing directions. ............................58 Figure 76. Mai Lumba. 1 and 4, 2-platform flake/blade cores. 2 and 3, 1-platform flake/blade cores. ......................58 Figure 77. Mai Lumba. 1 and 2, disc cores. 3 and 4, 1-platform flake/blade cores.........................................................58 Figure 78. Mai Lumba. 1-4, disc cores. ...................................................................................................................................58 Figure 79. Mai Lumba. 1 and 4, Levallois points. 2-3 and 5, Levallois flake/blades. ......................................................59 Figure 80. Mai Lumba. 1-6, Levallois flakes and flake/blades. ...........................................................................................59 Figure 81. Mai Lumba. 1-5, Levallois flakes. ..........................................................................................................................59 Figure 82. Mai Lumba. 1, flake/blade. 2-4, flakes. ...............................................................................................................59 Figure 83. Mai Lumba. 2, blade. 1, 3-4, flakes. ......................................................................................................................60 Figure 84. Mai Lumba. 1 and 2, convex and double sidescrapers. 3, pseudo-levallois point. 4-5, overshot flake/ blades. .....................................................................................................................................................................................60 iv

Figure 85. Mai Lumba. 1 and 5, sidescrapers on ventral surface. 2 and 3, straight and convex sidescrapers. 4, Levallois flake/blade. ...........................................................................................................................................................60 Figure 86. Mai Lumba. 1 and 3, broken ground stone axes. 4, Mai Lumba Tudo, double/convergent sidescraper. .60 Figure 87. Mai Lumba. 1, transverse convex sidescraper. 2, convergent sidescraper. 3, broken ground stone axe.....61 Figure 88. Mai Lumba. Biface...................................................................................................................................................61 Figure 89. Mai Lumba lease 988, ground stone axe...............................................................................................................61 Figure 90. Mai Lumba lease 988, ground stone axes. ............................................................................................................61 Figure 91. Mai Lumba, disc and 2-platform Levallois flake/blade core. ............................................................................62 Figure 92. Mai Lumba, disc and Levallois flake core. ............................................................................................................62 Figure 93. Mai Lumba, 2-platform flake/blade cores. ..........................................................................................................62 Figure 94. Mai Lumba, microsyenite initial core. ..................................................................................................................63 Figure 95. Mai Lumba, rare raw materials at the site. ..........................................................................................................63 Figure 96. Tibchi 1976 paddock, plan and sections. ..............................................................................................................64 Figure 97. Tibchi, general view of paddock, 2 July 1976. .....................................................................................................65 Figure 98. Tibchi, section 1, 2 July 1976. .................................................................................................................................66 Figure 99. Tibchi, tin-miners removing lower tin-wash, 2 July 1976. ................................................................................66 Figure 100. Tibchi, section 2, 2 July 1976. ...............................................................................................................................67 Figure 101. Yada Gungume. 1, biface. Tibchi. 2, 2-platform core. 3, sidescraper on ventral surface. .......................67 Figure 102. Tibchi. 1, disc core. 2, Levallois/disc core. 3 and 4, disc cores. ....................................................................68 Figure 103. Tibchi. 1, Levallois flake core. 2, Levallois/disc core. 3, 1-platform core/chopper. 4, 1-platform core. ...69 Figure 104. Tibchi. 1, 1-platform core. 2, Levallois blade. 3, Levallois flake/blade. 4, overshot Levallois flake/ blade........................................................................................................................................................................................69 Figure 105. Tibchi. 1, blade. 2 and 5, flake/blades. 3 and 7, Levallois flakes. 4, convex sidescraper. 6, Levallois blade. 8, pseudo-levallois point. ........................................................................................................................................69 Figure 106. Tibchi. 1, pseudo-levallois point. 2, 5 and 6, flakes. 3, canted sidescraper. 4, transverse convex sidescraper. 7, blade.............................................................................................................................................................69 Figure 107. Tibchi. 1, convergent sidescraper/limace. 2, transverse convex sidescraper. 3, sidescraper on ventral surface. 4, straight sidescraper with thinned back...........................................................................................70 Figure 108. Tibchi. 1, double sidescraper/limace. 2, canted sidescraper/core. 3, limace. ..........................................70 Figure 109. Tibchi. 1, limace. 2, concave sidescraper. 3, biface. .......................................................................................70 Figure 110. Tibchi. Canted sidescraper/limace. ...................................................................................................................70 Figure 111. Tibchi. 1, limace. 2 and 3, transverse convex sidescrapers. ..........................................................................71 Figure 112. Tibchi, limaces (L and R), dorsal view. ...............................................................................................................71 Figure 113. Tibchi, limace (L), side view. ................................................................................................................................71 Figure 114. Tibchi, sidescrapers (L + centre), Levallois blade and pseudo-levallois point (R). ......................................71 Figure 115. Yelwa 1978 lease 4311, plan and section. ...........................................................................................................72 Figure 116. Yelwa village, March 1977. ...................................................................................................................................73 Figure 117. Yelwa lease 4311, general view, 10 April 1978. ..................................................................................................74 Figure 118. Yelwa lease 4311, mining work in progress, 10 April 1978..............................................................................74 Figure 119. Yelwa, Jos Museum. 1, limace. 3 and 4, transverse convex sidescrapers. 2, YMP 1962, blade. ...............74 Figure 120. Yelwa, Jos Museum. YMP LG 1962. 1, initial core. 2, 1-platform flake/blade core. Yelwa 56/31. 3, disc core. 4, convergent sidescraper. ................................................................................................................................75 Figure 121. Yelwa 1978. 1, 2-platform core. 2, disc core. 3, 1-platform flake/blade core. ...........................................75 Figure 122. Yelwa 1978. 1, 1-platform flake/blade core. 2, flake.......................................................................................75 Figure 123. Yelwa 1978. Initial/preparatory core, in situ in tinwash. ...............................................................................75 Figure 124. Ningi Hills: Physical Geography...........................................................................................................................76 Figure 125. Ningi Hills: Geology. ..............................................................................................................................................77 Figure 126. Koluki, 2 May 1978. ................................................................................................................................................78 Figure 127. Koluki, 2 May 1978. ................................................................................................................................................79 Figure 128. Old Ningi, 9 May 1978............................................................................................................................................79 Figure 129. Jigawa, 8 May 1978. ................................................................................................................................................80 v

Figure 130. Jigawa. 1, Levallois flake/blade core. 2 and 5, disc cores. 3 and 6, 1-platform flake/blade cores. 4, flake/blade. ............................................................................................................................................................................81 Figure 131. Jigawa. 1 and 2, disc cores. ..................................................................................................................................81 Figure 132. Jigawa. 1, disc core. 2 and 4, Levallois flakes. 3, Levallois blade. 5, flake. 6 and 7, blades......................81 Figure 133. Koluki. 1, 1-platform flake/blade core. Tsofon Ningi. 2, Levallois flake/blade core. Jigawa. 3 and 6, flakes. 4, straight sidescraper on Levallois flake/blade. 5, bifacial sidescraper. .......................................................81 Figure 134. Tabela. 1 and 4, 1-platform flake/blade cores. 2, disc core. 3, Levallois flake/blade core. 5, Levallois blade........................................................................................................................................................................................82 Figure 135. Tabela. 1, convex sidescraper. 2, disc core. 3, flake, in situ upper tinwash. 4, endscraper. 5, Levallois flake. ........................................................................................................................................................................................82 Figure 136. Jigawa, disc cores. ..................................................................................................................................................82 Figure 137. Banke, general view looking north, 27 March 1977. ........................................................................................83 Figure 138. Banke, map showing position of leases on river Marwan Karini. ..................................................................83 Figure 139. Banke, plan of paddocks in lease 13468 over the period 1976-1978, with sections 1-3 and location of samples for radiocarbon dating. ........................................................................................................................................84 Figure 140. Banke, NW and SE paddocks, 29 December 1976. .............................................................................................85 Figure 141. Banke, SW extension paddock, 8 April 1978. .....................................................................................................85 Figure 142. Banke, quartz handaxe (July 1976) and perforated stone artefact (June 1977). ..........................................86 Figure 143. Banke, SE paddock 29 December 1976, clay horizon above the upper tin-wash. ........................................86 Figure 144. Banke, SE paddock 27 March 1977, water pump in action. .............................................................................87 Figure 145. Banke, SE paddock 27 March 1977, clay horizon and lower tin-wash. ..........................................................87 Figure 146. Banke, SE paddock 27 March 1977, tree trunk fallen from clay horizon above the upper tin-wash. .......88 Figure 147. Banke, sections 1 and 2 (SE paddock) and 3 (SW extension paddock). .........................................................89 Figure 148. Banke, section 3 (SW extension paddock) with tree trunk above the upper tin-wash. .............................90 Figure 149. Banke, two cleavers found in April 1978. ...........................................................................................................91 Figure 150. Banke. 1 and 4, 1-platform flake/blade cores. 2 and 3, disc cores. 5, blade with marginal retouch/ sidescraper. ............................................................................................................................................................................91 Figure 151. Banke. 1, disc core. 2, endscraper. 3, ground stone axe in course of manufacture. 4, broken half of ground stone axe in course of manufacture. 5, Levallois blade. 6, notch/denticulate. 7, transverse sidescraper. ............................................................................................................................................................................91 Figure 152. Banke. 1, apparent roughout for ground stone axe. 2, 1-platform initial core. 3, perforated stone artefact. ..................................................................................................................................................................................92 Figure 153. Banke. Cleaver. ......................................................................................................................................................92 Figure 154. Saminaka, April 1982. ............................................................................................................................................93 Figure 155. Saminaka map (after Opadeji 2001, Fig. 2).........................................................................................................94 Figure 156. Saminaka sketch of cross-section (after Opadeji 2001, Fig. 13). .....................................................................94 Figure 157. Sketch map of the Nok Valley (after Fagg 1956, Fig. 9). ...................................................................................97 Figure 158. Hypothetical section of the Nok Valley (after Fagg 1956, Fig. 8). ..................................................................98 Figure 159. Nok valley, 27 March 1976. .................................................................................................................................100 Figure 160. Nok abandoned paddock, 27 March 1976.........................................................................................................100 Figure 161. Pingell, April 1976................................................................................................................................................101 Figure 162. Pingell, April 1976................................................................................................................................................102 Figure 163. Distribution of Sangoan sites in Northern Nigeria (adapted from Soper, 1965, 184-185)........................104 Figure 164. Sangoan tools (descriptions after Soper, 1965, Figs. E, G, and J). 1, Keffi, quartz push-plane. 2, Yumu, quartz handaxe. 3, Nassarawa, quartz handaxe. 4, Jebba, large quartzite pick. .....................................................105 Figure 165. Jebba. Locality 4, December 1983. .....................................................................................................................106 Figure 166. Jebba. Locality 3, December 1983. .....................................................................................................................107 Figure 167. Jebba. Locality 3, artefact horizon indicated by Arthur Harding. ...............................................................107 Figure 168. Jebba. Locality 3, artefact horizon. ...................................................................................................................108 Figure 169. Jebba. 1-platform quartzite core. ......................................................................................................................108 Figure 170. Jebba. sidescraper, pseudo-levallois point, two flakes. ................................................................................108 Figure 171. Jebba. two flake/blade cores, one round scraper...........................................................................................109 vi

Figure 172. Jebba. quartzite biface pick, quartz uniface pick. ..........................................................................................109 Figure 173. Jebba. quartzite pebbles/choppers. .................................................................................................................109 Figure 174. Geological Map of Ibadan (after Burke and Durotoye, and Kiladejo, 1980, Fig. 2). ...................................110 Figure 175. Sites in the Ibadan area (after Bagodo, 2012, Fig. 15). ...................................................................................111 Figure 176. Bodija Railway Cutting, 13 December 1979. ....................................................................................................112 Figure 177. Bodija Formation type section in the Bodija Railway Cutting, Ibadan (after Burke and Durotoye, 1971, Fig. 6).....................................................................................................................................................................................113 Figure 178. Bodija Railway Cutting, particle size distribution for Orita and Agodi members and weathered in situ gneiss (after Durotoye 1972, and Kiladejo, 1980, Fig. 11)..............................................................................................114 Figure 179. Sites in the vicinity of Ajibode (after Momin, 1995, Fig. 1, and Bagodo, 2012, Fig. 13). ...........................115 Figure 180. Olude-Araromi, general view, sections left of the road, 13 December 1979...............................................118 Figure 181. Bodija Formation sections at Olude-Araromi (after Kiladejo, 1980, Fig. 8). ...............................................119 Figure 182. Palaeolithic sites in Northern Cameroun (modified after Marliac 2006, Carte 9 bis). ..............................124 Figure 183. Northern Cameroun. Traverse number 4 (after Marliac 2006, Figure 2). ...................................................125 Figure 184. Mindif 80. 1, 2-platform flake/blade core. 2 and 4, Levallois points. 3, Levallois point core. 5, flake. 6 and 7, Levallois blades. [PAJ 1986 Fig. 9.6: 4-7]. ..........................................................................................................126 Figure 185. Mindif 80. 1, Levallois flake. 2, Levallois blade. 3, flake from disc core. 4, core edge removal flake. 5, flake. 6, flake from disc core. 7, convex sidescraper on Levallois point. 8, chopper. [PAJ 1986 Fig. 9.6: 1-3]. ........126 Figure 186. Dent de Mindif, 9 February 1980. ......................................................................................................................127 Figure 187. Dent de Mindif, 9 February 1980. ......................................................................................................................127 Figure 188. Dent de Mindif, ferruginous nodule horizon. .................................................................................................128 Figure 189. Dent de Mindif, ferruginous nodule horizon and artefacts. .........................................................................128 Figure 190. Katchel, 9 February 1980. ...................................................................................................................................129 Figure 191. Katchel, 9 February 1980. ...................................................................................................................................129 Figure 192. Northern Cameroun. Traverse number 1 (after Marliac 2006, Figure 3). ...................................................130 Figure 193. Stratigraphy of Mayo Louti site (after Digara 1988: 80). ...............................................................................131 Figure 194. Mayo Louti, general view, 11 February 1980. ..................................................................................................132 Figure 195. Mayo Louti, ‘basse terrasse’, 11 February 1980. ................................................................................................133 Figure 196. Mayo Louti, ‘moyennes terrasses’, 11 February 1980. .......................................................................................133 Figure 197. Mayo Louti, Douroumian ‘glacis’, 11 February 1980. ......................................................................................134 Figure 198. Mayo Louti. Levallois flake/blade, pseudo-levallois point, disc cores. ......................................................134 Figure 199. Mayo Louti. 1, initial core. 2 and 4, 1-platform flake/blade cores. 3, 1-platform core/chopper..........134 Figure 200. Mayo Louti. 1, 1-platform flake/blade core. 2, Levallois flake core. 3, Levallois point core. 4, disc core. 5, Levallois blade. .....................................................................................................................................................135 Figure 201. Mayo Louti. 1, Levallois blade core. 2, Levallois blade. 3, Levallois point. 4, Levallois flake/point. 5 and 6, Levallois point cores. 7, Levallois flake/point. ..................................................................................................135 Figure 202. Mayo Louti. 1, Levallois flake/point core. 2, Levallois blade. 3, Levallois flake core. 4, Levallois blade. 5 and 6, Levallois flake/points. .........................................................................................................................................135 Figure 203. Mayo Louti. 1, flake. 2, concave sidescraper on Levallois flake. 3, straight sidescraper. 4, pseudolevallois point. 5, convex sidescraper. ............................................................................................................................135 Figure 204. Mayo Louti. 1 and 6, concave sidescrapers. 2, concave-convex sidescraper. 3 and 4, pseudo-levallois points. 5, convex sidescraper. 7, endscraper.................................................................................................................136 Figure 205. Mayo Louti. 1, retouched Levallois point/Mousterian point. 3, convex sidescraper. Douroum. 2, flake. 4, Levallois blade. 5, Levallois flake. ....................................................................................................................136 Figure 206. Stratigraphy of Nassarao site (after Digara 1988: 186). .................................................................................136 Figure 207. Douroum, 8 February 1980. ................................................................................................................................137 Figure 208. Douroum, 8 February 1980. ................................................................................................................................137 Figure 209. Douroum. 1 and 3, disc cores. Sanguéré. 2, straight sidescraper. ............................................................138 Figure 210. Kossi, 6 February 1980.........................................................................................................................................138 Figure 211. Kossi, 6 February 1980.........................................................................................................................................139 Figure 212. Kossi, above the Mayo Kébi, 6 February 1980. .................................................................................................139

vii

Figure 213. Tongo. 1, 1-platform flake core. 2, Levallois flake/point core. Kossi. 3, flake. 4, convex sidescraper. 5, concave sidescraper. .......................................................................................................................................................140 Figure 214. Sanguéré, 5 February 1980. ................................................................................................................................140 Figure 215. Sanguéré, 5 February 1980. ................................................................................................................................140 Figure 216. Sanguéré. 1, 2-platform Levallois blade core. 2, disc/Levallois core. ........................................................141 Figure 217. Sanguéré. 1, Levallois flake core. 2, disc core. ...............................................................................................141 Figure 218. Sanguéré. 1, Levallois flake core. 2, Levallois blade. 3, Levallois blade core. .........................................141 Figure 219. Sanguéré. 1, Levallois point core. 2, flake. 3, concave sidescraper...........................................................141 Figure 220. Tongo. 1, disc core. 2, Levallois point. 3, flake. Sanguéré. 4, transverse convex sidescraper. 5, Levallois flake. 6, Levallois blade. ....................................................................................................................................142 Figure 221. Sanguéré, disc and Levallois cores. ...................................................................................................................142 Figure 222. Sanguéré, Levallois cores. ..................................................................................................................................142 Figure 223. Tongo, 5 February 1980. ......................................................................................................................................143 Figure 224. Tongo, 5 February 1980. ......................................................................................................................................143 Figure 225. Southern Cameroun. 1, Geological section at Locality 156. 2, Generalized geological section at Locality 114. 3, Locality 156, unifacial pick. (after Omi 1977, Figs. 24, 25.2, and 32). .............................................144 Figure 226. Southern Cameroun. Locality 114, pick. (after Omi 1977, Fig. 38). .............................................................145 Figure 227. Southern Cameroun. Locality 114. 1 and 2, picks. 3, unifacial scraper. (after Omi 1977, Fig. 39). ........146 Figure 228. Adrar Bous stratigraphic sections (after Clark et al., 2008, Fig. 4.1). ...........................................................148 Figure 229. Bilma stratigraphic section (after Maley et al., 1971, Fig. 5). 1, Holocene layers. 2, Calcareous diatomites. 3, Silémi terrace, including the following sequence from the top: (e) sand with calcareous nodules (d) lacustrine limestone with reed prints (Gif-1788 33,400±2500 BP) (c) coarse sandstone (in situ Levallois industry) (b) white aeolian sand (a) coarse sandstone (in situ Acheulean industry)..............................150 Figure 230. Map of Ghana, with principal sites indicated. ................................................................................................152 Figure 231. Asokrochona railway cutting, 28 December 1980. .........................................................................................154 Figure 232. Asokrochona railway cutting (after Davies 1967, Fig.1). ...............................................................................155 Figure 233. Asokrochona generalised section (after Nygaard and Talbot 1976, Fig.2). ................................................156 Figure 234. Ghana Nautical College, 28 December 1980.....................................................................................................157 Figure 235. Asokrochona [Oliver Davies]. 1, 2 and 4, disc cores. 3, Levallois flake/blade. ..........................................159 Figure 236. Asokrochona [Oliver Davies]. 1, core-axe. 2, blade. 3, convex sidescraper. [QCG Fig. 67.18, 21, 3]. ....159 Figure 237. Asokrochona [Oliver Davies]. 1, handaxe. [CQG Fig. 66.1]. 2, Levallois flake/point. .............................160 Figure 238. Asokrochona. 1, initial core. 2, disc core. 3, 2-platform flake/blade core. 4, 1-platform flake/blade core. [1, test pits I-IV, 2-4 Area II]. ...................................................................................................................................160 Figure 239. Asokrochona. 1, 3, 5 and 6, disc cores. 2, notch. 4, sidescraper on Levallois flake. [1 and 3-6, Area II, 4 Area I]. ...............................................................................................................................................................................160 Figure 240. Asokrochona. 1 and 5, bifaces in course of manufacture. 2, point/convergent sidescraper. 3, disc core. 4, convex sidescraper. 6, convergent sidescraper. 7, awl. [Area II]. ...............................................................160 Figure 241. Asokrochona. 1, convergent sidescraper. 2, denticulate. 3 and 5, chopping tools. 4, chopper. [Area II]. ...161 Figure 242. Asokrochona. 1, pick. 2, transverse convex sidescraper. [Area II]. ...........................................................161 Figure 243. Asokrochona. 1, handaxe. 2, point. [Area II].................................................................................................162 Figure 244. Asokrochona. 1, point. 2, concave sidescraper/notch. 3, awl. [Area II]. Tema [Oliver Davies]. 4, Levallois flake. 5, blade. 6, disc core. [QCG Fig. 57.7, 56.9, 56.7]. 7, disc/bipolar core [Tema West I 1976 Nygaard]. ..............................................................................................................................................................................162 Figure 245. Tema [Oliver Davies]. 1, handaxe/pick. 2, Levallois flake. 5, pseudo-levallois point. [QCG Fig. 56.16 and 17]. 3, convex sidescraper. 4, chopper. 6, chopping tool. [Tema West II 1976 Nygaard]. .............................162 Figure 246. Asokrochona quartz bifaces. ..............................................................................................................................162 Figure 247. Hohoe handaxe/pick. ..........................................................................................................................................164 Figure 248. Chawenu [Oliver Davies]. 1, Levallois flake/blade core. 2, Levallois flake. 3, biface in course of manufacture. 4, pseudo-levallois point. 5, endscraper. 6, chopping tool. .............................................................164 Figure 249. Chawenu [Oliver Davies]. 1 and 2, disc cores. Hohoe [Oliver Davies]. 3, handaxe/pick. .......................164 Figure 250. Ghana Nautical College. 1, disc core. 2, double sidescraper. [Nygaard 1976]. Narago [Oliver Davies]. 3, pseudo-levallois point. 4, retouched Levallois point. 5, Levallois flake. 6, retouched Levallois flake. [QCG Fig. 62.6, 62.8]. .............................................................................................................................................................................165 viii

Figure 251. Birimi location and topographic map (after Quickert et al. 2003, Fig. 1). ...................................................166 Figure 252. Birimi MSA profile east wall of main gully (after Quickert et al. 2003, Fig. 2). ...........................................167 Figure 253. Abidjan area, with Attinguié and Anyama. .....................................................................................................168 Figure 254. South-west Ivory Coast, with Gouabuo and Sablière. ....................................................................................168 Figure 255. Plan of Anyama (after Guédé and Tastet, 1986, Fig. 1c). ................................................................................170 Figure 256. Anyama, view to quarry looking west from roadside, 7 November 1983. ..................................................171 Figure 257. Anyama, west face of quarry with excavated area at top, 7 November 1983. ............................................172 Figure 258. Anyama, excavated area surface view to north, 7 November 1983. ............................................................172 Figure 259. Stratigraphy of Anyama (after Liubin and Guédé, 2000, Fig. 6.I). ................................................................173 Figure 260. Sierra Leone map. ................................................................................................................................................176 Figure 261. Yengema mining area. ........................................................................................................................................177 Figure 262. Yengema NDMC golf course, 16 October 1986. ...............................................................................................178 Figure 263. Yengema rest house and cave, 17 October 1986. ............................................................................................179 Figure 264. Cross section of a typical valley (after H.J.E. Haggard). .................................................................................179 Figure 265. Gbogbora river east of Fotingaya, 17 October 1986. .......................................................................................181 Figure 266. Gbogbora river ‘cooperative mining site’, 17 October 1986. .........................................................................181 Figure 267. Yengema broken bifaces. ....................................................................................................................................183 Figure 268. Yengema broken bifaces. ....................................................................................................................................183 Figure 269. Yengema complete bifaces. ................................................................................................................................183 Figure 270. Ground stone axe found by Mr A. Nabby in Yengema area. .........................................................................184 Figure 271. Quartz handaxe found by Mr H.H. Jackson in Koidu area. ............................................................................184 Figure 272. Map of Mali with Dogon plateau and sites. .....................................................................................................185 Figure 273. Ounjougou location and geomorphology (after Soriano et al. 2010a, Fig. 1). ............................................186 Figure 274. Yamé confluence of four rivers, 12 February 2001. ........................................................................................187 Figure 275. Yamé view upstream. ..........................................................................................................................................187 Figure 276. Oumounaama 4. ...................................................................................................................................................188 Figure 277. Oumounaama 5. ...................................................................................................................................................188 Figure 278. Kokolo 3. ................................................................................................................................................................189 Figure 279. Kokolo 3. ................................................................................................................................................................189 Figure 280. Kokolo 2. ................................................................................................................................................................190 Figure 281. Yamé campement, 14 February 2001. ...............................................................................................................190 Figure 282. Ounjougou archaeological levels (after Chevrier et al. 2018, Fig. 10). .........................................................191 Figure 283. Bandiagara escarpment and Songona (after Rasse et al. 2012, Photo 1). ....................................................192 Figure 284. Map of Senegal, showing position of more detailed charts of Senegal and Falémé river valleys and Cape Verde. ..........................................................................................................................................................................193 Figure 285. Cape Verde peninsula. ........................................................................................................................................194 Figure 286. Falémé river valley, recently excavated sites (after Lebrun et al. 2017, Fig. 1)...........................................195 Figure 287. Senegal river valley, recently excavated sites (after Scerri et al. 2015, Fig. 5). ..........................................196 Figure 288. Tiémassas excavation section (after Descamps 1979, Fig. 29). .....................................................................198 Figure 289. Tiémassas ravine first stop, 20 October 1988. .................................................................................................198 Figure 290. Tiémassas ravine first stop.................................................................................................................................199 Figure 291. Tiémassas marigot, 20 October 1988 ..................................................................................................................199 Figure 292. Tiémassas marigot. ...............................................................................................................................................200 Figure 293. Tiémassas bridge section, sables/gravillons. .....................................................................................................200 Figure 294. Tiémassas bridge section, detail. ......................................................................................................................201 Figure 295. Tiémassas disc cores found at site. ...................................................................................................................201 Figure 296. Sébikotane ravine entrance, 15 October 1988. ................................................................................................203 Figure 297. Sébikotane ravine entrance. ..............................................................................................................................203 Figure 298. Sébikotane ravine east side. ...............................................................................................................................204 Figure 299. Sébikotane biface on eastern slope...................................................................................................................204 ix

Figure 300. Sébikotane ravine west side and Wm Ponty school. ......................................................................................205 Figure 301. Fann biface findspots (after Descamps 1979, Fig. 29).....................................................................................206 Figure 302. Biface Faculté de Droit 1959 (after Descamps 1979, Fig. 8). .........................................................................207 Figure 303. Bafoulabé. 1 and 4, disc/Levallois cores. 2, pseudo-levallois point. 3, disc core. 5, Levallois point. 6 and 7, Levallois flake/points. .........................................................................................................................................207 Figure 304. Bafoulabé. 1, canted sidescraper. 2 - 4, Levallois points. 5, Levallois flake/point. 6, Levallois flake/ blade. 8, flake/blade. Sébikotane. 7, 2-platform core. ...............................................................................................207 Figure 305. Bafoulabé. 1 and 2, flake/blades. 3 and 4, flakes. Sébikotane. 5 and 6, 2-platform cores. 7 and 8, Levallois points. 9, transverse convex sidescraper. ......................................................................................................208 Figure 306. Sébikotane. 1, double sidescraper. 2, convex sidescraper. 3, bifacial point. 4, canted sidescraper. 5, 8 and 9, convergent sidescrapers. 6, tanged point base. 7, bifacial foliate point. 10, transverse convex sidescraper. 11, notch. .......................................................................................................................................................208 Figure 307. Sébikotane. 1, chopper/axe. 2, endscraper. 3, retouched Levallois point...............................................208 Figure 308. Previously known sites along Senegal and Falémé river valleys (after Camara and Duboscq 1987, Fig. 1). ...209 Figure 309. Diagrammatic section through Falémé river deposits (after Camara and Duboscq 1984, Fig. 4). ..........209 Figure 310. Takoutala. 1, Levallois core. 2, straight/concave sidescraper. 3, Levallois flake. 4, Levallois flake/ blade. 5 and 6, disc cores...................................................................................................................................................210 Figure 311. Takoutala. 1, convergent sidescraper. 2, canted sidescraper. 3, Levallois point. Badoye. 4, convex sidescraper. ..........................................................................................................................................................................210 Figure 312. Map of Guinea. .....................................................................................................................................................213 Figure 313. Fouta Djalon. Plan and section of site A1. (after Boriskovsky and Soloviev 1978, Fig. 10). .....................214 Figure 314. Fouta Djalon. Artefacts from sites A1 and A7. (after Boriskovsky and Soloviev 1978, Fig. 9). ...............215 Figure 315. African evolutionary and paleoclimatic changes (after de Menocal, 2011). ..............................................227 Figure 316. The roots of modern human behaviour (after Gibbons, 2018). ....................................................................233 Figure 317. Middle Palaeolithic-Middle Stone Age regional variants in Africa (after Clark, 1992, Fig. 1). ...............233

x

List of Tables Table 1. The pluvial and inter-pluvial system as assumed to exist in Africa (after Clark, 1959, Table 2). ......................8 Table 2. Zenabi Section 1 stratigraphy 1976-1978. ................................................................................................................34 Table 3. Zenabi artefact inventory............................................................................................................................................37 Table 4. Zenabi artefact measurements...................................................................................................................................45 Table 5. Mai Lumba lease 988. Stratigraphy, sections 1-4. ...................................................................................................50 Table 6. Mai Lumba and Yada Gungume. Artefact inventory. .............................................................................................57 Table 7. Tibchi 1976 paddock. Stratigraphy sections 1 and 2. ..............................................................................................67 Table 8. Yelwa and Tibchi. Artefact inventory. ......................................................................................................................68 Table 9. Yelwa lease 4311. Stratigraphy. .................................................................................................................................73 Table 10. Ningi Hills. Artefact inventory. ...............................................................................................................................78 Table 11. Banke lease 13468. Stratigraphy, sections 1-3. .....................................................................................................90 Table 12. Banke lease 13468. Artefact inventory. ..................................................................................................................92 Table 13. Saminaka Stone Age (SSA) Artefact Inventory. .....................................................................................................95 Table 14. Nigerian Sangoan assemblages (after Soper, 1965, Table III). ...........................................................................104 Table 15. Bodija Formation. Stratigraphic column (after Durotoye, 1976, Table 1). .....................................................113 Table 16. Ajibode UMF site. Stratigraphy of Trench D. .....................................................................................................116 Table 17. Ajibode UMF site. Dimensions of flakes, cores, points, and picks. ..................................................................117 Table 18. Olude-Araromi. Stratigraphy with major units identified. ..............................................................................118 Table 19. Mayo Louti and Nassarao. Artefact inventory. ....................................................................................................132 Table 20. Seggédim. Artefact inventory. ................................................................................................................................150 Table 21. Asokrochona artefact totals....................................................................................................................................158 Table 22. Asokrochona tool classes ........................................................................................................................................158 Table 23. Anyama I-IV layer D and Gouabuo I and II ...........................................................................................................174 Table 24. Anyama I-III layer C and Sablière ..........................................................................................................................175 Table 25. Yengema. Dimensions (means + SDs) of bifacial artefacts. ................................................................................182 Table 26. Four artefact inventories from Senegal ................................................................................................................197 Table 27. Fouta Djalon 14 sites summary statistics. ............................................................................................................213

xi

xii

Preface This book has been a long time in the making. Its origins go back to the author’s fieldwork in Northern Nigeria, conducted for the most part in the 1970’s, when he was a staff member of the University of Ibadan. The work was supported logistically and financially by the University and also by the Nigerian Federal Department of Antiquities (now the National Commission for Museums and Monuments). Thanks go to all those involved in the fieldwork, particularly to our regular driver Bernard Njoku and to Umaru Gol, my indispensable right hand man. In the field he knew exactly what was required, he could always be relied upon, and his occasional pithy remarks on life in general always stuck in the mind. The Bisichi Jantar Mining Company, on whose leases we worked, provided help in many ways, including the provision of accommodation in their rural rest houses at Lerebi and Kalatu. Thanks go to the General Manager Mr Thomson, and to his local managers in the field, Messrs Coker and Busari. My subsequent visits to the other countries of West Africa were financed by myself, apart from three (much appreciated) small grants from the Boise Fund, the British Prehistoric Society and the Cambridge Philosophical Society, to assist me respectively in Senegal and Sierra Leone. I have always taken the view, going back to my Ph.D. days, that there is no substitute for a personal acquaintance with the countries and with the material which are the object of your study. In carrying out these visits, I had help from many colleagues and institutions. In Cameroun, I thank especially Alain Marliac, without whose friendly cooperation I could have done nothing, and also Mohammadou Eldridge, then at the museum in Garoua. Professor John Sutton kindly facilitated my visit to Ghana, and thanks go to those who helped me locate the material I needed in the Department of Archaeology storeroom, Messrs Agyei-Henaku, Adjedu, and Murey. In the Ivory Coast, I had a great deal of help from François Guédé Yodé, and I thank Professor J.P. Tastet and Robert Chenorkian for accompanying me to Anyama. I later had the opportunity to discuss the Sangoan industry from this site with Professor V.P. Liubin, a man whose devotion to the study of prehistory was unparalleled. My visit to the diamond mines of Sierra Leone would have been impossible but for the assistance of all the staff of the NDMC, particularly Professor Victor Strasser-King, Messrs Koroma and Mantell, and the local staff who showed me round the mining areas. I am grateful to Mrs Dorothy Cummings, then Director of the Cotton Tree Museum in Freetown, for allowing me access to the collections there. In Senegal, I was fortunate to benefit from the great kindness and hospitality of Abdoulaye Camara and his family, and our visits to some of the sites in the vicinity of Dakar were greatly facilitated by M. L. Hébrard. In Mali in 2001, I was able to join a small group led by Professor Rudolph Kuper on a visit to Ounjougou and other sites, after the PAC meeting in Bamako, when we were shown great hospitality by Eric Huysecom and his team. I subsequently was able to discuss the sequence at Ounjougou with Sylvain Soriano on the occasion of his visit to Sheffield in 2009. For many details in all these countries I am of course dependent on published literature, and also on a number of unpublished materials which I have been able to consult. Most of the photographs in this work are my own, as are the drawings, unless otherwise indicated. My drawings no doubt are far from perfect, but in executing them I have endeavoured to follow the example of the (far superior) Pierre Laurent, who was kind enough to spend some time with me in Les Eyzies, a long time ago, showing me how he did it. The purpose of this work is fairly self-evident. It is to make known the Palaeolithic sequence in West Africa, insofar as it concerns the Middle Stone Age, in which is included the Sangoan, in the form of a monograph. The sequence in West Africa is not widely known to scholars working in other parts of the continent, so much so that it is sometimes assumed that there is no such sequence. There are many reasons for this neglect, partly because there is a scarcity of spectacular finds such as have been made elsewhere, but not least because of inadequate (or no) publication of such finds as have been made. The author is not altogether free of blame in this regard. A number of his reports concerning particular aspects have been published, relating both to Nigeria and to other countries, but some of these are in relatively inaccessible journals, and this is the first general account of the subject which he has been able to produce. Life itself got in the way, so I regret the delay, but I hope that this book will nonetheless serve its intended purpose. xiii

xiv

Chapter 1

The Middle Stone Age in West Africa: Introduction In order to comprehend fully the archaeology of Nigeria in particular and West Africa in general, it is necessary to have a broad understanding of the region’s geography and environment, including its environmental history so far as that is possible. In addition, something should be said about the institutional framework within which archaeology has been conducted in the region, and some definitions of archaeological terminology are required. This introduction is intended to provide the necessary background on these topics, before proceeding to a detailed discussion of the archaeological evidence, in a country-by-country fashion. Geography Politically speaking, West Africa is generally taken to correspond to the 15 countries now grouped together as ECOWAS (Economic Community of West African States). This comprises the former British colonies of Gambia, Ghana, Sierra Leone, and Nigeria, the former states of French West Africa (except for Mauritania), Liberia, Cape Verde, and Guinea-Bissau (Rochebrune and Sablayrolles, 2000). For the purposes of this work, Cameroun and Chad (formerly part of French Equatorial Africa) are also included, since so far as prehistory is concerned, they have close links to the countries to their west. The current boundaries of all these countries are shown in the map at Figure 1. The map at Figure 2 shows the physical geography of the region. Here as in the other maps the northern boundary is taken to be the Tropic of Cancer (23.5° N). For the most part, as remarked by Harrison Church (1963), the land lies between 600 and 1600 feet (or about 200-500 metres) and constitutes a ‘worn monotonous and fairly level surface’. Notable areas of higher relief include the Fouta Djalon and Guinea Highlands, the Jos Plateau, the Adamawa, Bamenda, and Cameroun Highlands, and, in the north, the high massifs of the Aïr Mountains and the Adrar des Iforas. A number of important rivers flow through the region, most notable the river Niger, which is some 2600 miles (or about 4000 km) long. The Upper Niger was not originally connected with the Lower Niger, but joined it in the Quaternary thanks to a considerable flow of water down the course of the Tilemsi and the partial silting up of the Araouane Lake. The Inland Delta on the Upper Niger is still a reminder of this former lake. Other important water courses include the Senegal, Volta, and Benue rivers, as well as the Chari which drains into Lake Chad. Nonetheless, as Harrison Church (1963) puts it, neither the areas of higher land nor the rivers ‘do much to upset the [north-south] zonal arrangement of climatic and vegetational belts’ which are such a characteristic and dominant feature of the West African landscape. Geology and Geomorphology The present physical geography of the region evidently reflects its underlying geology to a considerable extent. The geological map at Figure 3 is based upon the ‘esquisse structurale’ produced by Furon and his colleagues in 1958, as well as the book by Furon himself (Furon et al., 1958; Furon, 1963). Some of the details may well have been subject to revision in the meantime (Schlüter, 2006) but in its general lines this map is perfectly adequate for our purposes. Attention may be drawn to some outstanding features of the map, as follows. In the first place, it is obvious that Pre-Cambrian rocks are widespread, occupying about one third of the whole area. They are folded in places, often in a northeast to south-west direction, as in the case of the Atacora anticline. Palaeozoic deposits are also widely present, notably in Ghana (where the Voltaian Supergroup is alternatively regarded as Proterozoic), in southern Mali and in Guinea, as well as further north where they form part of the Taoudéni syncline (one of the largest such features in the world). The Benue Trough contains marine Cretaceous sediments, which also occur west of the Adrar des Iforas, both witnessing an extensive marine transgression which 1

Figure 1. West Africa: Political boundaries.

The Middle Stone Age of Nigeria in its West African Context

2

Figure 2. West Africa: Geographical features.

The Middle Stone Age in West Africa: Introduction

3

Figure 3. West Africa: Geology.

The Middle Stone Age of Nigeria in its West African Context

4

The Middle Stone Age in West Africa: Introduction

once formed a ‘Saharan sea’ (Furon, 1963, Fig. 1) connected with the Tethys ocean in the area of what is now the Mediterranean. About 1000 km long, the Trough also forms part of a broader Central African Rift system. The Tertiary and Quaternary sediments are of more direct concern to us. In general (apart from an important occurrence around Thiès in Senegal) they are of Continental origin. In addition to significant coastal presence (notably in Senegal, the Ivory Coast, and the Niger delta), the large areas of such deposits in the once vast Chad basin, and to a lesser extent along the upper Niger, are of obvious importance. Igneous rocks (both extrusive and intrusive) are also of major importance, particularly in the area of the Cameroun Highlands, where there is an impressive system of faults (Furon, 1963, Fig. 21). The system extends out to sea, as far as São Tomé and Príncipe. Finally, clearly marked on the map, are the Younger Granite volcanic occurrences in the area around the Jos Plateau, in the Aïr, and the Adrar des Iforas. They are of particular significance in relation to the Nigerian MSA, and they will be referred to in detail later. Indispensable works of reference for interpreting the landforms which developed in this environment are provided by Thomas (1974) and Tricart (1972), including the sometimes tricky question of the equivalence (or non-equivalence) of the respective terms in English and French. Vegetation and Climate A classic account of the vegetation of Nigeria was written by R.W.J. Keay (1965, 3rd edition) and in this case it may provide a suitable starting point for a consideration of the vegetation of West Africa as a whole. A simplified version of Keay’s map is at Figure 4, and a map of the entire region is at Figure 5, the latter based upon the work of Aubréville et al. (1958).

Figure 4. Vegetation zones of Nigeria (after Keay, 1965).

5

Figure 5. Vegetation zones of West Africa.

The Middle Stone Age of Nigeria in its West African Context

6

The Middle Stone Age in West Africa: Introduction

It can be seen that the broad lines of Keay’s system do continue throughout the region, not surprising, since the principle of division into Sahel, Sudan and Guinea Savanna zones was already proposed by the French botanist Auguste Chevalier in 1900, and it has proved reliable. As Keay points out, the vegetation of Nigeria (and West Africa as a whole) is determined by climate, in particular by the mean annual rainfall and the severity of the dry season. Forest (High Forest or Rain Forest) is defined as vegetation dominated by woody species in open or closed canopy, from which grasses are virtually absent. Most of the trees are not fire-tolerant. Savanna has grass dominant in the field layer, but also has woody species with some degree of fire-tolerance. Derived savanna, shown as a separate zone on both maps, indicates a belt of country where, as Keay put it, ‘the combined effects of native agriculture and grass fires can bring about, and indeed undoubtedly have brought about, the degradation of forest to a savanna type’. The purely climatic limit of continuous forest (ignoring the effects of degradation) is determined by two factors, a mean annual rainfall of at least 48 inches (1219 mm), and a lowest mean monthly relative humidity of not less than 70%. Mangrove (Rhizophora) and fresh water swamp are not indicated separately on the map of Nigeria, but they are on the map of West Africa. Keay, unlike Chevalier and other authors, divided the Guinea Savanna into a southern and a northern zone, although they have here been amalgamated into one. Clearly the northern zone is drier than the southern, and is characterised by such trees as Isoberlinia doka, Isoberlinia dalzielii, Monotes kerstingii, and Uapaca togoensis. The Sudan zone is drier still, with much of the soil being loose and sandy, and there are far more thorny plants than before, most of them species of Acacia. The 20 inches (508 mm) isohyet more or less marks the boundary between the Sudan and Sahel zones, with various types of Acacia again abundant in the Sahel. So far as Nigeria is concerned, montane vegetation is confined to a small area on the boundary with Cameroun (as well as in Cameroun itself), although as shown on the map of the entire region, it does also exist in patches in the Fouta Djalon and Guinea Highlands. The lower slopes of the mountains, up to about 6000 feet (1830 metres), are forested but thereafter the forest gives way to grassland, which is of steppe rather than savanna type. Keay is at pains to emphasise that the ‘ideal’ vegetation pattern has been considerably modified by human action everywhere (not only in the derived savanna) but this no doubt concerns at most the last few thousand years. Further information to supplement the above is given by Dalziel (1937), Keay at al. (1964), Lawson (1966), and Hopkins (1974). The broader African picture is described in detail by White (1986), which has the advantage of providing French equivalents for the relevant English terminology. The close connection between vegetation and climate is clear from the foregoing account, but as anyone who has lived in West Africa will testify, the climatic conditions through the year (as reflected in the vegetation) are anything but static. The seasonal contrasts, in terms of average monthly rainfall, are strong (Thompson, 1975, Fig. 2). Two opposing air masses confront each other throughout the year, equatorial maritime from the south-west giving a maximum rainy season in August, and tropical continental from the north-east (the harmattan) producing a dry season which is at its strongest in January. The moving frontier between the two is the Inter-Tropical Convergence Zone. In January the position of this front is at about 5-7° N and in August it is at about 17-21° N. The rain accompanying the movement of the front is usually associated with line squalls, normally arriving from the east (Hopkins, 1974). Apart from the mean annual rainfall, its seasonal distribution is equally important. Harrison Church (1963) remarks that 4 inches (102 mm) monthly rainfall is sufficient for most plant growth. His Fig. 19 (‘Number of months with at least 4 inches of rain’) presents a map showing how many months in the year fulfil this requirement in the region as a whole, based on readings from 44 weather stations. The resulting measurements reveal four east-west trending bands, with respective totals of 1-3, 3-5, 5-7, and >7 months. Not surprisingly, these bands bear a close resemblance to the vegetation zones just discussed: Sahel, Sudan Savanna, Guinea Savanna, and Rain Forest respectively. These contemporary clues to the inter-connection between vegetation and climate are surely also of relevance when considering past environmental conditions. Climatic and Environmental History Those investigating the history and archaeology of West Africa have never doubted that its climate varied over time, but the models used or assumed to explain it have themselves been subject to drastic 7

The Middle Stone Age of Nigeria in its West African Context

Table 1. The pluvial and inter-pluvial system as assumed to exist in Africa (after Clark, 1959, Table 2). Geological stage

Climate

Recent

Present Climate 2nd Post-Pluvial Wet Phase

Epi-Pleistocene Upper Pleistocene

Middle Pleistocene

Lower Pleistocene

Climatic stage Present Nakuran

Dry 1st Post-Pluvial Wet Phase

Makalian

Dry Fourth Pluvial

Gamblian

Very Dry Third Pluvial

Kanjeran

Dry Second Pluvial

Kamasian

Very Dry First Pluvial

Kageran

change. What has been written, in particular by scholars of an earlier generation, cannot be understood (or reappraised) if these assumptions are not clearly understood. This question was addressed in a recent paper (Allsworth-Jones, 2016), so the main points only will here be summarised. In the decades up to and after the second world war, there was a widespread assumption that, broadly speaking, pluvial (or rainy) periods in Sub-Saharan Africa corresponded to glacial episodes in the temperate northern hemisphere (Oakley, 1964: 85; Clark, 1959: 3335, Table 2). Clark’s diagram, reproduced here as Table 1, provides a handy summary.

It can be seen that the pluvial periods were divided from each other by dry or very dry phases. The names of the pluvials, as well as inter-pluvials and post-pluvials, were taken from localities in East Africa (Bishop, 1967). Thus the Gamblian or fourth pluvial corresponded to the last glacial period. It was preceded by the third inter-pluvial, or very dry period, corresponding to the last interglacial. Two wet phases of lesser intensity referred to as the first and second post-pluvials (or Makalian and Nakuran) came after the Gamblian, and were separated from it and from each other by two arid episodes. The Makalian was regarded as Epi-Pleistocene, but the Nakuran was Holocene. So far as the archaeological record is concerned, it was assumed that the Middle Stone Age in Sub-Saharan Africa corresponded chronologically to the Upper Palaeolithic in North Africa and Europe (Clark, 1970, Figure 1, Chronological table of African prehistory). Hence it was a ‘laggard’ and not a ‘leader’ (Clark, 1975). This scheme was assumed to hold good for West Africa, and it is reflected in the writings of Oliver Davies, who was concerned above all with Ghana, and in the report by Geoffrey Bond (1956), who was invited by Bernard Fagg to examine the sequences which he had identified at Zenabi and Nok in northern Nigeria. On the basis of radiocarbon dates obtained at these sites, it was suggested that ‘the Middle Stone Age may have survived into the Makalian interval in this part of Africa’ (Barendsen et al. 1957), and this was not regarded as anything unusual. The circumstances at these sites are considered in more detail below. The idea that pluvials could provide a continent-wide standard of comparison and that they could be equated with glaciations in the northern hemisphere did not, however, survive further investigation. Karl Butzer, after a careful examination of the available evidence, came to the unequivocal conclusion that ‘the African pluvial record is far too complex to serve as a basis for stratigraphic correlation’ (Butzer 1971: 350-351). The classical East African pluvial chronology, and terminology, was ‘based on false premises and incorrect deductions; it should therefore be abandoned entirely’. This did not mean that there were no drier and wetter phases which could be discerned in the various African regions through time, nor that correlation on geological grounds was inadmissible as such. In particular, he agreed there was good evidence for a prolonged and widespread dry episode corresponding to the last glacial period in many parts of the continent (cf. Butzer 1978: 208), and two wetter phases in the earlier part of the Holocene could be clearly discerned in the Sahara. In fact, many of the phenomena once classified as Gamblian (including the type site) could be regarded as part of this post-Pleistocene sequence of events. The idea of a ‘dry phase south of the Sahara 20,000 years ago’ was also championed by Burke and his colleagues (Burke et al. 1971, 1972). As they said, this scenario is the opposite of what was previously accepted, and it is of particular relevance in West Africa, where for example there is clear evidence that the lower valley of the Senegal river was blocked by great dunes during a time period referred to by Francophone authors as the Ogolian.

8

The Middle Stone Age in West Africa: Introduction

The loss of pluvials as a means of continent-wide correlation has been more than compensated for by the application of palaeomagnetism and oxygen isotope analysis that together have produced a chronostratigraphy which in principle has world-wide validity (Imbrie, 1979; Smart and Frances, 1991; Lowe and Walker, 1997). Palaeomagnetic stratigraphy involves the measurement of the natural remanent magnetism of sediments or rocks. Bernhard Brunhes in 1906 and Motonori Matuyama in 1929 observed that cooling lava flows acquired a direction of magnetization parallel to the earth’s magnetic field. To their surprise, the direction of some ancient lava flows showed not ‘normal’ but ‘reversed’ polarity, in other words there must have been a dramatic switch when (as it were) north became south and south became north. Subsequent work by many others showed that over the past two million years there have been at least four episodes of reversed polarity separated by periods of normality. We can reliably estimate the times of transition thanks to potassium-argon (K-Ar) and argon-argon (Ar-Ar) dating of several of the key locations. One important reversal took place about 735,000 years ago, and in honour of the first researchers this has been named the Brunhes-Matuyama boundary. Several archaeological sites in East Africa, where volcanic eruptions were frequent, have been dated in this way. Signs of polarity may also be detected in the sediments of marine cores, and in this way it acts as a vital adjunct to oxygen isotope analysis. There are two main isotopes of oxygen, 18O and 16O, the former being heavier (as well as less abundant) than the latter. Both occur in marine cores, in differing frequencies, since they were incorporated into the shells of foraminifera, and these in turn reflected the ratios of the two isotopes which were prevalent in the sea water at the time. Nicholas Shackleton and his colleagues (following on earlier work by Emiliani) showed why the variations occurred and what they meant in terms of Pleistocene geochronology. They showed that a mechanism was at work which reflected the alternation of glacial and interglacial events. The lighter isotope 16O is more easily picked up by evaporation, and during colder periods it tended to be precipitated and then trapped in the extended glaciers. During warmer periods when the ice sheets melted the 16O would be returned to the oceans. Thus, during the cold periods the oceans were isotopically ‘heavier’ and in the warm periods they were isotopically ‘lighter’. This was demonstrated by Shackleton and Opdyke (1973) on the basis of Pacific core V28-238, which also had palaeomagnetic evidence showing the Brunhes-Matuyama boundary. The ratios of the two isotopes are measured not in absolute terms but as relative deviations (δ18O per mil) from a laboratory standard value. A succession of peaks and troughs were detected, and these were labelled Oxygen Isotope Stages 9

Figure 6. Oxygen isotope stages and the palaeomagnetic time scale (after Smart and Frances, 1991, Fig. 9.1).

The Middle Stone Age of Nigeria in its West African Context

1-22, up to and beyond the Brunhes-Matuyama boundary, where the odd numbers correspond to warmer periods and the even numbers to colder ones. The authors concluded (Shackleton and Opdyke, 1973: 48) that their scheme was sound, inasmuch as it depended on ‘a phenomenon, isotopic change in the oceans, that must occur essentially synchronously’. ‘It is highly unlikely that any superior stratigraphic subdivision of the Pleistocene will ever emerge. We propose that the stages set up in this core be adopted as standard for the latter half of the Pleistocene’. This is how it has turned out to be. The oxygen isotope curve was later extended further back into the Pleistocene, and this, together with a comparison to the palaeomagnetic time scale, is shown in Figure 6 (Smart and Frances, 1991, Fig. 9.1). It has also been demonstrated that the regularities observed in such diagrams can be correlated with long-term changes in the earth’s axis and its orbit around the sun, thus vindicating the general theory advanced in the early years of the 20th century by Milutin Milankovitch (Imbrie, 1979). In the following account, the evidence currently available for dating the MSA and Sangoan sites in West Africa will be summarised and assessed, but for the moment we are concerned only with the climatic and environmental aspect. The impact made by the study of marine cores, and the importance of movement in the Inter-Tropical Convergence Zone, may be illustrated by reference to the sequence of events now established for the end of the Pleistocene and the beginning of the Holocene in the region. Thurstan Shaw, in his study of the LSA at Iwo Eleru in the south of Nigeria (7° 26’ N 5° 7’ E) had proposed that its occupation began in savanna conditions and persisted in a forested environment (Shaw and Daniels, 1984; Allsworth-Jones et al., 2010). The uncalibrated radiocarbon dates for the occupation of the cave range from 11,200±200 to 3465±65 BP, marking a change from an aceramic to a ceramic way of life. The results of the marine core studies and the climatic reconstructions based upon them bear out the accuracy of his hypothesis. Thus, deMenocal and his colleagues have established a framework for the nature and longevity of what they term the African Humid Period, which followed the end of the Pleistocene, on the basis of the marine core at ODP site 658C off Cap Blanc, Mauritania (deMenocal et al., 2000). Humid conditions initially commenced about 14,800 years ago, with the main episode of the AHP occurring between about 9,000 and 5,500 years ago. It is likely that during the preceding Late Glacial Maximum (equivalent as we have seen to the Ogolian) the southern boundary of the Sahara may have been situated at about 14°N, whereas in the early Holocene the northern boundary of the forest may have reached as far as 10-12°N (Dupont et al., 2000). Lézine and Cazet (2005, Fig. 5) have proposed a step-wise model for the northward expansion of the forest in West Africa during the AHP, extending from about 11,600 to 9,300 years ago. There was a similarly uneven retreat at the end of the period. There is some disagreement about the extent to which the forest was reduced to fragments along the coast during the LGM, but the principle of its expansion is not in doubt. The advance of the forest in the AHP was mirrored by a rise in lake levels, including Lake Chad (Leblanc et al., 2006). The lake may have begun to fill as far back as 10,160±160 BP and may not finally have begun to retreat until 3,000±110 BP. It is estimated that it will have been at its maximum extent between about 7,700 and 5,500 BP, i.e., about 8,500 to 6,300 years ago, when calibrated. Lake Mega Chad therefore constitutes one of the most convincing cases demonstrating an early Holocene climatic optimum, and an eventual decline to conditions more resembling those of the present. This story is however now dwarfed by dramatic new evidence from Lake Bosumtwi in Southern Ghana (6° 30’ N 1° 25’ W) (Miller, 2013; Miller and Gosling, 2014). This lake occupies a 1.08±0.04 million year old meteorite impact crater 11 km in diameter. The lake itself at present is 8.5 km wide and 76 metres deep. Since the time of impact, 294 metres of sediment have accumulated in the centre of the basin. It was first investigated by Mike Talbot and his colleagues, who obtained radiocarbon dates on exposed lacustrine deposits on the shores of the lake and also on material from three cores which were put down in 1976 (Talbot, 1983; Talbot and Delibrias, 1980; Talbot et al., 1984). The longest of these cores was 17 metres, and a succession of lake level and vegetational changes was plotted, going back for 27,500 years (Talbot, 1983, Fig. 1). Two marked Holocene regressions were observed, but the main regression occurred prior to 12,500 BP, during which time the vegetation in the area of the lake was characterised as wooded grassland of montane character. In 2004, further cores were put down as part of an International Continental Drilling Program, this time extending to the full depth of the deposits, and a much longer record of 10

The Middle Stone Age in West Africa: Introduction

Figure 7. Palynological and stratigraphic record at lake Bosumtwi (after Miller and Gosling, 2014, Fig. 2).

climatic change has been obtained, particularly on the basis of core BOS04-5B. The results so far have been summarised in Charlotte Miller’s thesis, where particular attention is paid to the fossil pollen record (Miller, 2013). 135 new dates (most of them radiocarbon) were obtained for this core, and a chronology for the last 520,000 years has been estimated based on a linear interpolation between the radiocarbon dates and an Ar-Ar date on impact glass at the base. The sequence thus established extends through 150 metres of sediment (Miller and Gosling, 2014, Fig. 2) (shown here as Figure 7). With 216 pollen taxa identified to at least the family level, this is said to be the ‘longest terrestrial pollen record’ so far obtained in Africa. According to Miller, grass pollen was the dominant component throughout, but there were six prolonged periods of forest expansion, which on the basis of comparison to marine cores are equated with interglacial episodes. In the diagram, the Bosumtwi forest zones (BF) are shown as dark and pale grey horizontal bars. Bosumtwi forest zone 1 (BF1) is equated with the Holocene and BF2 (at a depth of 35.85 to 28 metres) with the last interglacial period (oxygen isotope stage 5e). There are two optically stimulated luminescence (OSL) dates from core 5B for this episode of 11

The Middle Stone Age of Nigeria in its West African Context

112,415±8010 and 78,620±5590 BP. Not that everything is completely straightforward. During this episode, as expected in terms of the model, grass pollen drops from 61 to 5.6% and there is a corresponding rise in woodland taxa. Nonetheless, there are also indications that this long wet period was interrupted by a ‘megadrought’ when the lake was completely dry, as indicated by the red bar showing a barren interval (BI) at the appropriate point in the diagram. What is clear in general is that we now have a reliable standard of comparison covering much of the Quaternary in West Africa, which is as good as any elsewhere on the continent. What is still missing is a direct link between this sequence and the archaeological record. Archaeological perspectives This work concerns both the Middle Stone Age (MSA) and the Sangoan, the latter because so far as West Africa is concerned, its occurrences, when stratigraphically verified, bear a close relationship to the MSA, and in one way or another it has long been assumed (rightly or wrongly) that it was somehow intermediate between the Acheulean and the Middle Stone Age. Both expressions require definition. The term MSA was originally proposed by Goodwin on the basis of the archaeological materials then known in South Africa (Goodwin, 1928). Typologically, it was said to be characterised by scrapers and points, but the two essential elements in Goodwin’s definition were technological: the frequent presence of flakes with facetted butts, and ‘a tendency to convergent rather than parallel flaking’ on the dorsal surface of the flakes. From the first, there were elements of comparison between the MSA and the Middle Palaeolithic of Europe, the Near East, and North Africa, since, following Burkitt’s suggestion, Goodwin accepted that it showed ‘strong Mousterian influence’. As mentioned already, it was once assumed that the MSA was contemporary with the Upper Palaeolithic in Europe, but, irrespective of that, I long ago came to the conclusion – on the basis of the material from Nigeria and Cameroun – that ‘at a techno-complex level, making allowance for lesser taxonomic differences, Middle Palaeolithic and MSA are really identical’ (Allsworth-Jones, 1986: 166). That is not to say that the MSA is an undifferentiated whole. In relation to East Africa, J.D. Clark pointed out some time ago (1988) that the ‘beginnings of regional identity’ could be discerned at this time. He referred to ‘a number of spatially adapted variants or facies’ which in his view were closely related to different environments and habitats, as well as to such factors as raw material and available resources. It is likely that the same was true in West Africa. The methodology here adopted for the study of these industries is based essentially on the work of François Bordes (1961, 1972: 48-54, 152-157). He prepared a list of 63 types, of which the first three were unretouched Levallois flakes and points. Generally similar but shorter lists are given here for the various sites studied. They are not exactly uniform, because they depend in some cases on material inventoried by others, and there are a number of idiosyncrasies which need to be taken into account. One particularity is the appearance of artefacts which I have referred to as limaces, which are especially characteristic of Zenabi, Tibchi, and Yelwa, in northern Nigeria. By limace (literally ‘slug’) is meant an ‘ellipsoidal unifacial tool’ (Bordes, 1972), objects described in more detail as ‘racloirs convergents doubles’ typically ‘épaisses et symétriques, parfois trapues et larges’ (Bordes, 1961: 23). As emphasized in the account of the Nigerian sites, the artefacts given this name are on the boundary between cores and steeply retouched tools, and probably they extend the meaning of the term beyond what was intended by Bordes, but they are unquestionably distinctive and deserve a separate appellation of some kind. To express the percentage contribution of unretouched Levallois flakes and points to any particular assemblage, Bordes suggested the use of a Levallois typological index (ILty), as distinct from a Levallois technological index (IL) which includes all Levallois blanks, whether retouched or not. Other indices can be calculated, such as (technically) the percentage contribution of facetted and dihedral platforms to the totality of all blank platforms (IFl) or (typologically) the percentage of sidescrapers (IR) in any given assemblage. These indices have been used on occasion in this work, and they can be useful in distinguishing one kind of industry from another. Bordes also suggested a number of categories of cores, among them Levallois, disc, and prismatic. The Levallois cores were actually divided into three types: flake, blade, and point (Bordes, 1961: 71-72). A Levallois flake was defined in a simple and sweeping fashion, as follows (Bordes, 1961: 14): ‘éclat à forme prédéterminée par une préparation spéciale 12

The Middle Stone Age in West Africa: Introduction

du nucléus avant enlèvement de cet éclat’. For Levallois points and blades, Bordes maintained that the same principle applied, but he admitted that the preparation of the core was somewhat distinct. In the first case, its preparation permitted the production of ‘un éclat triangulaire, obtenu d’un seul coup, sans retouches’. In the second case, a series of ‘longs éclats étroits’ were obtained, ‘parallèles au lieu d’être centripètes’. These definitions and distinctions have also been observed in this work. Since Bordes’s time, the Levallois concept has been further refined, notably by Eric Boëda (1990, 1994). He has emphasized the variability inherent in the concept, and in particular he has distinguished two main types, which he has termed preferential and recurrent (récurrent: ‘série, dont chaque terme est une fonction des termes immédiatement précédents’). The two types, as defined by Boëda, are illustrated at Figure 8. a and b show the preferential process, a being the first removal, b the second. As Boëda points out, the second removal cannot take place without a reworking of the removal surface. c and d show the recurrent procedure, where the core is struck from one platform. Again there are two stages, the second of which Figure 8. Preferential and recurrent Levallois techniques (after Boëda, 1994, Fig. 176). requires reworking as before, but at each stage several Levallois blanks can be produced. In both cases, it is the surface which is exploited, hence the procedure is different from that characteristic of the Upper Palaeolithic - which Boëda terms volumetric - where the creation of a crested guide flake (lame à crête) is the starting point of the process (Boëda, 1990, Fig. 4). Specifically in the African context, further varieties of the Levallois technique have been suggested, in particular (following a proposal originally put forward by J. and G. Guichard) concerning so-called Nubian cores (Vermeersch and Van Peer, 1988; Van Peer, 1991). These cores are illustrated at Figure 9, 1 and 2. The Nubian I method [shown at 2] enables the production of points or pointed flakes by creating a distal ridge, from a platform at the opposite end to the one which is intended for the removal of the Levallois blank. The Nubian II method [shown at 1] also achieves a central ridge, in this case by means of a series of transversal scars. In both cases the cores are approximately triangular. Figure 9, 3 and 4, shows (by contrast) a ‘classic’ Levallois core, and a so-called Halfan core. As originally defined by A.E. Marks, this type of core was seen as a means for the production of secondary Levallois flakes, i.e., flakes exhibiting a large negative of the first Levallois removal on their dorsal surface, although in Van Peers’ view it is no more than a variant of the ‘classic’ method. While as the name implies, these methods are important in North-east Africa, they were so far as I am aware not employed in West Africa. 13

The Middle Stone Age of Nigeria in its West African Context

Figure 9. Nubian cores (1 and 2), ‘classic’ and Halfan cores (3 and 4) (after Van Peer, 1991, Fig. 3).

While the definition of the MSA might be relatively straightforward, that of the Sangoan is not. Some of the relevant points were considered before (Allsworth-Jones, 1987), so they will here be summarized as succinctly as possible. The name Sangoan comes from Sango Bay, on the west side of Lake Victoria in East Africa, where significant collections were first made on the hills above the Bay by E.J. Wayland in 1920 (Cole, 1967). Subsequently the term has been much more widely used, and it is fair to say that it has become a ‘very loosely defined entity’ (McBrearty, 1987). At least in the period when the fieldwork described in this volume was carried out, by this author and others, the main standards of reference were supplied by the work done by J.D. Clark in north-eastern Angola and at Kalambo Falls (Clark 1963, 1964, 1969, 1974). All the finds and sections in Angola were located as a result of diamond mining operations carried out along the river valleys of Lunda province, in various horizons of what were referred to as Kalahari Redistributed Sands. The successive archaeological entities recognized were labeled Sangoan/Lower Lupemban, Upper Lupemban, Lupembo-Tshitolian, and Tshitolian, all of which names recur in West African writings on the subject, particularly those of Oliver Davies. There were few individual collections of any size, those from Musolexi, Camafufo, and Catongula being particularly important. Two radiocarbon dates for the Sangoan/ Lower Lupemban from Mufo were published, respectively 38,000±2500 BP (UCLA-168) and >34,000 BP (UCLA169), and one for the Upper Lupemban of 14,503±560 BP (C-581) from the same locality. Clark summed up the post-Acheulean cultural development that took place in north-eastern Angola as showing ‘continuing and general similarity of form which indicates that it represents a gradual autochthonous evolution’ with a number of common factors binding all these cultures together (Clark 1963: 184). At Kalambo Falls the situation was better in that one had a single site and a clear stratigraphy (Clark, 1969). A number of radiocarbon dates were obtained, in the range from 46,100±3500 BP for the Sangoan Industrial Complex to 27,500±2300 BP for the Lupemban Industrial Complex (Clark, 1974, Table 10). The methodology adopted for the study of the artefacts attributed to the Sangoan was quite different to that employed for the MSA. The broad classes were as defined by M.R. Kleindienst (1962), namely, large cutting tools (LCT), heavy duty tools (HD), and light duty tools (LD), and most of the individual artefact classes were also as defined by her, apart from those specific to the Sangoan (Clark and Kleindienst, 1974). Thus, Kleindienst (1962) originally defined picks in a rather general way as ‘sturdy tools with a minimum of overall trimming, but with emphasis upon a point as such’, but this definition was later made more restrictive. Clark made a clear distinction between handaxes, picks, and core-scrapers, 14

The Middle Stone Age in West Africa: Introduction

although he admitted that in the past it had proved very difficult to describe core-scrapers adequately (Clark, 1963: 50). They were described by Kleindienst (1962) as ‘high-backed tools characterized by steep trimming from a flat surface along some segment of the circumference’. Core-axes also posed a problem. According to Clark, the Sangoan core-axe was ‘usually carefully worked, either bifacially or unifacially, only at one end’, but ‘the most evolved forms are elongated with roughly parallel sides and sometimes may have both ends retouched’ (Clark, 1964: 317, plate 11). Lanceolates and other pointed forms, which ‘always show a thin section and careful secondary retouch over one or both faces’ (Clark, 1963: 51), were generally included among the light duty tools, although this was not invariably the case. The ambiguities and implied very broad range of material which could be included in some of these artefact classes have given rise to endless difficulties for the archaeologists working on materials which they have classified as Sangoan in West Africa. This particularly applies to Asokrochona in Ghana and to Anyama in the Ivory Coast, as discussed later in this work, but also to the less well known occurrences in Nigeria and Cameroun. The difficulties also reflect a tendency earlier remarked upon by Clark (1963: 190) ‘for some prehistorians to describe any crude post-Acheulean industry of early Later Pleistocene times as Sangoan on the basis that it contains only crude and heavy elements … ignoring the fact that the assemblage is often very incomplete’. Institutional Framework West Africa, so far as official languages go, was and is divided into Francophone and Anglophone countries, reflecting their colonial histories, and this division affects the way archaeology was and is conducted as well, not least in respect of the terminology employed (de Barros, 1990; Kense, 1990). French West Africa was quite centralized, and much of the research work carried out, before and after independence, was based upon the Institut Fondamental d’Afrique Noire (IFAN) in Dakar, originally founded in 1938. There were branches in other West African cities, and later universities were also created, in Dakar in 1957 and in Abidjan in 1964, but IFAN has retained a pre-eminent position. Much research work has also been carried out, in various disciplines including archaeology, thanks to French metropolitan institutions, notably the Office de la Recherche Scientifique et Technique Outre-mer (ORSTOM) created in 1943 (and now renamed the Institut de Recherche pour le Développement, or IRD). Both IFAN and ORSTOM sponsored a wide range of publications, as well as specialists in the field, such as Alain Marliac, who worked in Cameroun for many years, exclusively for ORSTOM. There was no parallel to this centralized system, or to research sponsored from the metropolis, in Anglophone West Africa, particularly Ghana and Nigeria. In both countries, independently, museums and government archaeological services were founded, but much depended on the initiative of colonial officers who persuaded their superiors of the merits of what they advocated. The University College of the Gold Coast (now the University of Ghana) in Legon was founded in 1948, and the National Museum of Ghana in 1957, at the time of independence. Thurstan Shaw actually worked, and conducted excavations, in Ghana prior to that, while based at the Achimota College. The Nigerian Antiquities Service was created in 1943, and the Jos Museum was opened in 1952 (Shaw, 1969; Njoku, 1978). Bernard Fagg, who investigated Nok and Zenabi among other sites, served as Director of Antiquities from 1957 to 1966. The University of Ibadan was founded (like Legon) in 1948, but the Archaeology Department did not come into existence until 1970, with Thurstan Shaw as its Head. The author’s own fieldwork was for the most part conducted while he was a member of staff of the University of Ibadan, with financial and logistical assistance from the Department, as well as from the Federal Department of Antiquities (now the National Commission for Museums and Monuments). It could not have been done otherwise. In describing and analyzing the MSA in Nigeria and the other countries of West Africa, for the purposes of this volume, it was decided that, after an examination of the material from Nigeria itself, the other relevant sites should be examined country by country. This is not a division which has much to do with prehistory, but it has a lot to do with present day reality, including the Francophone-Anglophone division referred to above, and the way in which archaeology is conducted in those countries. A general summary will then highlight the overall picture emerging, including some recent developments which have modified quite considerably the pattern previously thought to exist.

15

The Middle Stone Age of Nigeria in its West African Context

References Geography Harrison Church, R.J. 1963. West Africa. London: Longmans, Green and Co. Rochebrune, R. de, Sablayrolles, J. (eds) 2000. L’Atlas de L’Afrique. Paris: Les Éditions du Jaguar. Geology and Geomorphology Furon, R. 1963. Geology of Africa. (trans. A. Hallam, L.A. Stevens). Edinburgh and London: Oliver and Boyd. Furon, R., Daumain, G., Dubertret, L., Lelubre, M. 1958. Esquisse structurale provisoire de l’Afrique. Congrès Géologique International, Association des Services Géologiques Africains, Paris. Schlüter, T. 2006. Geological Atlas of Africa. Berlin, Heidelberg, New York: Springer-Verlag. Thomas, M.F. 1974. Tropical Geomorphology. London: Macmillan. Tricart, J. 1972. The Landforms of the Humid Tropics, Forests and Savannas. (trans. C.J.K. de Jonge). London: Longman. Vegetation and Climate Aubréville, A., Duvigneaud, P., Hoyle, A.C., Keay, R.W.J., Mendonça, F.A., Pichi-Sermolli, R.E.G. 1958. Vegetation map of Africa south of the Tropic of Cancer. London: Oxford University Press. Dalziel, J.M. 1937. The Useful Plants of West Tropical Africa. London: The Crown Agents for the Colonies. Hopkins, B. 1974. Forest and Savanna. Ibadan and London: Heinemann. Keay, R.W.J. 1965. An Outline of Nigerian Vegetation. (3rd ed.). Ibadan: Federal Department of Forestry Research. Keay, R.W.J., Onochie, C.F.A., Stanfield, D.P. 1964. Nigerian Trees. Vols. I and II. Ibadan: Federal Department of Forestry Research. Lawson, G.W. 1966. Plant Life in West Africa. London: Oxford University Press. Thompson, B.W. 1975. Africa: The Climatic Background. Ibadan: Oxford University Press. White, F. 1986. La Végétation de l’Afrique. Paris: ORSTOM-UNESCO. Climatic and Environmental History Allsworth-Jones, P., Harvati, K., Stringer, C. 2010. The archaeological context of the Iwo Eleru cranium from Nigeria and preliminary results of new morphometric studies. In West African Archaeology: New developments, New perspectives: 29-42. Oxford: British Archaeological Reports International Series, 2164. Allsworth-Jones, P. 2016. Changing frameworks for the Quaternary in West Africa. In ed. P.A. Oyelaran, R.A. Alabi and P.A. Adeonipekun, Human Palaeoecology in Africa: Essays in honour of M. Adebisi Sowunmi: 214-237. Ibadan: The Palynological Association of Nigeria, Ibadan University Printery. Barendsen, G.W., Deevey, E.S., Gralenski, L.J. 1957. Yale natural radiocarbon measurements III. Science, 126: 908-919. Bishop, W.W. 1967. Annotated Lexicon of Quaternary Stratigraphical Nomenclature in East Africa. In W.W. Bishop, J.D. Clark (eds). Background to Evolution in Africa: 375-395. Chicago and London: The University of Chicago Press. Burke, K., Durotoye, A.B., Whiteman, A.J. 1971. A Dry Phase South of the Sahara 20,000 Years Ago. West African Journal of Archaeology, 1: 1-8. Burke, K.C., Durotoye, A.B. 1972. The Quaternary in Nigeria. in ed. T.F.J. Dessauvagie and A.J. Whiteman, African Geology, Ibadan 1970: 325-347. Ibadan: Department of Geology, University of Ibadan. Butzer, K.W. 1971. Environment and Archeology. Chicago: Aldene Publishing Company. Butzer, K.W. 1978. Climate patterns in an un-glaciated continent. The Geographical Magazine, LI.3: 201-208. Clark, J.D. 1959. The Prehistory of Southern Africa. Harmondsworth: Penguin Books. Clark, J.D. 1970. The Prehistory of Africa. London: Thames and Hudson. Clark, J.D. 1975. Africa in prehistory: peripheral or paramount? Man, 10 (N.S.): 175-198.

16

The Middle Stone Age in West Africa: Introduction

deMenocal, P., Ortiz, J., Guilderson, T., Adkins, J., Sarnthein, M., Baker, L., Yarusinsky, M. 2000. Abrupt onset and termination of the African Humid Period: rapid climate responses to gradual insolation forcing. Quaternary Science Reviews, 19: 347-361. Dupont, L.M., Weinelt, M. 1996. Vegetation history of the savanna corridor between the Guinean and the Congolian rain forest during the last 150,000 years. Vegetation History and Archaeobotany, 5: 273292. Dupont, L.M., Jahns, S., Marret, F., Ning, S. 2000. Vegetation change in equatorial West Africa: time-slices for the last 150 ka. Palaeogeography, Palaeoclimatology, Palaeoecology, 155: 95-122. Imbie, J. and K.P. 1979. Ice Ages. Cambridge and London: Harvard University Press. Leblanc, M., Favreau, G., Maley, J., Nazoumou, Y., Leduc, C., Stagnitti, F., van Oevelen, P.J., Delclaux, F., Lemoalle. J. 2006. Reconstruction of Megalake Chad using Shuttle Radar Topographic Mission data. 2006. Palaeogeography, Palaeoclimatology, Palaeoecology, 239: 16-27. Lézine, A-M., Cazet. J-P. 2005. High-resolution pollen record from core KW31, Gulf of Guinea, documents the history of the lowland forests of West Equatorial Africa since 40,000 yr ago. Quaternary Research, 64: 432-443. Lowe, J.J., Walker, M.J.C. 1997. Reconstructing Quaternary Environments. Harlow: Addison Wesley Longman. Miller, C.S. 2013. 520,000 years of environmental change in West Africa. PhD dissertation, The Open University, UK. Miller, C.S., Gosling, W.D. 2014. Quaternary forest associations in lowland tropical West Africa. Quaternary Science Reviews, 84: 7-25. Oakley, K.P. 1964. Frameworks for Dating Fossil Man. London: Weidenfeld and Nicolson. Shackleton, N.J., Opdyke, N.D. 1973. Oxygen isotope and paleomagnetic stratigraphy of equatorial Pacific core V28-238: oxygen isotope temperatures and ice volumes on a 105 and 106 year scale. Quaternary Research, 3: 39-55. Shaw, T., Daniels, S.G.H. 1984. Excavations at Iwo Eleru, Ondo State, Nigeria. West African Journal of Archaeology, volume 14. Smart, P.L., Frances, P.D. 1991. Quaternary Dating Methods – A User’s Guide. Technical Guide No. 4. Cambridge: Quaternary Research Association. Talbot, M.R. 1983. Lake Bosumtwi, Ghana. Nyame Akuma, 23: 11-12. Talbot, M.R., Delibrias, G. 1980. A new late Pleistocene-Holocene water-level curve for Lake Bosumtwi, Ghana. Earth and Planetary Science Letters, 47: 336-344. Talbot, M.R., Livingstone, D.A., Palmer, P.G., Maley, J., Melack, J.M., Delibrias, G., Gullik-Sen, S. 1984. Preliminary results from sediment cores from Lake Bosumtwi, Ghana. In ed. J.A. Coetzee, E.M. van Zinderen Bakker, Palaeoecology of Africa, 16: 173-192. Rotterdam: A.A. Balkema. Archaeological perspectives Allsworth-Jones, P. 1986. Middle Stone Age and Middle Palaeolithic: the evidence from Nigeria and Cameroun. in ed. G.N. Bailey and P. Callow, Stone Age Prehistory: 153-168. Cambridge: Cambridge University Press. Allsworth-Jones, P. 1987. The Earliest Human Settlement in West Africa and the Sahara. West African Journal of Archaeology, 17: 87-129. [reprinted in WAJA vol. 36 (2006): 126-177]. Bishop, W.W., Clark, J.D. (eds) 1967. Background to Evolution in Africa. Discussions on Terminology, 861875. Chicago and London: The University of Chicago Press. Boëda, E. 1990. De la surface au volume: analyse des conceptions des débitages Levallois et laminaire. In ed. C. Farizy, Paléolithique moyen recent et Paléolithique supérieur ancien en Europe. Colloque international de Nemours, 9-11 mai 1988. Mémoires du Musée de Préhistoire d’Ile de France, 3: 63-68. Boëda, E. 1994. Le concept Levallois: variabilité des méthodes. Paris: CNRS. Bordes, F. 1961. Typologie du Paléolithique ancien et moyen. Bordeaux: Imprimeries Delmas. Bordes, F. 1972. A Tale of Two Caves. New York: Harper and Row. Clark, J.D. 1963. Prehistoric cultures of northeast Angola and their significance in tropical Africa. Parts I and II. Museu do Dundo: Diamang, Publicações Culturais 62.

17

The Middle Stone Age of Nigeria in its West African Context

Clark, J.D. 1964. The Sangoan Culture of Equatoria: the implications of its stone equipment. In ed. E. Ripoll Perelló, Miscelanea en Homenaje al Abate Henri Breuil, vol. I, Barcelona: 309-325. Clark, J.D. 1969. Kalambo Falls Prehistoric Site, volume I. Cambridge: Cambridge University Press. Clark, J.D. 1974. Kalambo Falls Prehistoric Site, volume II. Cambridge: Cambridge University Press. Clark, J.D. 1988. The Middle Stone Age of East Africa and the Beginnings of Regional Identity. Journal of World Prehistory, 2 (3): 235-305. Clark, J.D., Kleindienst, M.R. 1974. The Stone Age cultural sequence: terminology, typology and raw material. In J.D. Clark (ed.), Kalambo Falls Prehistoric Site II, The late prehistoric cultures: 71-106. Cambridge: Cambridge University Press. Cole, G.H. 1967. The Later Acheulian and Sangoan of Southern Uganda. In W. W. Bishop, J.D. Clark (eds) 1967. Background to Evolution in Africa: 481-528. Chicago and London: The University of Chicago Press. Goodwin, A.J.H. 1928. An Introduction to the Middle Stone Age in South Africa. South African Journal of Science, 25: 410-418. Kleindienst, M.R. 1962. Components of the East African Acheulian assemblage: an analytic approach. In Actes du IV Congrès Panafricain de Préhistoire et de l’Étude du Quaternaire, section III: 81-105. McBrearty, S. 1987. Une évaluation du Sangoen: son âge, son environnement et son rapport avec l’origine de l’Homo sapiens. L’Anthropologie, 91: 497-510. Van Peer, P. 1991. Interassemblage Variability and Levallois Styles: The Case of the Northern African Middle Palaeolithic. Journal of Anthropological Archaeology, 10: 107-151. Vermeersch, P.M., Van Peer, P. 1988. The Early Upper Paleolithic in Egypt: 1-22. In ed. J.F. Hoffecker, C.A. Wolf. The Early Upper Paleolithic: Evidence from Europe and the Near East. Oxford: British Archaeological Reports International Series 437. Institutional Framework Barros, P. de 1990. Changing paradigms, goals and methods in the archaeology of francophone West Africa. In ed. P. Robertshaw, A History of African Archaeology: 155-172. London: James Currey, and Portsmouth: Heinemann Educational Books. Kense, F.J. 1990. Archaeology in anglophone West Africa. In ed. P. Robertshaw, A History of African Archaeology: 135-154. London: James Currey, and Portsmouth: Heinemann Educational Books. Njoku, O.O. 1978. 25 Years of Jos Museum. Jos: Federal Department of Antiquities, National Museum. Shaw, T. 1969. Archaeology in Nigeria. Antiquity, XLIII: 187-199.

18

Chapter 2

The Middle Stone Age of Nigeria The author’s own work in Nigeria took place for the most part in a study area north of the Jos Plateau, as shown in Figure 10. A detailed account will be given of the sites investigated in this area: Zenabi, Mai Lumba, Tibchi, Yelwa, Banke, and the Ningi Hills (the latter accounting for the north-eastern extension of the study area as shown). In addition, work done at other sites will be summarised, in particular Nok, Pingell, Rop, and Saminaka. A number of sites described as Sangoan have been identified in other parts of northern Nigeria, and these will also be considered. The situation in southern Nigeria is somewhat different, in terms of its research history and the nature of the material recovered, for the most part in the vicinity of Ibadan, and at Asejire. For the purposes of this account therefore northern and southern Nigeria will be treated separately.

Figure 10. The study area north of the Jos Plateau.

19

The Middle Stone Age of Nigeria in its West African Context

Northern Nigeria In terms of geographical situation, raw materials, and mode of occurrence, the archaeological sites in the vicinity of the Jos Plateau have a close relationship to what is referred to by geologists as the Younger Granite Province of Nigeria. The general characteristics of this province are well known and have been described by a number of authors (Turner, 1972, 1973, 1976; Bennett et al., 1984; Ngako et al., 2006). These works build upon detailed studies of individual formations, which go back for many years, e.g. Falconer and Raeburn (1923), Bain (1934), MacLeod et al. (1971), Thorp (1975), Jacobson and MacLeod (1977), Turner and Bowden (1979). The reason for such an interest is mainly due to the fact that the formations in the province have been a significant source of economically valuable minerals, notably cassiterite, the exploitation of which has had a considerable effect on Nigeria for more than a hundred years (Freund, 1981). An understanding of these factors, aside from their intrinsic interest, is relevant when seeking to understand the archaeological sites and how they came to be what they are. The Nigerian Younger Granite province extends over an area of some 400 x 150 km centred on the Jos Plateau (Figure 11). It consists of up to 50 named ring complexes, which have been intruded into the basement. Ring complexes of a similar type extend into Niger, where the Aïr and Damagaram-Mounio formations cover an even larger area (Figure 12). Using the Rb-Sr (rubidium-strontium) technique, it has been ascertained that the age of these formations in Nigeria decreases in a north-south direction, from Daura in the north at 313±13 myrs to Mada in the south at 150±6 myrs ago (Bennett et al., 1984, Appendix). By contrast, the 28 ring complexes in the Aïr mountains date to c. 480-400 myrs, and those in the Damagaram-Mounio area to c. 330-260 myrs ago (Ngako et al., 2006). Up to 60 analogous formations in Cameroun, extending along the so-called Cameroun Line, are much later, all ≤ 66 myrs old. As MacLeod et al. (1971, 55) remark, the Nigerian ring complexes possess an ‘essential unity’ characterised by a gradation from glassy rhyolites, through porphyries, to coarse-grained granites - hence it is possible to provide a single general account of their mode of occurrence. As described by Turner, there were three main stages in the development of these complexes. (1) The early volcanic stage. High volcanic mountains were intruded into domed or arched basement rocks. Early and late rhyolites from this intrusion have been recognised, the former often deposited as ash flows rather than lavas, the latter more crystal-rich. High-level magma chambers developed beneath the mountains, but over time they were much eroded, and today the volcanic rocks survive only within the confines of the ring structures, which developed next. (2) The caldera and ring-dyke stage. The underlying magma chambers were partially emptied thanks to the eruption of pyroclastic flows, along generally concentric fault lines, which led to the collapse of the central block and the formation of a caldera in each case. Marginal granite-porphyry ring-dykes formed along the fault lines, and the interior of the space was partially invaded by intrusions associated with them, such as sills and laccoliths. The resistant ring-dykes today are the most obviously visible features of the Younger Granite Province, and give it its characteristic appearance, often with sets of overlapping and successive ring-dykes. (3) The waning igneous stage. The space within the ring-dykes was filled by further intrusions, mainly biotite and riebeckite granites, the characteristics of which reflected the gradual cooling of the underlying magma. Today the biotite granites account for 45% of the total area of the Younger Granite Province, but it will be observed from the map at Figure 11 that they are more widespread in the south, including the Jos-Bukuru area, than they are in the north. This is attributed to the fact that the northern complexes have been less deeply eroded than those in the south. From the point of view of economic geology, this stage is the most important, since biotite granite in particular was affected by mineralisation. Cassiterite (SnO2) was precipitated in greisen veins occupying joints within this rock, whereas columbite (Fe2+Nb2O6) was disseminated throughout its groundmass. Both were subject to erosion.

20

The Middle Stone Age of Nigeria

Figure 11. The Younger Granite Ring Complexes of Nigeria (after Turner 1976, Fig. 1).

21

The Middle Stone Age of Nigeria in its West African Context

Three of the ring complexes – Liruei, Banke, and Tibchi – are of particular interest to us, because the principal archaeological sites are directly linked to them. More detailed maps of the area north of the Jos Plateau, in terms of physical geography and geology, but excluding the Ningi Hills, are at Figures 13 and 14, the geological map being particularly relevant. Liruei is described by Turner (1972, Fig. 6) as a ‘typical ring complex’, the sequence of events in the creation of which follows the usual order, from early rhyolites, through a granite porphyry ring-dyke stage, to biotite and riebeckite granites as the final phase. The archaeological sites of Yelwa and Zenabi are right up against the north western and northern sides of the granite porphyry ring-dyke where the Mallam (or Baba) and Gaiya rivers descend to the plain beneath. The two highest hills in the south and east, Dutsen Shetu and Dutsen Ginshi, are rare identifiable remnants of volcanic plugs. The biotite granites are much eroded and form a basin in the centre of the complex. Here Thorp (1975) identified two successive depositional sequences in the valley alluvials, each commencing with a cassiterite

Figure 12. The Younger Granite Ring Complexes of Niger (after Ngako et al., 2006, Fig. 3).

22

The Middle Stone Age of Nigeria

Figure 13. The area north of the Jos Plateau: Physical Geography.

bearing coarse gravel, which he correlated with the sequence revealed at Zenabi. In his view, erosion will have occurred in the late Pleistocene and the Holocene in unstable ‘morphodynamic’ phases, characterised by aridity, vegetation reduction, and increased surface run-off, in a process which he compared with that identified by Burke and Durotoye (1971) in the Ibadan area. He did not suggest any more specific date for these occurrences. The Banke complex map prepared by Jacobson and MacLeod (1977) has been somewhat revised by Bennett et al. (1984, Fig. 2) but the essentials are not affected. Again the sequence proceeds from precaldera tuffs at Dutsen Burtu and Dutsen Zanzari via a granite porphyry ring-dyke stage to biotite granite in the centre and south-west of the complex. It will be observed that the granite porphyry ring-dyke is discontinuous, although the existence of a full ring-shaped fault can be inferred. The Damau hills to the north, composed of ignimbrites, constitute an extensive dissected plateau, and are today completely deserted. Up to one third of the area enclosed within the ring-fault consists of basement rocks, a flat area to the south and east, where tin mining has been taking place along the

23

The Middle Stone Age of Nigeria in its West African Context

Figure 14. The area north of the Jos Plateau: Geology.

course of two rivers. In this case, unusually, greisen veins occur not only in the biotite granite but also at various places in the basement. The Tibchi complex was studied in detail by E.C. Ike (1979) for his doctoral thesis, and we profited from meeting him and consulting him in the field. In this case basement rocks account for two thirds of the area enclosed within the granite porphyry ring-dyke, and it is here that the sites of Mai Lumba and Tibchi are situated. There is an extensive central hilly area of biotite granite, containing multiple greisens, including the Kogo lode, which is up to 14 metres wide. Volcanic rocks to the north-west, consisting mainly of porphyritic rhyolite lavas and poorly welded tuffs, commenced the sequence, and they no doubt provided the bulk of the raw material used at the two sites. If geology and geomorphology are sufficient to account for the importance of ring-complexes in the story of the archaeological sites in Northern Nigeria, the matter of their discovery depends upon their economic exploitation, above all alluvial mining for the recovery of cassiterite. Without that, they would have remained unknown. As emphasised by Freund (1981), tin mining and smelting was practised in pre-colonial times, with smelting operations centred on Ririwain Kano (in the centre of the Liruei Hills)

24

The Middle Stone Age of Nigeria

and Ririwain Dalma (north-west of Jos). Smelted tin ‘straws’ were traded as far as Zinder, but this remained essentially a small scale regional enterprise. With the imposition of colonial rule, the situation changed. The local smelters were closed down, and large scale exploitation by European firms commenced, with the issuing of exclusive prospecting licences (EPLs) and mining leases, the latter usually granted for renewable periods of 21 years. Throughout the colonial period, no tin ore was smelted in Nigeria, it was all shipped out as cassiterite, for the most part to Liverpool, until 1962. This process was much enhanced following the completion of railway lines to Bukuru in 1915 and to Jos in 1927. A number of different firms were involved, but the largest among them became ATMN (Amalgamated Tin Mines of Nigeria) which by 1939 controlled 26% of the leased mining land and 47% of tin ore production as a whole. This work was concentrated above all on the Jos Plateau, the area with the greatest abundance of biotite granite, on what became known as the Plateau Minesfield.

Figure 15. Alluvial tin-mining at Nok (after Fagg 1977, Frontispiece).

ATMN introduced some elements of mechanisation in the 1930s, in the form of dredges, pumps, and draglines. Nonetheless, hand labour as a method still accounted for up to half of the total (in terms of yards moved) in the years after the second world war, and during that period the number of workers taken on annually seldom fell below 50,000 (Freund, 1981, tables 8.5 and 8.6). In general, as Freund puts it, the tin mines of the Jos Plateau employed ‘the largest labour force of any enterprise in colonial West Africa’. Most of these labourers were migrant workers, a so-called ‘floating population’, but their headmen were much more permanent. These men might control gangs of up to 40 persons, and were an indispensable part of the whole operation. Also important were ‘tributers’, a term which has its origin in the history of Cornish tin mining (Freund, 1981: 106, note 108). These men were essentially piece-workers, labouring on their own account, and selling the recovered tin-ore to the licensee of the property (Fagg, 1977: 15, note 6). A paddock operated by tributers is shown on Bond’s map of Zenabi, which he prepared in 1948. The manner in which hand labourers worked at that time (with pick and shovel, and head-pan) is clear from the photograph taken by Fagg at Nok (Figure 15). Fagg stated that so far as Nok is concerned it was the tributers who made ‘most of the significant archaeological discoveries … often in the remotest places’. Large scale use of draglines would clearly be inimical to this process. From our point of view, it is fortunate that in the 1970s in the more remote areas, such as those studied here, the use of hand labour was paramount, and the company concerned, Bisichi-Jantar, was welcoming and encouraged our work. That too is part of the story. A consideration site by site follows.

25

The Middle Stone Age of Nigeria in its West African Context

Zenabi Introduction Middle Stone Age artefacts were first observed in the outwash gravels of the Gaiya river below the Zenabi Falls in northern Nigeria (10° 47’ N, 8° 46’ E) by Bernard Fagg in 1947 (Fagg, 1956). Then as now the artefacts were being turned up as a result of tin-mining activities. Fagg and Bond returned to the site in 1948, and Bond was able to sketch its geological history (Bond, 1956). Fagg collected further artefacts during one or two additional visits. Both Fagg and Bond referred to the site as Zenebi, but a slight correction of orthography seems necessary, in that the form Zenabi now appears on the relevant 1:50,000 and 1:100,000 maps published by the Nigerian Federal Surveys (1965) and this form has been used in other more recent works (Ajakaiye, 1968; Thorp, 1975). For the sake of consistency therefore it is also used here. Following the work of Fagg and Bond, the next investigations were conducted by Robert Soper in two short visits to the site in 1962 and 1964 (Soper, 1965). He was able to make certain stratigraphic observations both at Zenabi and at nearby Yelwa, as well as making an inventory of the artefacts then available in the Jos Museum. In 1968 Angela Fagg (Mrs A. Rackham) visited the area above the Zenabi Falls and collected a number of artefacts along the Gaiya river in the vicinity of Zenabi Tudo. Dr F. Anozie made a study of certain of the finds in the Jos Museum in the early 1970s (Anozie, 1975); but no further fieldwork was done until March 1976, when I observed that large quantities of artefacts were being turned up in the paddock known as Fulata on the east bank of the Gaiya river not far from the foot of the Falls. It was decided to include the site in a programme of investigation into the nature and conditions of occurrence of the Nigerian Middle Stone Age in the area north of Jos, where a number of sites have been affected by tin mining operations in this part of the Nigerian Younger Granite Province. Excavation and site planning was commenced in July 1976 and continued in May 1978, as described in two preliminary publications (Allsworth-Jones, 1981 and 1987) and in a paper presented to the 8th Panafrican Congress in Nairobi in 1977 (Allsworth-Jones, 1980). Site situation and stratigraphy (1) Earlier investigations According to Bond, both the Gaiya and the Baba rivers, falling steeply from the northern front of the Liruei Hills, had created massive alluvial outwash fans, and it was from these that the cultural material had been recovered. The plains consisted of granite and gneiss, both strongly lateritised. The river channel at Zenabi had cut through the laterite crust and the alluvial material rested in this channel. The outwash fans were described as being in three parts superimposed one upon the other, as indicated in his sketch map (Figure 16; it should be noted that this map is oriented with the north side at the base). Bond commented that all the deposits were coarse and bouldery near the foot of the scarp but became finer as they fanned out onto the plains. Three paddocks or mining areas were indicated on the map, but not named in the published version. They had ‘exposed the succession in some detail’ but owing to the nature of the deposits the thicknesses were variable. Bond’s description of the ‘Zenebi sequence’ (1956: 192-193) appears to be a synthesis of all his observations, as follows. (1) Brown gritty sands and gravels, a recent outwash fan, exposed only near the foot of the Falls. It does not contain any unrolled cultural material. (2) Red gritty sands and gravels, with a tin-wash at the base containing MSA artefacts. To all appearances, this is the position of the industry at Fulata, but in Bond’s time it was exposed in the ‘Tributers’ Paddock’. The red series was separated from the deposits both above and below by what was described as a ‘non-sequence’. (3) Buff false-bedded gritty sands, the lowest fan with the most complex stratigraphy, said to be exposed in the Main Paddock. An ‘upper’ tin-wash with MSA artefacts occurred in gravel below the buff-coloured 26

The Middle Stone Age of Nigeria

Figure 16. Geological sketch map of Zenabi (after Bond, 1956, 201).

sands, but was underlain by further deposits below a ‘minor erosion surface’. The deposits here were listed as blue clay (4 feet thick = 1.2 m), lenses of peat and fossil tree trunks, gravel with a ‘lower’ tin-wash (5 feet thick = 1.5 m), and sand (4 feet thick = 1.2 m) above an erosion surface and decomposed granite bedrock. It is these lowest deposits which yielded a radiocarbon date of 5440±100 BP (Y-142-7) (Barendsen et al., 1957), described as having been obtained on ‘wood from a large log imbedded in older tin-bearing alluvium’ collected by Fagg and Bond in 1948. The log in question is in fact the one illustrated by Fagg (Fagg, 1956, Fig. 3; here Figure 17) from ‘Zenabi No. 1 paddock’, shown as occurring below the ‘buffcoloured sediments’ (3) (Bernard Fagg, personal communication). The implication of the phraseology used is that Paddock No. 1 and the Main Paddock are one and the same thing. Bond commented that there must have been a break between (2) and (3), the red and buff-coloured sands, but he regarded it as ‘minor’ because ‘a similar flake culture is found in situ in both deposits’. Both were ascribed to the Gamblian Pluvial, although as the laboratory pointed out, the date obtained rather suggested the Makalian (Barendsen et al., 1957). Bond observed that the majority of the cores and flakes were ‘fresh or only lightly rolled’ and took this as evidence that they were contemporary with their enclosing 27

The Middle Stone Age of Nigeria in its West African Context

deposits. Since there was ‘no evidence of handaxes, even in the lowest gravel’ the whole series was assigned to the MSA. When working in the archives of the Jos museum, Anozie (1975) located three sections from Zenabi, which he assumed originated with Bond, although they may possibly be due to Soper (1965). The sections are labelled as follows: A, nearest the valley; B, Main Paddock; C, north of the road (as shown at Figure 18). Both A and B contain two MSA horizons, the situation in B resembling that described by Figure 17. Carbonised tree trunk in Zenabi No.1 Paddock (after Fagg, 1956, Fig. 6). Bond for the Main Paddock in that there are two gravels with MSA, separated by a layer of grey-blue clay. C, north of the road, contains only one MSA horizon. This locality (not described in detail by Bond, although the position of a paddock at this point is indicated on his map) is otherwise known as Karara. The situation is in agreement with

Figure 18. Zenabi sections A, B, and C (after Anozie, 1975, Fig. 16). [A] 1 topsoil and sand. 2 grey-brown consolidated gravel (MSA). 3 fine sand. 4 grey-brown consolidated gravel (MSA). 5 fine sand. [B] 1 topsoil. 2 loose sand. 3 sticky blue clay. 4 gravel (MSA). 5 grey-blue clay. 6 grey-brown consolidated gravel (MSA), above granite bedrock. [C] 1 reddish clay. 2 tin-bearing gravel (MSA). 3 grey clay.

28

The Middle Stone Age of Nigeria

Soper’s account (personal communication) that in 1962 in Zenabi Main Paddock (B) he could distinguish an upper and a lower gravel, both with artefacts. He suggested that the single gravel exposed at Zenabi north paddock (C) could be equated with the upper gravel in the Main Paddock, although only the lower gravel was observed in this Paddock in 1964. The fact that the sections in the various paddocks should have differed somewhat over time is not surprising, given the nature of the deposits and the manner in which they were exposed. In situ finds so far have been very few, almost all of them due to Soper’s observations in 1962-64, the total of such finds amounting to 140 pieces (6 tools, 4 cores, 87 flake/blades, 43 fragments). The remainder of the finds from the earlier investigations, although they are quite abundant, have only a general provenance. (2) Results of work done in 1976-1978 On my first visit to the village of Zenabi on 23 March 1976, some labourers were observed working in the area between the village and the Falls. One of them accompanied me to the paddock known as Fulata (Figure 19), where the ground was strewn with artefacts, and a retouched stone tool was immediately observed in situ in the exposed side of the paddock (Figures 20 and 21). It is a double sidescraper on a Levallois flake/blade (Figure 31.3). Subsequent work at the site took place between July 8 and 16 1976 and then between May 16 and 26 in 1978, with an additional brief stop on 26 March 1977. The stark contrast between the way the site looks in the dry and in the rainy season can be seen from the landscape in Figures 22 and 23. A good deal of time was expended on preparing a plan of the site. A contoured map showing the mining leases 3741, 9282, and 5174 (based on documentation kindly provided by the Bisichi Jantar company) is at Figure 24. It has the advantage of indicating the original course of the Gaiya river below the Falls. A more detailed plan, based on our own survey work, is at Figure 25.

Figure 19. Zenabi Fulata, 23 March 1976.

29

The Middle Stone Age of Nigeria in its West African Context

Figure 20. Zenabi Fulata, 23 March 1976, in situ find.

Figure 21. Zenabi Fulata, 23 March 1976, in situ find.

Figure 22. Zenabi Falls, March 1976.

The former course of the river is indicated, on its way through a very large and deep flooded paddock, which presumably corresponds to the former Main Paddock. The whole ground has been much disturbed by ditches and embankments. The many heaps of stones, the result of former mining work, revealed 30

The Middle Stone Age of Nigeria

Figure 23. Zenabi Falls, July 1976.

some artefacts, as well as many scorpions. The position of our sections 1, 2, and 3 is indicated, the first two in the paddock named Fulata, the third near the edge of the river closer to the bridge. The drawn sections are shown in Figure 26, with a corresponding description in Table 2. Photographs of the work in progress in 1976 and of sections 1 and 2 are at Figs. 27-29. The upper part of section 1 (as well as section 2) was traced and studied in that year, whereas the sounding beneath the level of the paddock was put down in 1978, when section 3 was also cut. The upper part of section 1 is 5 metres thick, whereas the sounding was extended down for a further 4 metres. As can be seen from the drawn sections, the uppermost deposits consist of a series of red or reddish-yellow sands interspersed with thin gravel bands. Layer 2, beneath topsoil layer 1, contains numerous termite tunnels, and a few potsherds were recovered at the top. Layer 6, yellowish-red gravel, 1-2 metres thick, was identified as the Upper Tin-wash, its presence explaining why the miners took the paddock down to that level and no further. Beneath red sand, in the sounding, there was a layer of yellowish-red mottled clay (9) and then yellowish-red sand and gravel (10) constituting the Lower Tin-wash, again over 1 metre thick. These two basal layers were also found in section 3. It had been hoped to find a sample of carbonised wood at this locality, but this hope was not realised. A former overseer of the Bisichi Jantar company, Mr Daudu, indicated various places where he thought such samples might be found, but they would all be at great depths and would be at risk of flooding from the nearby old waterlogged paddocks. As in Soper’s case, the numbers of artefacts found in situ at Zenabi during our investigations were few, despite the encouraging first encounter on 23 March 1976. During a brief visit to the site on 26 March 1977, it was observed that a certain amount of material had fallen from our section of the previous year and was in a heap at the base. This material included 3 tools, 3 cores, 16 flake/blades, and 3 fragments. They were clearly derived from the exposed Upper Tin-wash, and they have the same characteristics as those recovered from the floor of the paddock the previous year, which were collected in large numbers. 31

The Middle Stone Age of Nigeria in its West African Context

Figure 24. Zenabi contour map showing mining leases 3741, 9282, and 5174.

In addition, a small series of artefacts was located in situ at the top and at the base of the Lower Tin-wash (layer 10) the combined figures for which are as follows: 2 cores or core fragments, 7 flake/blades, and 4 fragments. None of these finds can be considered really diagnostic, but they are enough to confirm that another industry is present in the Lower Tin-wash. It can be seen that section 1 bears some resemblance to section B recorded by Anozie on the basis of the Jos museum records as being from the Main Paddock (Figure 18), although one would not expect an exact parallel. Two tin-washes were also noted above the Zenabi Falls by Mr D.W Hannaford, the Manager of the Gold and Base company at Ririwai, who pointed them out to me on 22 March 1976. They occur in the upper part of the Gaiya river, where it flows through biotite granite, but the deposits contain no admixture of coarse gravels or artefacts. The situation is somewhat different immediately above the Falls, in the vicinity of Zenabi Tudo village, where artefacts were located by Angela Fagg in 1968 and later. This entire area has been extensively disturbed by local tin-miners, as shown in Figure 30, and it was concluded that no useful excavation could be carried out here. 32

The Middle Stone Age of Nigeria

Figure 25. Zenabi plan as surveyed in 1976-1978 with position of sections 1, 2, and 3.

33

The Middle Stone Age of Nigeria in its West African Context

Figure 26. Zenabi sections 1-3, 1976 and 1978. Table 2. Zenabi Section 1 stratigraphy 1976-1978. Layer No.

Description

Munsell colour

1

Dark brown topsoil

7.5 YR 3/2

2

Light red sand

2.5 YR 6/8

3

Reddish yellow sand

5 YR 7/8

4

Pink sands

7.5 YR 7/4

4a

Reddish yellow fine gritty sand

7.5 YR 6/6

5 i-iv

Yellowish red gravel bands

5 YR 5/8

6

Yellowish red gravel Upper Tin-wash

5 YR 5/6

7 i-ii

Yellowish red gravel bands

5 YR 5/8

8

Red sand

2.5 YR 5/8

9

Yellowish red mottled clay

5 YR 5/8

10

Yellowish red sand and gravel Lower Tin-wash

5 YR 5/6

Layers 1-6 in section 2 correspond to the above. Section 3 equates only to the lower part of the sequence, where yellowish red layer II (5 YR 5/6) corresponds to 9 and red layer III (2.5 YR 5/8) corresponds to 10.

34

The Middle Stone Age of Nigeria

Figure 27. Zenabi Fulata, work in progress, July 1976.

Figure 28. Zenabi Fulata, section1.

Figure 29. Zenabi Fulata, section 2.

35

The Middle Stone Age of Nigeria in its West African Context

Figure 30. Zenabi Tudo, March 1976.

(3) 1976-78 and Jos Museum collections The collection of artefacts made in Fulata in 1976 was systematic. Since the paddock does not reach to bedrock and only one tin-wash was exposed, the artefacts do have a degree of localisation both in space and in terms of vertical stratigraphy. The industry itself appears to be remarkably homogeneous, and it was felt that it could serve as a reliable basis of comparison with the finds already known from the site. Such a basis is needed because ever since Bond’s time it has been clear that several localities and more than one assemblage are involved. For this reason, a study was made of the material stored in the Jos museum over a period of weeks in March and April 1977. This provided the foundation for the comparative study presented to the 8th Panafrican Congress in Nairobi in the same year (AllsworthJones, 1980), summarised here in Table 3. The total given for Fulata differs marginally from that given in Nairobi because the finds fallen from the section in 1977 have been added in, but otherwise the figures are unchanged. According to the rather incomplete notes in the museum archives, it seems that the material from Zenabi was collected by a number of different people over many years, beginning with Fagg and Bond in 1947 and 1948. Bernard Fagg made additional visits in 1956 and 1962, and Angela Fagg also collected material from 1968 onwards. Soper was active in the area in 1962-1964, so this much at least is known. We possess a varying amount of information about the separate areas listed in the Table, as follows. (a) Karara, a paddock a quarter or half a mile north of the bridge (as shown in Bond’s sketch map), to which have been added the finds from Soper’s north paddock (ZNP); (b) areas A-E, about which nothing is known apart from their numbers; (c) the Main Paddock (ZMP), and (d) Paddock No. 1. The position regarding No. 1 Paddock is not satisfactory, in that whereas in 1962-64 Soper recorded no less than 1815 pieces from here (36 tools, 160 cores, and 36

The Middle Stone Age of Nigeria

1619 flakes and fragments) (Soper, personal communication) the museum’s catalogued collection now contains very few. No attempt has been made here to include the data from Zenabi No. 1 Paddock, since in the existing state of things it is obviously incomplete. In addition, it should be recalled that there is a certain amount of ambiguity over the relationship between the Main and No. 1 Paddock. Otherwise, there is no reason to think that the figures given here are seriously incomplete. (d) Zenabi Tudo, usually clearly labelled as such. The main typological and technological characteristics of the assemblages are summarised in Table 3; some of the typological indices have only a relative value in view of the small numbers of tools involved. Table 3. Zenabi artefact inventory.

Fulata

Karara

A-E

Main Paddock

Tudo

Total

Levallois flakes

3

14

4

3

1

25

Levallois blades

-

5

-

1

3

9

Zenabi Tools

Levallois points

1

1

-

2

1

5

Pseudo-levallois points

2

3

1

2

1

9

Limaces

8

1

1

-

-

10

Straight

14

3

9

1

-

27

Convex

13

4

4

3

1

25

Sidescrapers

Concave

8

3

3

-

1

15

Double

10

4

3

1

-

18

Convergent

7

3

1

3

-

14

Canted

4

3

2

1

-

10

Transverse straight

4

1

1

-

-

6

Transverse convex

6

3

3

3

-

15

Transverse concave

1

1

-

1

-

3

On ventral surface

3

2

5

2

1

13

With alternate retouch

-

1

2

-

1

4

Total sidescrapers

70

28

33

15

4

150

Endscrapers

4

1

1

-

-

6

Burins

2

1

2

-

-

5

Awls

-

-

-

2

-

2

Notches

1

-

1

1

-

3

Denticulates

12

9

2

2

2

27

Various

7

6

5

1

3

22

Trihedrals

-

-

-

2

-

2

Handaxes

-

-

-

1

-

1

Total tools

110

69

50

32

15

276

Initial preparatory

2

5

6

3

-

16

Irregular

4

11

5

2

1

23

Disc

1

23

11

5

9

49

Levallois flake/blade

-

3

4

4

-

11

Cores

37

The Middle Stone Age of Nigeria in its West African Context

Fulata

Karara

A-E

Main Paddock

Tudo

Total

-

2

-

-

-

2

1-platform flake/blade

11

6

10

4

4

35

2-platform flake/blade

1

-

1

1

-

3

Total cores

19

50

37

19

14

139

462

234

160

62

39

957

Cortical

95

20

28

14

4

161

Facetted

35

80

66

21

24

226

Zenabi Levallois point

Striking platforms (tools, flakes, blades) Plain

Dihedral

13

40

18

6

12

89

Removed

94

57

66

16

23

256

Total striking platforms

699

431

338

119

102

1689

ILty

3.64

28.99

8.00

18.75

14.13

IR

63.64

40.58

66.00

46.87

54.35

IFl

6.87

27.84

24.85

22.69

Indices

The typological categories employed are meant to coincide with those listed by François Bordes (1961). A word of explanation is required concerning the tool described here as a limace, as mentioned in the Introduction. ‘Les limaces typiques sont épaisses et symétriques, parfois trapues et larges’ (Bordes, 1961, 23; plate 13: 8, 9, 11). The pieces we have are perhaps strictly speaking more comparable to what Bordes referred to as proto-limaces (Bordes, 1961; plate 13: 10), the distinction between them and cores being sometimes quite difficult. In any case, they are an unavoidable and characteristic part of the inventory from Zenabi and other sites with which it can be compared. Some artefacts from Zenabi are illustrated in Figures 31-56. The artefacts in Figures 31-32 from Fulata and Karara were already illustrated in the Panafrican Congress volume (Allsworth-Jones, 1980). Figures 3335 are mainly artefacts found in 1977 and 1978, as fallen from the 1976 section or in situ in the 1978 deep sounding, respectively. Figures 55-56 illustrate other material from Fulata and Zenabi Tudo. The remainder of the illustrations are for the most part from the various localities mentioned in the Jos museum archives, organised in technological and typological groups. Where the provenance is quoted, this is as it appears in the museum records.

35.29

18.65

Figure 31. Zenabi Fulata. 1, limace. 2, angle burin. 3, double sidescraper on Levallois flake/blade. [PAJ 1980 Fig. 1. ABC].

38

The Middle Stone Age of Nigeria

Figure 32. Zenabi Fulata. 3, 1-platform flake/blade core. 5, endscraper. Zenabi Karara. 1, transverse convex sidescraper. 2, Levallois point. 4, disc core. [PAJ 1980 Fig. 2. ABCDE].

Figure 33. Zenabi Fulata 1976. 1 and 4, transverse and straight sidescrapers. Zenabi Fulata 1977, finds derived from 1976 section, upper tinwash. 2, limace. 3, flake.

Figure 35. Zenabi Fulata 1977, finds derived from 1976 section, upper tinwash. Initial/preparatory core.

Figure 34. Zenabi Fulata 1978. Lower tinwash, upper series. 2, plain platform flake. 3, waste flake. Base of deep sounding. 1, core-like chunk. 4, flake/blade. 5, waste flake.

Figure 36. Zenabi, Jos Museum. MP56/33. Initial Levallois core.

39

The Middle Stone Age of Nigeria in its West African Context

Figure 37. Zenabi, Jos Museum. MP56/33. Levallois flake/blade core.

Figure 38. Zenabi, Jos Museum. C. Levallois flake core.

Figure 39. Zenabi, Jos Museum. MP56/33. 1, Levallois core. Karara. 2, Levallois flake core.

Figure 40. Zenabi, Jos Museum. A47/10. 1, Levallois flake core. ZNP 1962. 2, Levallois point core. Z ‘west of spur’ 1948. 3, 1-platform flake/blade core. C47/10. 4, Levallois flake core.

(4) Nature of the industry Both typologically and technologically the industry from Zenabi Fulata makes a marked impression of homogeneity and idiosyncrasy. The number of unretouched Levallois tools is small, and there are no Levallois cores. 1-platform flake/blade cores are the most common. Plain striking platforms are predominant. Sidescrapers are the principal tool class, many of them massive and with a type of retouch which it is legitimate to refer to as ‘Quina’. This expression was used by Bordes to denote a form of steep or overhanging retouch which often occurs on thick blanks and probably is occasioned by that very fact (Bordes, 1961, 26: ‘retouche écailleuse souvent scalariforme … il suffit que le racloir soit fait sur éclat très épais 40

The Middle Stone Age of Nigeria

Figure 41. Zenabi, Jos Museum. A. 1-platform flake/ blade core. C. 2, 1-platform blade core.

Figure 42. Zenabi, Jos Museum. B47/10. 1 and 2, 1-platform flake/blade cores. Karara 1948. 3, 1-platform ‘boat-shape’ core. A ‘in situ’. 4, 2-platform flake/blade core.

Figure 43. Zenabi, Jos Museum. Karara ‘1/2 mile north of bridge’. Disc core.

Figure 44. Zenabi, Jos Museum. MP 1962 ‘L.G.’. Disc core.

41

The Middle Stone Age of Nigeria in its West African Context

Figure 45. Zenabi, Jos Museum. MP 1964. 1, disc core. MP56/34 ‘B.G.’. 2, disc core. Karara 1948. 3, disc core.

Figure 46. Zenabi, Jos Museum. Fulata. 1, Levallois flake. C47/10. 2, Levallois flake. Karara. 3, Levallois flake/point. A47/10. 4, Levallois blade.

Figure 47. Zenabi, Jos Museum. ZR ‘gravel’ 1962. 1, Levallois flake. Karara 1948. 2, Levallois flake. MP56/33. 3, Levallois blade. D. 4, retouched Levallois flake. ZNP 1962. 5, Levallois flake/blade.

Figure 48. Zenabi, Jos Museum. A47/10. 1, overshot blade. C. 2, overshot flake. ZMP 56/33. 3, pseudolevallois point. Karara. 4, pseudo-levallois point. 5, blade.

42

The Middle Stone Age of Nigeria

Figure 49. Zenabi, Jos Museum. A. 1, limace. Karara 1948. 2, limace. 3, flake.

Figure 50. Zenabi, Jos Museum. ZNP 1962. 1, limace. ZR 1962. 2, endscraper. Karara. 3, endscraper. A. 4, endscraper.

Figure 52. Zenabi, Jos Museum. ‘No.1 middle gravels I/S’. 1, convergent sidescraper with alternate retouch. ZR 1962 ‘I/S gravels’. 2, notch. ZMP. 3, canted sidescraper. Karara. 4, double sidescraper.

Figure 51. Zenabi, Jos Museum. MP56/33. 1, transverse convex sidescraper. Karara. 2, multiple canted sidescraper. A. 3, multiple dihedral burin. 4, burin on retouched truncation.

43

The Middle Stone Age of Nigeria in its West African Context

Figure 53. Zenabi, Jos Museum. MP56/33. 1, trihedral. Karara. 2, double sidescraper.

Figure 54. Zenabi, Jos Museum. ZMP. 1, trihedral. B. 2, canted sidescraper on Levallois flake.

Figure 55. Zenabi Fulata 1976. 1, canted sidescraper. 2, transverse convex sidescraper. 3, sidescraper on ventral surface. Zenabi Tudo 1968. 4, Levallois point.

Figure 56. Zenabi Tudo. 1, disc core. 2, pseudolevallois point. 3, Levallois flake. 4, Levallois blade.

pour qu’il prenne ce type’). It is in fact sometimes a problem at Fulata, as elsewhere at Zenabi, to draw the boundary between cores and steeply retouched tools, particularly in the case of those very distinctive objects which have been classified as limaces (Figures 31.1, 33.2). Both endscrapers and burins make an appearance, and although there are not many of them there is nothing crude or fortuitous about them. As Bond noted, the artefacts are almost uniformly fresh and unrolled, and even the smallest splinters seem to have been preserved. The assemblage thus constitutes an excellent standard for comparison. It can be said straightaway that elements reminiscent of Fulata do recur at all the other localities but quantitatively there are differences, and in some instances both typology and the state of preservation of certain artefacts suggest that more than one entity may be involved. 44

The Middle Stone Age of Nigeria

Zenabi Karara clearly reveals a difference of emphasis in respect to cores and débitage techniques. Disc cores are much in evidence, also Levallois flake and point cores; unretouched Levallois tools make up 29% of the total, and the proportion of facetted and dihedral striking platforms is greatly increased. Disc cores also occur in fair numbers at all other localities, and it cannot perhaps be excluded that their poor showing at Fulata is due to chance alone. The same argument might apply to the rather high Levallois percentage at Karara. In other respects there is much to link the two: the frequency and nature of the sidescrapers, the presence of limaces (Figures 49:2, 50:1) and similar difficulties in distinguishing between certain cores and tools, the occurrence of endscrapers and burins, and the general state of preservation of the majority of the artefacts. The same elements, although in different proportions, can be seen in areas A-E; there is no striking qualitative distinction. On the basis of Soper’s notes and the few pieces which I was able to see in the Jos museum, the same may have been largely true for Zenabi No.1 paddock; the majority of the tools, it seems, were sidescrapers, there were both disc and Levallois cores, but by far the greater part of the striking platforms were plain. When summarising for comparative purposes, in an earlier account, it was felt to be sufficient to combine the figures for Fulata and Karara only, in order to give a representative impression of the industry as a whole (Allsworth-Jones, 1986: 159-160, Table 9.1). The collection from Zenabi Main Paddock, however, contains certain unexpected features, along with familiar items such as massive sidescrapers comparable to those from Fulata or Karara. Some of the artefacts, including particularly Levallois flake/blades and cores, are definitely corroded and of a different state of preservation than the remainder; whereas two trihedrals (Figures 53.1, 54.1) and one perfectly characteristic handaxe (the provenance of which is clearly marked) are a distinctly new typological element. No handaxes have hitherto been reported at Zenabi (Soper, 1965). Unfortunately the in situ material is too sparse to document convincingly any directional change from lower to upper gravels. In general the material is fresh and sharp edged, with a consistent dominance of plain platform flakes. Soper does record, however, that the artefacts found by him in 1964 in the lower gravel included both fresh and abraded pieces. At Zenabi Tudo there are likewise a minority of artefacts, particularly unretouched Levallois tools, which are much corroded. The majority, including disc cores and sidescrapers, are quite sharp edged and generally comparable to the industry below the Falls. Anozie (1975) studied 474 artefacts from Karara then kept at the Jos museum. They included 29 tools, 71 cores, 18 blades, and 356 flakes. These totals differ somewhat from the figures given here. According to him, the modal figures for flake length, width, and thickness at the site were 70-79, 40-49, and 10-19 mm respectively. The histograms (Anozie, 1975: plate 140) show that the lengths range from 30-39 to 150-159 mm, and the widths from 20-29 to 130-139 mm. The cores have length modal values of 70-79 mm, with width modal values of 60-69 mm. They range from 40-49 to 200-209 mm, and 30-39 mm to 150159 mm, respectively. In general, the histograms are more or less positively skewed. For comparison, measurements of the tools and cores from the entire Zenabi Jos museum collection as recorded by myself in 1977 (but not Fulata) were analysed by Steve Daniels at the ABU Centre for Nigerian Cultural Studies in 1978, using his AQUA computer program. The results are summarised in Table 4. The totals, with 143 Table 4. Zenabi artefact measurements.

Zenabi

Mean (mm)

SD (mm)

Min (mm)

Max (mm)

length

77.48

27.81

37

272

width

65.76

23.73

25

148

thickness

26.71

14.47

7

93

length

90.26

34.38

35

208

width

71.77

25.36

30

182

thickness

47.74

21.09

9

115

N=143 Tools

N=122 Cores

45

The Middle Stone Age of Nigeria in its West African Context

tools and 122 cores, differ somewhat from those given here in Table 3, but they are very similar in their composition, the cores being divided into five categories rather than seven (with the amalgamation of all Levallois cores and all flake/blade cores). As might be expected, the mean dimensions of the initial or preparatory cores, and of the Levallois cores, were somewhat larger than the core averages, 135.67 x 97.73 x 67.47 mm and 104.50 x 84.20 x 51.30 mm respectively. Conclusion It seems that the greater part of the industry from Zenabi as a whole does constitute a homogeneous unit, on the whole comparable to that from Fulata, although there are quantitative differences of emphasis, particularly with regard to unretouched Levallois tools and disc cores, in the different areas. This industry (also represented at Yelwa and Tibchi) is markedly idiosyncratic and can be expected, in comparison with other finds already known elsewhere, e.g. at Mai Lumba, to constitute a special facies of the Nigerian MSA. At Zenabi Main Paddock (also possibly Zenabi Tudo) there are indications, however, that more than one entity may be represented at the site: the presence of a handaxe and corroded Levallois flake/blades and cores suggest that there may have been some directional change through time. It is clear that this account leaves many questions unanswered. Some, such as the meaning of so young a radiocarbon date supposedly prior to or associated with an MSA industry, and the precise relationship between the individual stratigraphic sequences described in different parts of the site, could only be settled by further fieldwork, but, in view of the fact that much of it is now abandoned and flooded, the prospects for such a re-examination are slight. In the meantime, the purpose of the re-examination of the artefacts in the Jos museum, which has been carried out, has been to establish more exactly what the nature of the industry is, since (despite the fact that Zenabi is the most frequently quoted Nigerian MSA site) only the most summary accounts of it have hitherto been published. In particular, the answers were sought to two questions: to what degree, in the light of the new finds from Fulata, could it be regarded as a single entity; and, with the available stratigraphic information, could any developmental tendencies in it be traced? Since the finds have been gathered over many years largely as a by-product of mining operations, and the majority lack definite stratigraphic provenance, any conclusions that can be reached are necessarily somewhat tentative. Mai Lumba Mai Lumba (10° 48’ N 8° 59’ E) is situated on the western side of the Tibchi and Yeli hills but within the ring complex where the Kuskerri river descends to the plain (Figure 57). Leases 15004, 13118, 988, 963, and 10796 are strung out along the river in a north-west/south-east direction. Angela Fagg reported in January 1976 that artefacts were being turned up as a result of mining operations in this area, and for this reason a first visit was made to the site on 24 March 1976. At that time it was found that mining was going on in lease 988, on the left hand side of the river, directed by Mr J.O. Coker on behalf of the Bisichi Jantar company (Figures 58 and 59). The paddock being worked was about 4-5 metres deep, with at least two bands of what appeared to be blue clay alternating with sands and gravels (Allsworth-Jones, 1981). We conducted more extensive work in lease 988 during an eight day period between 25 June and 8 July of that year. A large number of artefacts, presumed to be associated with the basal tin-wash, were collected from spoil heaps by the sluicing gullies, clearly the by-product of the mining work carried out at the time of my first visit. A somewhat smaller but still considerable number were collected in lease 10796, south of Mai Lumba village and upstream from 988. These originated not from a regular paddock such as operated by Mr Coker, but from a series of small scale works using the ‘banga’ system, whereby the river is dammed and diverted and the water used for sluicing, in order to recover cassiterite. One such small scale sluicing device, referred to as a ‘kafanchan’, in operation during our visit in lease 963, is shown at Fig. 60. A few surface finds were also made downstream from 46

The Middle Stone Age of Nigeria

Figure 57. Tibchi and Mai Lumba and their position in relation to the Tibchi and Yeli hills.

988, in the vicinity of lease 13113. The same was the case further upstream, above the steep escarpment of the Tibchi and Yeli hills, in lease 4788, near the village of Mai Lumba Tudo, where small scale ‘banga’ workings were in progress. Thus it seems that the whole of the upper course of the river Kuskerri, both above and below the escarpment, contains MSA style artefacts, to a greater or lesser degree. A further visit to the site was made in the period between 17 and 29 April 1978 (Allsworth-Jones, 1987). It transpired that very little tin mining had been carried out in this area since 1976. The paddock in lease 988 formerly operated by the Bisichi Jantar company and then flooded had in the meantime been 47

The Middle Stone Age of Nigeria in its West African Context

Figure 58. Mai Lumba lease 988, 24 March 1976.

Figure 59. Mai Lumba lease 988, 24 March 1976.

48

The Middle Stone Age of Nigeria

Figure 60. Mai Lumba lease 963, ‘kafanchan’ working, June 1976.

invaded by the river and largely destroyed (Figures 61 and 62). Nonetheless, the walls of the paddock on the landward side were largely intact, and it was decided to obtain a record of the stratigraphy here, as indicated in the plan at Figure 63. The paddock is shown in the south-west corner of lease 988, near its junction with lease 13113. Sections 1 and 2 are on its southern side, and the positions of sections 3 and 4, on the hillside to the south-west, are also indicated. A pump was employed in an attempt to control the water level (Figure 64). It was necessary to cut two sections, since the first was incomplete and disturbed by flowing water, but in the second two tin-washes were revealed (Figures 65-68; Figure 63; Table 5). Taking the two sections together, the composite record of the succession can be regarded as reasonably satisfactory. Section 1 shows a series of leached and mottled silts and clays, some of which may be quite recent, above a basal ‘blue’ clay and river bed gravel. Section 2 contains a yellowish-red sand layer up to 1.5 metres thick intersected by a yellowish-red gravelly sand horizon, clearly identified by the miners as an Upper Tin-wash. The pinkish grey coarse sand and gravel layer at the base constituted the Lower tin-wash. Artefacts found in situ in the Upper tin-wash included 1 miscellaneous retouched piece and 4 flake/blades, whereas those from the Lower tin-wash included 1 disc core and 1 facetted platform flake. Although the two series are not large, both are MSA in character and entirely typical for the site as a whole. It was also decided to investigate the nature of the red lateritic deposits on bluffs overlooking the plain at this point, with a view to reconstructing the history of erosion and sedimentation in the valley of the Kuskerri, as far as possible. Two further trenches (3 and 4) were dug for this purpose, one on the slope and one at the foot of the slope (Figures 69 and 70; Figure 63; Table 5). If the initial interpretation of the stratigraphy is correct, it may be that the red lateritic deposits constitute the remnants of an earlier valley filling which was eroded and truncated to produce the present valley structure. It may be that the situation is somewhat analogous to that at Zenabi, where Bond (1956) 49

The Middle Stone Age of Nigeria in its West African Context

Table 5. Mai Lumba lease 988. Stratigraphy, sections 1-4. Section 1 Layer No. Description

Munsell colour

1

Reddish brown topsoil

5 YR 5/4

2

Pinkish white leached silt and clay

5 YR 8/2

3

Dark reddish brown silt and clay

5 YR 3/2

4

Pinkish grey leached silt and clay with mottled reddish brown silt and clay

7.5 YR 7/2 + 5 YR 4/4

5

Reddish brown silt and clay with mottled grey silt and clay

5 YR 4/4 + 5 YR 6/1

6

Pinkish grey silt and clay

7.5 YR 7/2

7

Reddish brown silt and clay with mottled grey silt and clay

5 YR 4/4 + 5 YR 6/1

8

Yellowish red clay and silt with mottled dark grey clay and silt

5 YR 4/6 + 5 YR 4/1

9

Strong brown gravel

7.5 YR 5/6

10

Strong brown silt and clay with mottled grey silt and clay

7.5 YR 5/6 + 7.5 YR 5/0

11

Basal “blue” clay and river bed gravel

Section 2 Layer No.

Description

Munsell colour

1

Dark brown topsoil

7.5 YR 4/4

12

Reddish brown silt and clay

5 YR 4/4

13

Yellowish red sand with mottled red sand

5 YR 5/8 + 2.5 YR 4/8

14

Yellowish red gravelly sand with mottled red sand Upper Tin-wash

5 YR 5/8 + 2.5 YR 4/8

15

Yellowish red sand with mottled red sand

5 YR 5/8 + 2.5 YR 4/8

16

Pinkish grey coarse sand and gravel with stones and boulders Lower Tin-wash

5 YR 6/2

Layer No.

Description

Munsell colour

1

Red gravelly soil - laterite

2.5 YR 4/8

2

Red decomposed bedrock

2.5 YR 4/6

3

Reddish brown silt + angular coarse fraction

5 YR 5/3

Layer No.

Description

Munsell colour

1

Reddish grey topsoil

5 YR 5/2

2

Yellowish red subsoil

5 YR 5/6

3

Red decomposing bedrock

2.5 YR 4/8

Section 3

Section 4

observed that the Gaiya river had cut through a laterite crust, the outwash fans and their contained palaeolithic material therefore being younger than this crust. The lateritic deposits at Mai Lumba are however by no means as indurated as those at Zenabi. While we were at the site, further collections of artefacts were made in the vicinity of the paddock in lease 988. They mainly come from three heaps of stones derived from tin-mining operations in this area. The material did not differ markedly from that previously collected at the site, the artefacts in general being rather heavily worn or corroded, some quite deeply patinated. In that respect, there is a marked contrast compared with the material from Zenabi. On the previous occasion, E.C. Ike examined some of the artefacts collected and determined that the great majority were rhyolite, of which there is an extensive outcrop north of the site, extending all the way to Tibchi. It contains many localised varieties, and if a more exhaustive examination were carried out it would be possible to trace the exact geological origin of many of these pieces. Other raw materials identified include quartz, basalt, ignimbrite, microsyenite, and dolerite. A complete list of the finds from leases 988 and 10796, from both 1976 and 1978, is at Table 6. 50

The Middle Stone Age of Nigeria

Figure 61. Mai Lumba lease 988, 21 April 1978.

Figure 62. Mai Lumba lease 988, 21 April 1978.

51

Figure 63. Mai Lumba 1976-1978 lease 988, plan and sections 1-4.

The Middle Stone Age of Nigeria in its West African Context

52

The Middle Stone Age of Nigeria

Figure 64. Mai Lumba lease 988, pump in operation, 15 April 1978.

Figure 65. Mai Lumba lease 988, section 1, 23 April 1978.

53

The Middle Stone Age of Nigeria in its West African Context

Figure 66. Mai Lumba lease 988, section 1, 23 April 1978.

Figure 67. Mai Lumba lease 988, section 2, 24 April 1978.

54

The Middle Stone Age of Nigeria

Figure 68. Mai Lumba lease 988, section 2, 25 April 1978.

Figure 69. Mai Lumba lease 988, section 3, 29 April 1978.

55

The Middle Stone Age of Nigeria in its West African Context

With 129 chipped stone tools, 109 cores, and 1412 blanks with identified striking platforms, it is identical to the combined list published previously (Allsworth-Jones, 1986, Table 9.1). It does not include the stray finds made at various localities, nor those found at Mai Lumba Tudo. These accounted, respectively, for 7 tools, 5 cores, and 530 blanks with identified striking platforms, and for 5 tools, 2 cores, and 37 blanks with identified striking platforms. In their general appearance, these artefacts do not differ from those collected in leases 988 and 10796 and (with the exception of a few pieces mentioned below) there is an overall impression of great homogeneity in the industry throughout the Kuskerri valley. A number of artefact drawings are at Figures 71-88, and photographs at Figures 89-95. Some comments are in order. With respect to homogeneity, it will be noted that 5 ground stone axes were found, and in 1976 a few fragments of pottery also, which evidently came from the top of the sequence in lease 988. The exact relationship between the ground stone axes (Figures 86.1 and 3, 87.3, 89 and 90) Figure 70. Mai Lumba lease 988, section 3, 29 April 1978. and the rest of the inventory cannot be determined. Figure 72.2-5 shows the artefacts recovered in 1978 from the Upper and Lower Tin-wash in lease 998 section 2. Figure 86.4 shows a double/convergent sidescraper from Mai Lumba Tudo, which is typical for the site as a whole. One of the two bifaces found at the site is at Figure 88. Apart from sidescrapers, Levallois blanks and cores are predominant among the artefacts illustrated from the site. The indices for Mai Lumba (ILty 55.8, IR 27.9, IFl 31.8) are the same as those already published, indicating a pronounced contrast between Zenabi and Tibchi on the one hand, and this site on the other in terms of technology and typology (Allsworth-Jones, 1986: 160). As at Zenabi, two tin-washes were identified, but it is not known whether they are directly comparable to their counterparts at that site, and no organic material for radiocarbon dating was available at Mai Lumba. Hence it cannot be ascertained what the chronological relationship might be between the two sites, and it would be risky to assume that the greater degree of wear on the artefacts from Mai Lumba implies a greater antiquity. Yada Gungume Yada Gungume (10° 52’ N, 9° 54’ E) is on the north-east periphery of the Tibchi-Yeli hills where the river Yelda comes out onto the plains. A short inspection on 3 July 1976 revealed a number of scattered finds within lease 10542, along the course of the river, where there were a number of old abandoned paddocks (Allsworth-Jones, 1981). Excavation was out of the question, but technologically and typologically the 56

The Middle Stone Age of Nigeria

Table 6. Mai Lumba and Yada Gungume. Artefact inventory. Mai Lumba Yada Gungume Tools Levallois flakes 46 6 Levallois blades 15 Levallois points 11 Pseudo-levallois points 6 Limaces 1 Sidescrapers Straight 6 2 Convex 12 Concave 1 Double 5 Convergent 5 Canted 1 Transverse straight 1 Transverse convex 1 On ventral surface 4 1 Total sidescrapers 36 Endscrapers 1 Burins 1 Notches 3 Denticulates 2 Bifaces 2 2 Various 5 3 Total tools 129 14 Ground stone axes 5 Cores Initial/preparatory 8 Irregular 18 1 Disc 51 4 Levallois point 3 1 Levallois flake 11 1 Levallois blade 5 1-platform flake/blade 7 3 2-platform flake/blade 6 Total cores 109 10 Striking platforms (tools, flakes, blades) Plain 774 18 Cortical 104 3 Facetted 339 12 Dihedral 110 3 Removed 85 1 Total striking platforms 1412 37 Indices ILty 55.8 IR 27.9 IFl 31.8

Figure 71. Mai Lumba. 1 and 2, Levallois blade and flake/blade. 3, Levallois blade core. 4, convergent sidescraper. [PAJ 1986 Fig. 9.2].

Figure 72. Mai Lumba. 1, Levallois flake core [PAJ 1986 Fig. 9.4.1]. 2, flake/blade. 3 and 4, flakes. 5, disc core. 2-5, excavated 1978, 2-4 upper tinwash, 5 lower tinwash.

57

The Middle Stone Age of Nigeria in its West African Context

Figure 73. Mai Lumba. Initial core.

Figure 74. Mai Lumba. Levallois cores. 1, flake/point core. 2-3, flake cores. 4, 2-platform flake/blade core.

Figure 75. Mai Lumba. 1 and 2, 2-platform flake/blade cores, 2 struck from opposing directions.

Figure 76. Mai Lumba. 1 and 4, 2-platform flake/blade cores. 2 and 3, 1-platform flake/blade cores.

Figure 77. Mai Lumba. 1 and 2, disc cores. 3 and 4, 1-platform flake/blade cores.

Figure 78. Mai Lumba. 1-4, disc cores.

58

The Middle Stone Age of Nigeria

Figure 79. Mai Lumba. 1 and 4, Levallois points. 2-3 and 5, Levallois flake/blades.

Figure 80. Mai Lumba. 1-6, Levallois flakes and flake/ blades.

Figure 81. Mai Lumba. 1-5, Levallois flakes.

Figure 82. Mai Lumba. 1, flake/blade. 2-4, flakes.

59

The Middle Stone Age of Nigeria in its West African Context

Figure 83. Mai Lumba. 2, blade. 1, 3-4, flakes.

Figure 84. Mai Lumba. 1 and 2, convex and double sidescrapers. 3, pseudo-levallois point. 4-5, overshot flake/blades.

Figure 85. Mai Lumba. 1 and 5, sidescrapers on ventral surface. 2 and 3, straight and convex sidescrapers. 4, Levallois flake/blade.

Figure 86. Mai Lumba. 1 and 3, broken ground stone axes. 4, Mai Lumba Tudo, double/convergent sidescraper.

60

The Middle Stone Age of Nigeria

Figure 87. Mai Lumba. 1, transverse convex sidescraper. 2, convergent sidescraper. 3, broken ground stone axe.

Figure 89. Mai Lumba lease 988, ground stone axe.

Figure 88. Mai Lumba. Biface.

Figure 90. Mai Lumba lease 988, ground stone axes.

61

The Middle Stone Age of Nigeria in its West African Context

Figure 91. Mai Lumba, disc and 2-platform Levallois flake/blade core.

Figure 92. Mai Lumba, disc and Levallois flake core.

Figure 93. Mai Lumba, 2-platform flake/blade cores.

62

The Middle Stone Age of Nigeria

Figure 94. Mai Lumba, microsyenite initial core.

Figure 95. Mai Lumba, rare raw materials at the site.

63

Figure 96. Tibchi 1976 paddock, plan and sections.

The Middle Stone Age of Nigeria in its West African Context

64

The Middle Stone Age of Nigeria

collection of 61 finds (Table 6) is of some interest in the context of the Palaeolithic settlement of the area as a whole. One of the two bifaces (Figure 101.1) could well occur in an MSA context, but the other (Allsworth-Jones, 1981, Fig. 4.2) resembles a handaxe and may be indicative of an older horizon. Otherwise the artefacts, including Levallois flakes and cores, conform to the general MSA pattern as seen at Mai Lumba and Tibchi, the small number of finds not permitting any more definite conclusion. Rhyolite is prominent as a raw material. The river Yelda does not pass through rhyolite country, whereas, as pointed out by E.C. Ike, certain of the varieties represented are identical to those found at Tibchi. This does point to a targeted raw material exploitation strategy and possibly a connection between the two sites. Tibchi Tibchi (10° 51’ N, 8° 59’ E) is at the north-western edge of the Tibchi-Yeli complex, where the Tibchi river comes out onto the plain (Figure 57). It was first visited by our team on July 1 1976, following information received from E.C. Ike. A paddock south of the Bisichi Jantar lease 359 was being worked by a private contractor Mr Idi Jila, who kindly gave us access to the area. We made a plan of the paddock and recorded two sections over a two day period (Figures 96 and 97-100). The drawn sections are at Figure 96, with a description at Table 7. Two tin-washes were located, the lower being exceptionally stony, with large numbers of rounded boulders (Figure 99). One flake was found in situ, suggesting that this was the source of the artefacts discovered in the vicinity. These were located partly in this paddock and partly as surface finds in lease 359 itself. All together a collection was made of 4 tools, 4 cores, and 60 blanks with recognisable striking platforms (Allsworth-Jones, 1981). Ike commented that many of the artefacts were made of fine-grained rhyolite, of a type which can be found only in this part of the ring complex. A return visit in May 1978 revealed a rather different situation. Between 1976 and 1978, the paddock south of lease 359 had been considerably extended, and many heaps of stones derived from the paddock were discovered in the vicinity. These heaps contained abundant artefacts. It is possible that the reason for

Figure 97. Tibchi, general view of paddock, 2 July 1976.

65

The Middle Stone Age of Nigeria in its West African Context

this may lie in the nature of the tin-wash encountered by the miners. To judge from the appearance of the stones, the area exploited more recently may have contained smaller and more angular material, forming a more natural ‘home’ for the artefacts, than the large rounded boulders seen in 1976. In any case the much larger quantity of material available allowed the industry to be more comprehensively described, and also put it in a somewhat different light. Apart from the former paddock, a small new paddock had also been opened some 200 metres NNW of it, within the bounds of lease 359. This also produced a modest series of artefacts similar to the ones from the main site. All told, the material located at this time amounted to 42 tools, 48 cores, and 354 blanks with recognisable striking platforms (Allsworth-Jones, 1987). A section in the new paddock showed that a series of sands and clays overlay a single tinwash, with no visible signs of the upper tin-wash detected in 1976.

Figure 98. Tibchi, section 1, 2 July 1976.

Figure 99. Tibchi, tin-miners removing lower tin-wash, 2 July 1976.

66

The Middle Stone Age of Nigeria

Figure 101. Yada Gungume. 1, biface. Tibchi. 2, 2-platform core. 3, sidescraper on ventral surface.

Figure 100. Tibchi, section 2, 2 July 1976.

The combined artefact totals from 1976 and 1978 are recorded in Table 8. These totals are the same as those previously published (Allsworth-Jones, 1986, Table 9.1) except for the addition of two cores. The indices descriptive of the industry are also the same, with a Levallois typological index (ILty) of 15.22, a sidescraper index (IR) of 54.35, and a facetted and dihedral platform Table 7. Tibchi 1976 paddock. Stratigraphy sections 1 and 2. Section 1. Layer No. Description

Munsell colour

1

Dark brown topsoil

10 YR 4/3

2

Yellowish-red sand

5 YR 4/8

3

Red sand

2.5 YR 4/6

4

Yellowish-red sand, with black inclusions Upper tin-wash

5 YR 4/6

5

Light brownish-grey grit

10 YR 6/2

6

Grey gritty clay

10 YR 6/1

7

Light brownish-grey clay and stones Lower tin-wash

10 YR 6/2

Section 2. Layers 1, 2, and 6 in section 2 correspond to their counterparts in section 1. Layers 8 and 9 are stratigraphically equivalent to layers 3-5 in section 1, but are not identical. 8

Reddish-yellow silt

7.5 YR 6/6

9

Red gravel, with horizontal bedding

2.5 YR 4/8

67

The Middle Stone Age of Nigeria in its West African Context

Table 8. Yelwa and Tibchi. Artefact inventory. Tools

Yelwa

Tibchi

Levallois flakes

3

Levallois blades

4

Levallois points

1

Limaces

1

4

Straight

2

2

Convex

3

6

Sidescrapers

Concave

2

Double

1

2

Convergent

3

1

Canted

3

4

Transverse convex

4

4

Transverse concave

2

On ventral surface

3

With alternate retouch

1

1

Total sidescrapers

19

25

Notches

1

1

Denticulates

2

Bifaces

1

Various

6

Total tools

22

Figure 102. Tibchi. 1, disc core. 2, Levallois/disc core. 3 and 4, disc cores.

46

index (IFl) of 19.81. It was on the basis of these indices, in comparison with those from Mai Lumba (ILty 55.8, IR 27.9, IFl 31.8) that I suggested one might distinguish a ‘Mai Lumba’ versus a ‘Zenabi’ facies of the Nigerian Middle Stone Age (Allsworth-Jones, 1986:160). In the summary previously provided (AllsworthJones, 1986, Table 9.1) the indices for Zenabi were based on the results from Fulata and Karara only. If the combined indices for all the localities are now used (as in Table 3) it can be seen that Zenabi and Tibchi are practically identical (Zenabi ILty 14.13, IR 54.35, IFl 18.65), marginally more so than if the figures for Fulata and Karara alone are used. A selection of artefacts from the site have been drawn and are at Figures 101-111.

Cores Initial/preparatory

1

3

Irregular

4

8

Disc

5

22

Levallois flake/blade

1

6

1-platform flake/blade

5

13

2-platform flake/blade

2

Total cores

18

52

Striking platforms (tools, flakes, blades) Plain

79

256

Cortical

15

43

Facetted

16

60

Dihedral

3

22

Removed

24

33

Total striking platforms

137

414

Attention is drawn to the large numbers of limaces, the way in which they grade into other forms (Figures 107.1, 108.1 and 3, 109.1, 110, and 111.1) being particularly expressive. Photographs of a few of these artefacts are at Figures 112-114.

Indices ILty

15.22

IR

54.35

IFl

13.87

19.81

68

The Middle Stone Age of Nigeria

Figure 103. Tibchi. 1, Levallois flake core. 2, Levallois/disc core. 3, 1-platform core/chopper. 4, 1-platform core.

Figure 104. Tibchi. 1, 1-platform core. 2, Levallois blade. 3, Levallois flake/blade. 4, overshot Levallois flake/blade.

Figure 105. Tibchi. 1, blade. 2 and 5, flake/blades. 3 and 7, Figure 106. Tibchi. 1, pseudo-levallois point. 2, 5 and 6, flakes. Levallois flakes. 4, convex sidescraper. 6, Levallois blade. 8, 3, canted sidescraper. 4, transverse convex sidescraper. 7, pseudo-levallois point. blade.

69

The Middle Stone Age of Nigeria in its West African Context

Figure 107. Tibchi. 1, convergent sidescraper/limace. 2, transverse convex sidescraper. 3, sidescraper on ventral surface. 4, straight sidescraper with thinned back.

Figure 108. Tibchi. 1, double sidescraper/limace. 2, canted sidescraper/core. 3, limace.

Figure 109. Tibchi. 1, limace. 2, concave sidescraper. 3, biface.

Figure 110. Tibchi. Canted sidescraper/limace.

70

The Middle Stone Age of Nigeria

Figure 112. Tibchi, limaces (L and R), dorsal view.

It will be noted that, as at Zenabi and Mai Lumba, two tin-washes were identified at this site. The tinFigure 111. Tibchi. 1, limace. 2 and 3, transverse convex sidescrapers. wash shown in Figure 99 shows little resemblance to those already encountered, suggesting that its mode of emplacement was somewhat different. Nonetheless, there is a remarkable degree of resemblance between the industries of Zenabi and Tibchi, which does not seem to have anything to do with raw material constraints, or any obvious functional necessity. Yelwa Yelwa (10° 48’ N, 8° 44’ E) is about 5 km north-west of Zenabi, also at the foot of the Liruei Hills, where the Mallam (or Baba) river descends to the plain (Figures 115 and 116). Bond (1956) noted that there was an outwash fan similar to the one at Zenabi below the Falls, where the ‘red series’ was well exposed, but the exposures were not sufficiently extensive to allow the full succession to

Figure 113. Tibchi, limace (L), side view.

Figure 114. Tibchi, sidescrapers (L + centre), Levallois blade and pseudo-levallois point (R).

71

Figure 115. Yelwa 1978 lease 4311, plan and section.

The Middle Stone Age of Nigeria in its West African Context

72

The Middle Stone Age of Nigeria

Figure 116. Yelwa village, March 1977.

be worked out. In 1962-1964 Soper visited the site, where he suggested that two mineral-bearing gravels could be distinguished, and some artefacts were deposited in the museum in Jos (Allsworth-Jones, 1980). According to the museum archives, other finds from the site go back as far as 1947-1948 and 1956. In 1977, the total number of artefacts recorded in the archives amounted to 22 tools, 8 cores, and 97 blanks with recognisable striking platforms (Allsworth-Jones, 1980, Table 1). Several visits to the site were also made in the period 1976-1978. At first it appeared that all the former paddocks in the vicinity of the village had been abandoned and flooded, but on 26 March 1977 it was found that some small scale mining operations were going on north of the village in lease 11590. 12 artefacts were collected from workings in the river bed, including three cores (Allsworth-Jones, 1981). On 10 April 1978, it was ascertained that a new paddock had been opened by Mr J.O. Coker, Mallamawa area manager for the Bisichi Jantar company, along the west bank of the river in lease 4311, south-west of the village between it and the Falls (Allsworth-Jones, 1987). According to Mr Coker (and as could be seen in the paddock) there was a thick tin-wash at the base, and higher up in the section, a layer yielding columbite (Figures 117 and 118). Unfortunately, by the time we came to do our work at the site in May 1978, Mr Coker had already abandoned the paddock, in view of the early rains and the poor yield of cassiterite, and it was already flooded. However, it was still possible to do some work at the site. The position of the paddock within lease 4311 and in relation to the river was established by surveying, and a section was cut along the western river bank immediately adjacent to the paddock (Figure 115). The stratigraphic succession is the same as within the paddock (Table 9). Table 9. Yelwa lease 4311. Stratigraphy. Layer No.

Description

Munsell colour

1

Reddish-brown topsoil

5 YR 5/4

2

Red clay

2.5 YR 5/8

3

Red columbite layer : uneven bedding, coarse sand

2.5 YR 4/8

4

Pink clay, mottled white powdery

7.5 YR 8/4

5

Yellowish red very coarse sand

5 YR 5/6

6

Pinkish grey mottled clay, with gravel lines

5 YR 7/2

7

Basal tin-wash

73

The Middle Stone Age of Nigeria in its West African Context

Figure 117. Yelwa lease 4311, general view, 10 April 1978.

Figure 118. Yelwa lease 4311, mining work in progress, 10 April 1978.

74

Figure 119. Yelwa, Jos Museum. 1, limace. 3 and 4, transverse convex sidescrapers. 2, YMP 1962, blade.

The Middle Stone Age of Nigeria

Figure 121. Yelwa 1978. 1, 2-platform core. 2, disc core. 3, 1-platform flake/blade core.

Figure 120. Yelwa, Jos Museum. YMP LG 1962. 1, initial core. 2, 1-platform flake/blade core. Yelwa 56/31. 3, disc core. 4, convergent sidescraper.

Figure 123. Yelwa 1978. Initial/preparatory core, in situ in tinwash.

Figure 122. Yelwa 1978. 1, 1-platform flake/blade core. 2, flake.

Since the river was in flood, we were not able to reach the bottom of the basal tin-wash, but even in the small area opened up an initial or preparatory core was found in situ in this tin-wash. Other artefacts were located in the heaps of stones brought up from the paddock. They clearly come from the basal tinwash, since only this material was brought up for sluicing and the columbite layer had already been largely removed. In addition, a few stray finds were made within the area of the lease. All together there were seven cores and 40 blanks with recognisable striking platforms. 75

The Middle Stone Age of Nigeria in its West African Context

Figure 124. Ningi Hills: Physical Geography.

The complete list of finds, both from the Jos museum collections and from the more recent fieldwork, is at Table 8. No new tools have been added to those from the museum, but there are now 18 cores and 137 blanks with recognisable striking platforms. Representative artefacts are illustrated in Figures 119-123. Attention should be drawn to the characteristic limace at Figure 119.1. The initial or preparatory core found at the base of the section is at Figure 123. The condition of the artefacts is usually quite fresh. Perhaps unsurprisingly, the industry in general is closely comparable to that from Zenabi. As at Zenabi, there are two tin-washes, a distinguishing mark being that the upper tin-wash at Yelwa contains columbite rather then cassiterite. Ningi Hills The Ningi Hills – stretching in a long line to the north and north-east of the Tibchi-Yeli complex – have remained until recently a poorly explored region, at least so far as prehistoric archaeology is concerned. We do however now have an up to date geological survey of the entire area (Turner and Bowden, 1979) on the basis of which the maps here have been constructed (Figures 124 and 125). From this it is clear that the Ningi Hills consist of a number of overlapping ring structures, generally with granite porphyry ring dykes, 76

Figure 125. Ningi Hills: Geology.

The Middle Stone Age of Nigeria

containing extensive areas of rhyolite, as well as basalt, granophyre, and syenite. There are two small areas of biotite granite in the west, centred on Koluki, and a larger area in the east, forming a rough triangle between Tabela and Jigawa. The latter was already being exploited to some extent for cassiterite in 1923, to judge from the published map of that time (Falconer and Raeburn, 1923), and it was known that at least some small scale workings had continued since then. Since the geological conditions matched those already known in the area already investigated, and the Hills are not far from the Tibchi-Yeli complex, there seemed to be a reasonable chance that reconnaissance would reveal further MSA sites and industries. Hence an exploratory visit to the area was made in the period between 1 and 9 May 1978, and some information was obtained (Allsworth-Jones, 1987). This is no more than indicative but it is clear that the area will merit further work in the future. A number of sites producing a very small amount of material were located, rather larger collections being made at Jigawa and Tabela (Table 10). In the western part of the Ningi Hills, a short visit was made to Koluki, in the central part of the complex, by the Dogwalo river (Figures 126 and 127). It seems that no mining has been carried out in this area. A search along the bed of the river revealed a few artefacts, none really diagnostic (1 1-platform flake/blade core,

77

The Middle Stone Age of Nigeria in its West African Context

Table 10. Ningi Hills. Artefact inventory. Tools

Jigawa

Tabela

4

2

Levallois blades

2

1

Levallois points

1

Levallois flakes

and 15 flake/blades) (Figure 133.1). The stone walls shown in the photographs provide evidence of later prehistoric settlement which must have been quite significant. In the eastern part of the complex, the site of Old Ningi (Figure 128) was historically very important (Patton, 1975). It was not abandoned till the 1930s. Within the area of the old town, which is very rocky, there were a number of shallow diggings, and a handful of artefacts was recovered (1 initial or preparatory core, 1 Levallois core, and 1 flake) (Figure 133.2).

Sidescrapers Straight

4

Convex

1

Double

1

2

On ventral surface

2

Bifacial

1

Endscrapers

1

Various

3

Total tools

20

6

3

1

1

As mentioned, more extensive workings and a larger amount of material was discovered at Jigawa, on the northern side of the Hills (11° 7’N, 9° 25’E), and at Tabela, on the southern side, east of Old Ningi (10° 57’N, 9° 27’ E).

Cores Initial/preparatory Irregular

3

Disc

14

Levallois flake/blade

1

1-platform flake/blade

3

5

Total cores

24

12

It was discovered that active tin mining operations were being carried out at Jigawa by the D.B. Zang company, who maintained four leases along the Bakinpa river south of the village. The most recently worked paddock was situated in the northernmost lease 17005 nearest to the village. South of lease 17005 is lease 18130. In the northern part of this lease there were extensive abandoned workings, and in the southern part was found a recently operated pit in the bed of the Bakinpa river, which was absolutely dry at the time of our visit (Figure 129).

6

Striking platforms (tools, flakes, blades) Plain

130

39

Cortical

14

2

Facetted

21

10

Dihedral

13

4

Removed

18

9

Total striking platforms

196

64

Figure 126. Koluki, 2 May 1978.

78

The Middle Stone Age of Nigeria

Figure 127. Koluki, 2 May 1978.

Figure 128. Old Ningi, 9 May 1978.

79

The Middle Stone Age of Nigeria in its West African Context

Figure 129. Jigawa, 8 May 1978.

Directly south of lease 18130 is the area known as Kura, formerly the site of the Kura mining camp, now derelict. The map published in 1923 also recorded a mining camp at Suma, in the hills between Jigawa and Tabela, further up the Bakinpa river. This was now apparently being worked only during the rainy season, and we were not able to visit it. In the Jigawa area, it became immediately apparent that palaeolithic type artefacts were present, and in view of the total composition of the finds (Table 10) they can be regarded as MSA. In raw material, state of preservation, technology and typology, they are in fact closely comparable to the material already known from further south. Taking the finds from all locations together, there are 20 tools, 24 cores, and 196 blanks with identifiable striking platforms. Among the tools, there is quite a noticeable Levallois element, apart from numerous sidescrapers, and disc cores are both characteristic and frequent (Figures 130-132, 133.3-6, 136). There was found to be no current tin mining going on in Tabela, but formerly the Nigerian Tin and Exploration Company was active in the area, and there were large abandoned workings south-east of the village, in the vicinity of the Gandu river. A collection of artefacts was made from heaps of stones by these abandoned workings (Table 10), with a total of 6 tools, 12 cores, and 64 blanks with identifiable striking platforms. These finds are of a generally MSA character, with a number of Levallois flakes and blades, and disc cores (Figures 134 and 135). The material is in general somewhat rounded and patinated and made of rhyolite. There are obviously good possibilities for establishing a stratigraphic control at Tabela. Excellent sections are visible along the Gandu river bank. Beneath topsoil, there is a series of silts, and below these two tin-washes separated by a distinctive red earth layer. Two flakes were recovered in situ in the upper tin-wash, one of which is illustrated at Figure 135.3. Not much, but enough to indicate that the source of the industry could be securely established. Clearly therefore a great deal more could be done here, as at Jigawa and elsewhere in the Ningi Hills, to investigate the prehistoric settlement of the area.

80

The Middle Stone Age of Nigeria

Figure 130. Jigawa. 1, Levallois flake/blade core. 2 and 5, disc cores. 3 and 6, 1-platform flake/blade cores. 4, flake/blade.

Figure 131. Jigawa. 1 and 2, disc cores.

Figure 132. Jigawa. 1, disc core. 2 and 4, Levallois flakes. 3, Levallois blade. 5, flake. 6 and 7, blades.

Figure 133. Koluki. 1, 1-platform flake/blade core. Tsofon Ningi. 2, Levallois flake/blade core. Jigawa. 3 and 6, flakes. 4, straight sidescraper on Levallois flake/blade. 5, bifacial sidescraper.

81

The Middle Stone Age of Nigeria in its West African Context

Figure 134. Tabela. 1 and 4, 1-platform flake/blade cores. 2, disc core. 3, Levallois flake/blade core. 5, Levallois blade.

Figure 135. Tabela. 1, convex sidescraper. 2, disc core. 3, flake, in situ upper tinwash. 4, endscraper. 5, Levallois flake.

Banke Banke village (8° 50’N, 8° 30’E) is just south of the Banke Hills (Figure 137) and has provided a base for tin-mining operations conducted by the Bisichi Jantar company since the early 1970s. At the time of these investigations, there were four leases along the course of the Marwan Karini river to the west and south of the village (9303, 14825, 13468, and 9678) (Figure 138). Extensive work was going on only in lease 13468 (Figure 139). A series of short visits was made to the site, on the following occasions: (1) 21 July 1976, (2) 29 Figure 136. Jigawa, disc cores. December 1976, (3) 27 March 1977, (4) 28 May-6 June 1977, (5) 8-9 April 1978. The configuration of the paddock was continually changing during this time, as indicated in Figure 139. In the period between July 1976 and June 1977, the paddock was divided by the company into two parts (N.W. and S.E.) separated by a baulk (Figure 140). The N.W. part was already mostly flooded, whereas the S.E. part was being actively exploited by ‘No. 21 plant’. The walls of the paddock moved substantially over the period in question, as indicated on the plan. Later, the company’s operations moved to a third (S.W.) part of the paddock, referred to as ‘No. 21 plant extension’, and this was the one visited in April 1978 (Figure 141). The situation was briefly summarised in two preliminary reports (Allsworth-Jones, 1981 and 1987). All visits were facilitated by the company’s area manager Mr A.R. Busari. 82

The Middle Stone Age of Nigeria

Figure 137. Banke, general view looking north, 27 March 1977.

Figure 138. Banke, map showing position of leases on river Marwan Karini.

83

The Middle Stone Age of Nigeria in its West African Context

Figure 139. Banke, plan of paddocks in lease 13468 over the period 1976-1978, with sections 1-3 and location of samples for radiocarbon dating.

On the occasion of the first visit (July 1976) Mr Busari explained that there were two tin-washes at a depth between about 15 and 18 feet (4.5-5.5 metres). Large carbonised pieces of wood were associated with the upper tin-wash. Some unstratified artefacts (including a quartz handaxe, Figure 142, on the left) were collected. Two samples of wood were taken for radiocarbon dating, as indicated on the plan: (1) selected from pieces on the floor of the S.E. paddock dislodged during the previous 10 days, and (2) from an in situ location in the upper tin-wash in the wall of the largely flooded N.W. paddock. On the occasion of the second visit (December 1976) a clay horizon which commonly occurs at the site was observed in the wall of the S.E. paddock, above the upper tin-wash, carbonised pieces of wood frequently being found 84

The Middle Stone Age of Nigeria

Figure 140. Banke, NW and SE paddocks, 29 December 1976.

Figure 141. Banke, SW extension paddock, 8 April 1978.

85

The Middle Stone Age of Nigeria in its West African Context

at the junction of the two (Figure 143). Further artefacts, including another handaxe and two ground stone axes, were collected. In March 1977, the S.E. paddock was being actively exploited, as can be seen in Figure 144. Removal of the sediments using a water pump, while no doubt more convenient for the miners, is inimical for the determination of the stratigraphic provenance of any artefacts recovered, which is particularly serious at Banke in view of the diverse nature of the finds. The general stratigraphic situation can be seen in Figure 145, where the position of the lower tin-wash is indicated by Mr Busari. Two further samples were collected for radiocarbon dating: (3) a piece taken from a large carbonised tree trunk found in situ in the upper tin-wash, at the point indicated on the plan, and (4) a further such piece from the tree trunk shown in Figure 146, fallen at the time of our visit, but recovered a short time before from grey clay above the upper tin-wash. On the fourth visit to the site (May-June 1977) two sections were drawn on the east wall of the S.E paddock, at the positions indicated on the plan (Figure 147; Table 11). Section 1, slightly over 6 metres thick, contains a full record of the stratigraphy at that point. Section 2, to the south of it, was evidently disturbed in its upper part by former mining works, hence only the lower part was investigated in detail, a thickness of about 4 metres. The reddish yellow gravelly sand, a portion of which survives at the top of this section, evidently corresponds to layer 7 in section 1, so there is a link between the two. The base of section 2 (as

Figure 142. Banke, quartz handaxe (July 1976) and perforated stone artefact (June 1977).

Figure 143. Banke, SE paddock 29 December 1976, clay horizon above the upper tin-wash.

86

The Middle Stone Age of Nigeria

Figure 144. Banke, SE paddock 27 March 1977, water pump in action.

Figure 145. Banke, SE paddock 27 March 1977, clay horizon and lower tin-wash.

87

The Middle Stone Age of Nigeria in its West African Context

Figure 146. Banke, SE paddock 27 March 1977, tree trunk fallen from clay horizon above the upper tin-wash.

shown to scale in the diagram) is about 1 metre lower than section 1. In both sections, there are traces of both upper and lower tin-wash, though (as typical for the site in general) they are not identical. The lower tin-wash is coarser than the upper and according to Mr Busari is the principal source of cassiterite, as well as probably the bulk of the artefacts. In section 2, but not in section 1, the upper tin-wash is surmounted by a horizon of dark grey clay, similar to the one observed in the S.E. paddock on our second visit to the site in December 1976 (Figure 143). Mr Busari stated that a stratigraphic situation similar to this was encountered elsewhere on the site, although the thickness of the two tin-washes is not constant and sometimes only one is present. He also confirmed that (as we had already observed) carbonised wood occurred either in the upper tin-wash or immediately above it. Artefacts recovered on this occasion (none in situ) included a perforated stone (Figure 142, on the right) typical for the LSA in West Africa. The results available up to this point were reported to the X INQUA Congress meeting in Birmingham in that year (Allsworth-Jones, Catling, Switsur, 1977). On the last visit to the site (April 1978) it was found that the company’s operations had moved to a third (S.W.) part of the paddock, referred to as the ‘No. 21 plant extension’ area (Figure 141). Three samples of carbonised wood collected from the floor of the paddock were said by the overseer (Mr Dan Fulani) to have come from a position immediately above the upper tin-wash. A fourth piece of carbonised wood was observed in situ in the south wall of the paddock in this position and at the same level as a layer of dark greyish brown clay (Figure 148). This formed our radiocarbon dating sample (5). Section 3 was drawn at this point, as shown on the plan (Figures 139 and 147). As in sections 1 and 2, there were two superimposed tin-washes, but in this case they were separated from each other by a layer of reddish yellow sand and silt. A further collection of artefacts was made, all of them coming from this paddock as a result of the mining operations. The collection was heavily biased towards handaxes and cleavers, of which 9 were recovered (Figure 149). They are well made and highly characteristic, mostly on side-struck flakes, not all together unlike some examples from Mai-idon-Toro. 88

Figure 147. Banke, sections 1 and 2 (SE paddock) and 3 (SW extension paddock).

The Middle Stone Age of Nigeria

89

The Middle Stone Age of Nigeria in its West African Context

Table 11. Banke lease 13468. Stratigraphy, sections 1-3. Section 1 Layer No. 1 2 3 4 5 6 7 8 9 10 11 12 13

Description Light brown topsoil Brownish yellow sand and silt Pinkish white sand and silt Pinkish grey interbedded sands, silts, gravels Pinkish white laminated sands, silts, gravels Light grey sands, silts, and clays, with banded grey sands, silts, and clays Reddish yellow gravel and sand UT Reddish yellow sharp gravelly sand UT Reddish yellow medium sand UT Pinkish grey intercalated gravelly sand UT Pinkish grey fine sand LT Light brown coarse sandy gravel LT Pinkish grey sand

Munsell colour 7.5 YR 6/4 10 YR 7/6 7.5 YR 8/2 7.5 YR 7/2 7.5 YR 8/2 10 YR 7/1 + 2.5 YR 5/0 5 YR 7/8 5 YR 6/8 5 YR 6/8 7.5 YR 6/2 7.5 YR 6/2 7.5 YR 6/4 7.5 YR 6/2

Description Dark grey clay below reddish yellow gravelly sand (= section 1.7) UT Light brownish grey silty sand + light brownish grey sandy silt LT Light brownish grey silt and gravel

Munsell colour 10 YR 4/1 5 YR 7/8 10 YR 6/2 + 10 YR 6/1 10 YR 6/2

Description Dark greyish brown clay UT Pinkish grey sand Reddish yellow sand and silt LT Yellowish brown sand and gravel

Munsell colour 10 YR 4/2 7.5 YR 6/2 7.5 YR 6/8 10 YR 5/4

Section 2 Layer No. 1 2 3 Section 3 Layer No. 1 2 3 4

Figure 148. Banke, section 3 (SW extension paddock) with tree trunk above the upper tin-wash.

90

The Middle Stone Age of Nigeria

A complete record of all the artefacts recovered during our five visits to the site is at Table 12, with drawn illustrations at Figures 150-153. Typologically, 10 tools can be attributed to the MSA, 7 to the LSA, and 10 to the ESA. Such a pronounced mixture was not encountered at any other of our sites, but it is a salutary warning about what can happen when dealing with an industrially worked location of this kind. Attention has already been drawn to the use of a water pump which clearly would not permit the identification of provenance. It cannot be determined whether there was any kind of difference between the artefact content of the upper versus the lower tin-wash, the latter of which according to Mr Busari will have produced the majority of the material. Carbonised wood samples (1) and (2) recovered in 1976 from the upper tin-wash in the S.E. and the N.W. paddocks respectively produced radiocarbon dates of 4240±90 and 4130±90 BP. Sample (3) obtained in 1977 from another location in the S.E. paddock, also in the upper tin-wash, produced a radiocarbon date of 4710±65 BP [Q-2029]. On the other hand, the tree trunk (5) sampled in 1978 in section 3 in the S.W. paddock (Figures 147 and 148), which was clearly above the upper tin-

Figure 149. Banke, two cleavers found in April 1978.

Figure 150. Banke. 1 and 4, 1-platform flake/blade cores. 2 and 3, disc cores. 5, blade with marginal retouch/sidescraper.

Figure 151. Banke. 1, disc core. 2, endscraper. 3, ground stone axe in course of manufacture. 4, broken half of ground stone axe in course of manufacture. 5, Levallois blade. 6, notch/denticulate. 7, transverse sidescraper.

91

The Middle Stone Age of Nigeria in its West African Context

Table 12. Banke lease 13468. Artefact inventory. MSA Tools Levallois flakes

1

Sidescrapers Straight

1

Convex

4

Canted

1

Transverse

1

Denticulates

2

Total tools

10

LSA Perforated stones

2

Ground stone axes

5

Total tools

7

ESA Handaxes

7

Cleavers

3

Total tools

10

Cores Figure 152. Banke. 1, apparent roughout for ground stone axe. 2, 1-platform initial core. 3, perforated stone artefact.

Initial/preparatory

1

Irregular

1

Disc

3

Levallois flake/blade

1

1-platform flake/blade

3

Total cores

9

Striking platforms

Figure 153. Banke. Cleaver.

(tools, flakes, blades)

Plain

28

Cortical

1

Facetted

9

Dihedral

11

Removed

3

Total striking platforms

52

wash, produced a radiocarbon date of 2400±50 BP [Q-3034], somewhat younger than the others, which would make sense. Mr Busari said that dates of this age did not surprise him, on the basis of his own observations, since both the mineral and the stones in the lower tin-wash were sharp edged, and the upper tin-wash (in his words) had never had ‘time to settle’. The dates are of course also comparable to the one long known from Zenabi of 5440±100 BP. The prominence of thick clay horizons at both sites is another link between them, as is the frequent occurrence of large pieces of carbonised wood. If these dates can be accepted at face value, it seems that at some time in the Holocene in this area an 92

The Middle Stone Age of Nigeria

erosional event of some violence took place, more or less coincident with the emplacement of (at least) the upper tin-wash, which displaced large quantities of organic material. It is not at all clear how at this site, as distinct from the others we have studied, there was such a marked admixture of material of different ages, in particular artefacts which are strongly reminiscent of the Acheulean. It was in order to escape the ambiguities inherent in alluvial mining sites such as these that attention was devoted to another site in the study area, Saminaka, which it was hoped would provide a more certain provenance for the archaeological material. Saminaka Saminaka (10° 25’ N, 8° 42’ E) is a small town on the Jos-Kaduna road, where it crosses the river Rahama. Soper (1965) noticed that material of MSA character was weathering out in erosion gullies above the river north east of the town, and he thought that it would repay excavation, as a possible in situ settlement. A further visit to the site in March 1977 confirmed this conclusion (Allsworth-Jones, 1981), when a wide scatter of artefacts was similarly noticed in this area. The collection made on that occasion included 2 tools, 4 cores, and 5 flakes. The majority of the artefacts were made from rough mottled brown-white quartz, but there were also a few of whitish clear quartz, and a couple of other materials including rhyolite. This raised the question of possible long-range connections to sites in the tin-mining areas, but it was hoped that a site which had not been subject to alluvial reworking would afford more secure contextual information. A photograph of the site as it appeared in April 1982 is at Figure 154. The challenge was taken up by O.A. Opadeji, who investigated the area in 1996-1998 as part of a PhD subsequently published in Frankfurt (Opadeji, 2001). The author reported in detail on three or four locations, about 2 km north east of the town and within some hundreds of metres of each other, as indicated on the map at Figure 155.

Figure 154. Saminaka, April 1982.

93

The Middle Stone Age of Nigeria in its West African Context

Figure 155. Saminaka map (after Opadeji 2001, Fig. 2).

Figure 156. Saminaka sketch of cross-section (after Opadeji 2001, Fig. 13).

94

The Middle Stone Age of Nigeria

All are on the west bank of the river. Excavations were conducted at Saminaka Stone Age (SSA) site 1, with two trenches on the western slope and the summit of a rise overlooking the valley, and at SSA site 2 near the bank of the river itself, as shown in the sketched cross-section of the topography at Figure 156. In addition, a one metre wide cutting was made in the side of one of the eastward trending gullies north of SSA 1, where some artefacts had been observed eroding out of the gully wall. SSA 2 was in an area much disturbed by farmers, and was also subject to seasonal flooding, whereas SSA 1 was largely undisturbed, due to the presence of large granite boulders and an infertile soil. It produced the majority of the material excavated at the site. SSA 1 Trench 1 on the western slope covered an area of 4 m2 and was taken to a depth of 60 cm. Layer 2, a yellowish red horizon between 8 and 30 cm thick, constituted the main, and most prolific, cultural layer at the site. SSA 1 Trench 2, a test pit 1 m2 in extent at the top of the rise, was taken to a depth of 50 cm, but produced relatively little material. The same was true for SSA 2, also a 1 m2 test pit, taken to a depth of 40 cm, some 13 metres from the riverbank. The excavated gully section was cut back for 50 cm and had a maximum depth of 1 metre. As the author emphasises, there is no reason to think that the archaeological material excavated on the rise was redeposited, whatever may have been the situation on the floodplain. It is unfortunate that the only piece of charcoal discovered was evidently intrusive, since it produced a radiocarbon date of no more than 200±35 BP (Utc-5748). It was also not possible to carry out any successful refitting of material. But that does not invalidate the fact that this turned out to be, as had been hoped, an in situ non-alluvial MSA site, yielding stratified archaeological material in a good context. The artefact inventory is summarised in Table 13, based on figures provided by the author (Opadeji, 2001, Tables 3-6; contrary to his statement on page 49, I have not been able to study the material myself). By chunks are meant stone pieces more than 2 cm in maximum dimension, with no evidence of human interference. Chips are small removals from struck stones, less than 2 cm in maximum dimension, presumed to be the by-product of core preparation. It will be noted that in SSA 1 Trench 1 quartz amounts to 3715 out of 3810 pieces, or 97.5% the total, the great majority consisting of chips and chunks. The proportions are reversed if tools and cores alone are considered, since non-quartz materials amount to 30 out of 45 pieces, or 66.7% of the total. The figures for the remaining trenches are not divided by raw material, but the recorded non-quartz items amount to no more than 15 pieces out of a total of 382. The author’s detailed consideration of technology and typology is confined to SSA 1 Trench 1, Table 13. Saminaka Stone Age (SSA) Artefact Inventory.

SSA 1 Trench 1

flakes

tools

cores

chunks

chips

total

quartz

11

1

-

439

251

702

non-quartz

5

-

-

-

1

6

all

16

1

-

439

252

708

quartz

79

8

6

1434

1486

3013

non-quartz

46

11

19

8

5

89

all

125

19

25

1442

1491

3102

total

141

20

25

1881

1743

3810

Others

flakes

tools

cores

chunks

chips

total

SSA 1 Trench 2

16

-

5

69

53

143

SSA 2

1

-

-

80

3

84

Gully 1 Cutting

3

2

2

132

16

155

total

20

2

7

281

72

382

surface

excavated

95

The Middle Stone Age of Nigeria in its West African Context

but for unknown reasons the totals given for flakes, tools, and cores are slightly at variance with the above (Opadeji, 2001, Tables 7-11). They amount to flakes 143, tools 28, and cores 27. Apart from 4 Levallois flakes, the tools include 3 notches, 16 scrapers, 1 denticulate, and 4 points. The cores include 1 Levallois, 8 discoid, and 18 others (globular, irregular, pyramidal, and rectangular). 74 (51.7%) of the identified striking platforms are plain whereas 21 (14.7%) are facetted. 121 (84.6%) of the flakes have no traces of cortex, which is taken as an indication that they were initially worked elsewhere prior to their transport to this site. The mean dimensions for flakes and cores were calculated as follows (in cm). Flakes: length 3.0, width 2.16, thickness 0.74. Cores: length 5.5, width 4.64, thickness 2.96. On the basis of limited measured samples, it was noted that these dimensions are less than those of their counterparts at Banke, Tibchi, and Mai Lumba. The same would apply to Zenabi, since as noted here the modal values for length, width, and thickness for 356 flakes given by Anozie (1975) amount to 7.0-7.9, 4.0-4.9, and 1.01.9 cm respectively, whereas the corresponding mean figures for the length, width, and thickness of 122 cores calculated by Steve Daniels amount to 9.0, 7.2. and 4.8 cm. The relatively small size of the artefacts from Saminaka is attributed to the greater intensity of use at the site, which in turn is taken as a reflection of its distance from the relevant raw material sources. This is backed up by a petrological analysis of 48 artefacts undertaken by Professor Brey of the Geology Department of Frankfurt University (Opadeji, 2001, Table 32). Of these, 34 were varieties of rhyolite, 12 were identified as quartz, and 2 as granite. The rhyolite was present in two varieties, which by comparison to published sources (Bain 1934, Jacobson et al. 1977) could be traced to the ring complexes at Kudaru, and Banke or Liruei. These would have been more accessible than Saiya-Shokobo because of the obstacle posed by the river Karami and its tributaries (Figure 14). 27 Kudaru type artefacts accounted for most of the total, whereas 7 could be traced to Banke or Liruei. Kudaru is 40 km from Saminaka, so the artefacts must have been transported for at least this distance, probably, it is suggested, in the form of pre-formed cores (Opadeji, 2001: 96-97). It should be noted however that this interpretation applies only to the rhyolite component, since quartz, the other raw material used at the site, was available in abundance locally. In his overall interpretation, Opadeji suggests that this site functioned as a campsite, not a factory, and that this function was facilitated by its situation along the banks of the Rahama river. It surely formed part of the overall settlement pattern which elsewhere has been noted at the tin-mining sites, but its excavation has demonstrated that in situ MSA material in non-alluvial contexts certainly does exist in Northern Nigeria. The details given, and the amount of material recovered, are not sufficient to permit a detailed comparison to the other sites investigated in the study area, but the proportions of flakes, tools and cores, in comparison to chips and chunks are of course much more representative of an undisturbed site than any of the alluvial-based collections. While radiocarbon dating did not succeed in this instance, there is perhaps no reason to think that other dating methods might not do better at this site in future. The above account concludes our survey of the sites investigated in the study area north of the Jos Plateau. The other sites to be mentioned, in the vicinity of the Plateau, are no less significant, particularly from a historical angle, and since they are all alluvial sites connected with tin mining, they are essentially similar to the majority of those already discussed. Nok Nok (9° 30’ N, 8° 10’ E) is south-west of the Jos Plateau and west of the Kagoro Hills. It is well known as the site where the Nok iron age culture and its characteristic art style was first recognised and after which it is named. Less well known is that fact that Acheulean and MSA artefacts were also found at the site, and recognised as such by Bernard Fagg (1956, 1977). Fagg visited the site on numerous occasions from 1944 onwards. His sketch map of the valley with the westward flowing Nok river is at Figure 157 (Fagg, 1956, Fig. 9) and his ‘hypothetical section’ is at Figure 158 (Fagg, 1956, Fig. 8).

96

The Middle Stone Age of Nigeria

Figure 157. Sketch map of the Nok Valley (after Fagg 1956, Fig. 9).

97

Figure 158. Hypothetical section of the Nok Valley (after Fagg 1956, Fig. 8).

The Middle Stone Age of Nigeria in its West African Context

98

The Middle Stone Age of Nigeria

He stated that ‘three distinct and important archaeological periods were preserved in the river gravels and were being brought to light by mining operations’ (Fagg, 1956: 209). The Big Paddock (BP) is shown on the map as having produced a Middle Stone Age industry, unrolled. This is described in more detail as ‘a quartz blade industry in which the prepared platform technique is used, though it is not highly developed’ (Fagg, 1956: 214). This material occurred in tin-bearing gravels, overlain by sandy loam, and then by an upper tin-wash (Fagg, 1956, Fig. 11, a photograph taken in May 1948). In Fagg’s section a terrace containing the MSA blade industry ‘Q’ is hypothetically shown as backed up against a post-Acheulean erosion scarp, and preceding the alluvial deposits ‘X’ with the Nok Culture remains. There are ??? indicating the uncertain position of ‘Q’ and it is stated that the MSA industry is ‘thought to exist but not yet clearly identified’. Bond visited the site with Fagg in 1948 and provided a ‘combined succession’ which corresponds to Fagg’s threefold division as follows (Bond, 1956: 193-194). (1) False-bedded sands stained yellow (12 feet thick = 3.7 m), followed by blue clay with sand lenses (6 feet thick = 1.8 m), and an upper tin-wash in ‘impersistent gravel’ (up to 1 foot thick = 0.3 m). The blue clay and the upper tin-wash were associated with pottery and figurines (corresponding to Fagg’s ‘X’). There was then a (?) erosion surface. (2) Greenish sands, sterile (12 feet thick = 3.7 m), followed by a middle tin-wash in gravel (1.5 feet thick = 0.45 m), with Middle Stone Age. This was underlain by another erosion surface marked by ‘lateritisation’. (3) Sands and gravels (up to 40 feet = 12 m) maximum, with a lower tin-wash at the base. This was characterised as an Acheulean horizon (corresponding to Fagg’s ‘B’) above a further erosion surface and then bedrock (Fagg’s ‘A’). Bond emphasised that phase (3) had been detected in the ‘Loto’ site only (on a tributary of the Nok river), whereas phases (1) and (2) had been located only in the Big Paddock. He stated that ‘the base of the lateritised series is not reached in the Big Paddock, where the lowest tin-wash gravel is that which contains Middle Stone Age flake tools’ (Bond, 1956: 194). This corresponds to the situation recorded in Fagg’s photograph mentioned above. In his general summary, Bond (1956: 198) ascribed phase (2) at Nok to the Gamblian Pluvial. Material for radiocarbon dating was obtained at a somewhat later date, in 1951 and 1956, and is said to have come from the Main Paddock, which presumably is the same as the Big Paddock (Barendsen et al., 1957). Three dates on wood labelled Nok G, D, and H (Y-474 1750±50 BP, Y-142-4 2873±70 BP, and Y-475 4060±140 BP) came from clay and sand overlying the ‘basal gravel of the youngest alluvial body’. This provided an acceptable age estimate for the pottery and figurines (Fagg’s ‘X’), the ‘most probable’ date considered to be Y-142-4. Carbonised wood from the ‘oldest alluvium, associated with Acheulean artefacts’ (Nok E, Y-142-8) produced an infinite date of >39,000 BP. If the Main and Big Paddocks are indeed one and the same thing, this indicates that in the years after 1948 the deepest deposits were reached at that site. In addition, there is a date labelled Nok C, described as having been obtained on ‘wood from basal tin-bearing gravel, at (the) main figurine horizon, overlying (an) erosion surface cut in older alluvium’. The first date obtained was Y-142-3 5490±85 BP. Another portion of the same specimen was dated separately with the result Y-1423’ 5660±90 BP. The average age published for Nok C therefore came to 5575±65 BP. It is a date which has commonly been quoted as applying to the MSA in Nigeria. In the commentary accompanying the dates, it is said that the sample ‘must have been redeposited from older sediments’, this argument being supported by a comparison to Zenabi where a date of 5440±100 BP was also obtained on an ‘older tin-bearing alluvium’. At best therefore this attribution is somewhat hypothetical. As at Zenabi, the laboratory suggested that the period concerned should be the Makalian rather than the Gamblian Pluvial (Barendsen et al., 1957). The material from Nok stored in the Jos Museum and attributed to the MSA was studied by Robert Soper (1965: 190-191). He listed 2069 artefacts, with percentage numbers, from which it is possible to derive actual figures. On this basis, it can be said that the collection included 83 tools, 158 cores, and 99

The Middle Stone Age of Nigeria in its West African Context

Figure 159. Nok valley, 27 March 1976.

Figure 160. Nok abandoned paddock, 27 March 1976.

100

The Middle Stone Age of Nigeria

1828 flakes and flake-blades. Of the cores, 95 or 60% were said to be ‘prepared’, although only 766 or 42% of the flakes had prepared striking platforms. A much smaller number of objects coming from the Big Paddock alone were studied by Anozie (1975: 100-109). Apart from two quartz hammerstones, the total given by him amounts to 827 artefacts, including 42 tools, 39 cores, and 746 flakes. According to him, 16 of the cores could be classified as Levallois. The raw material employed consisted of both quartz and rhyolite. Both authors are broadly in agreement about the nature of the tools at the site. Sidescrapers were predominant. There were some unifacial and bifacial points, and also a few small bifaces, although in Soper’s view, the latter could possibly be an admixture from the Acheulean. Two photographs of the Nok valley and an abandoned paddock (the latter with a characteristic stepped configuration) are at Figures 159 and 160, showing the way they looked in March 1976. Pingell Pingell (10° 20’ N, 9° 05’ E) in the Dagga Allah hills east of Jos was mentioned by Soper (1965) as having produced MSA type material, but is best known as an Acheulean site. It was another tin-mining location, and according to him the archaeological material was collected in 1955-1956 by Mr R.J.P. Hiscock, one of the managers at the site. Soper describes the deposits as a series of buried river channels in the vicinity of the Garawa river. The uppermost series consisted of loose bedded sands with semi-carbonised logs (assumed to be quite recent) above a lower series with in situ material contained within a tin-wash or washes. The Acheulean type material was described as abraded, most considerably so, but it occurred ‘in association with fresher artefacts probably assignable to the MSA’. Because of possible admixture, Soper confined his analysis to 233 of the larger tools which he thought could confidently be regarded as Acheulean (Soper, 1965, Table II). Anozie (1975) also studied some of this material kept in the Jos museum, again treating most of it as Acheulean. He located seven drawn sections in the archives, dating to 1956, so presumably all were produced by Hiscock. Of these, three showed both an upper and a lower

Figure 161. Pingell, April 1976.

101

The Middle Stone Age of Nigeria in its West African Context

Figure 162. Pingell, April 1976.

tin-wash (Anozie, 1975, plates 9-11). A visit to the site in April 1976 revealed that the old workings were completely flooded (Allsworth-Jones, 1981; Figures 161 and 162). Some more recent small-scale works were examined, and they produced quite numerous artefacts of MSA type (n=33), which generally were not rolled (Allsworth-Jones, 1981, Table 5, Fig. 5.9 and 10). They included Levallois flakes and sidescrapers, as also recorded by Anozie. At that point there was no sign of any Acheulean. Rop Rop (9° 30’ N, 8° 55’ E) south of Jos on the Plateau is best known as the site of a Late Stone Age rock shelter excavated by Bernard Fagg and his successors. But in the vicinity, as recorded by Soper (1965), ‘a small hand-worked mining paddock produced an industry made in chalcedony with a few quartz pieces, from the tin-wash of a channel cut into a relatively recent basalt flow’. The industry included Acheulean type material but also some ‘flakes and cores made by the prepared core technique’. Soper suggested that it could represent a ‘transitional stage’ between the Acheulean and the MSA, although mechanical mixing could not be excluded. The overlying channel contained (as so often the case) ‘water-logged wood’, and Soper hoped to obtain a radiocarbon date for it, but it seems that this never happened. No doubt there may have been other occurrences of this sort. Fagg (1956) noted that some of the first stone implements to be identified on the Plateau, including handaxes as well as ‘typical Levallois flakes and tortoise cores’, were found in the Delimi river deposits in Jos itself, but those mines were already ‘worked out and no longer available for study’. Soper (1965) listed 12 other occurrences in Northern Nigeria, observed by him in the course of his reconnaissance, which could possibly be the source of MSA type material, but (apart from Saminaka) they have not been further investigated, so one cannot be sure of the situation.

102

The Middle Stone Age of Nigeria

Summary The above account of sites north of the Jos Plateau and in its vicinity relies for the most part on material recovered in alluvial contexts, revealed by industrial workings. There is nothing inherently unusual in this situation, which also occurs elsewhere in the old world, sometimes in connection with sites which are really well known. But it does pose inevitable limitations, particularly in relation to possible admixture, which is most clearly seen at Banke. At Nok and Pingell, there are indications of a stratigraphically distinct Acheulean horizon, but for the most part these sites have produced material which is predominantly or entirely MSA in character. It is abundant and varied, mostly reliant on local raw materials from the ring complexes characteristic for the area. The existence of this raw material source was clearly well known to the makers and no doubt encouraged settlement. The evidence from Saminaka demonstrates that they were prepared to transport such material for tens of kilometres and use it in preference to quartz which was near at hand. The degree of variation in the MSA industries, most clearly shown in the contrast between Mai Lumba on the one hand and Zenabi, Tibchi, and Yelwa on the other suggests that this was not entirely random, and reflects significant distinctions not dependent on raw materials or functional requirements. Unfortunately it cannot be determined whether these differences owe anything to chronological factors. So far, we only have radiocarbon dating for some of these sites, notably Zenabi, Nok, and Banke. All these dates were obtained on wood charcoal, sometimes from large logs embedded in the deposits, and they are all Holocene in age. In almost all cases, we can observe that at the alluvial sites there were two tin washes, which surely must reflect erosive processes (as suggested by Thorp at Liruei) which dislodged the archaeological material. The tin washes are not identical, as can be observed most clearly at Tibchi, so whether there were in fact more than two erosive phases cannot for the moment be determined. Obviously radiocarbon dating itself has limitations, and in these days other methods could be used to directly date the deposits. They may be older, even considerably older, than the radiocarbon dates suggest. Nonetheless, that would not alter the fact that the archaeological material has been displaced, so any dates achieved would relate to the displacement and not to the artefacts themselves. With all these limitations, the expressive nature of the MSA in this area cannot be ignored, and it surely forms a significant part of the overall pattern existing in West Africa at the time. Sangoan The Sangoan is somewhat distinct from what one might refer to as a normal MSA, but occurrences classified in this way have also been claimed to exist in northern Nigeria. Oliver Davies (1957) was the first to draw attention to the existence of what he termed Sangoan sites in Northern Nigeria. During his wide-ranging reconnaissance of this region in the early 1960’s, Robert Soper (1965) revisited certain of Davies’s sites, and discovered others, which he amalgamated together into a Southern, a Northern, and a Western group. The Southern group is concentrated in the area of Abuja, Keffi, and Nassarawa, the Northern group along the tributaries of the Sokoto river, and the Western group along the Niger river in the vicinity of Bussa and Yelwa. According to Soper, there were 19 sites in his Southern group, 15 in the Northern group, and 21 in the Western group. He named and gave a number to only a few of these sites, and not all of the others appeared as symbols on his map of the region, but his map does allow their broad areas of occurrence to be discerned. The approximate limits of the three groups, and the names of the principal sites, are indicated here at Figure 163, based on the information provided by Soper. Davies had drawn attention to the fact that some of his sites appeared to be concentrated along the line of contact between the basement complex of Northern Nigeria and the Cretaceous sediments of the Niger and Benue valleys, and this concentration is reflected in the geography of the Southern group. Most of Soper’s collection consists of surface material, the number of stratified occurrences being very few. He summarised the characteristics of the Southern and Northern groups in a table, reproduced here as Table 14. Considering the large number of sites from which this material is drawn, it has no more than an indicative value. Prominent among the tools recognised are scrapers, picks, choppers, and handaxes,

103

The Middle Stone Age of Nigeria in its West African Context

Figure 163. Distribution of Sangoan sites in Northern Nigeria (adapted from Soper, 1965, 184-185). Table 14. Nigerian Sangoan assemblages (after Soper, 1965, Table III). Class

Southern Group

but also an artefact type which he characterised as a ‘push-plane’. He described this as a ‘roughly circular humped scraper with a flat base and almost vertical working-edge’ (Soper, 1965: 184). To judge from the illustration (Figure 164.1) and the description, this category of artefact is not far from the ‘limaces’ identified by this author as characteristic of some Nigerian MSA occurrences. Also illustrated are handaxes and a pick chosen by Soper as characteristic of the groups identified by him (Figure 164.2-4).

Northern Group

Tools Scrapers

13

30

6

9

Picks

25

6

Choppers

17

37

Handaxes

5

14

Miscellaneous

4

5

70

101

Push-planes

Total tools

One handaxe is from Yumu, one of the sites in the Western group. Soper regarded these sites as somewhat different from the others, since the artefacts were made for the most part on small round pebbles, but they could still be regarded as Sangoan in a broad sense. There were no indications of a possible chronological ordering

Cores conical

9

8

other

38

35

Total cores

47

43

Flakes

43

77

Total artefacts

160

221

104

The Middle Stone Age of Nigeria

Figure 164. Sangoan tools (descriptions after Soper, 1965, Figs. E, G, and J). 1, Keffi, quartz push-plane. 2, Yumu, quartz handaxe. 3, Nassarawa, quartz handaxe. 4, Jebba, large quartzite pick.

105

The Middle Stone Age of Nigeria in its West African Context

of the industries discovered, apart from the fact that both at Yumu and at Jebba the artefacts were found on river terraces between 12 and 16 metres above the current level of the Niger. Soper took this as an indication that they corresponded to a high sea-stand which he equated with the last interglacial. Some more details were given about the situation at Jebba (9° 08’ N, 4 50’ E) which must rank among the more interesting of the sites discovered. According to Soper, the most important find spot was about 5 km north-east of the town, where a large borrow pit had exposed an ‘ill-sorted lateritised gravel’ containing rounded quartzite pebbles and granite boulders. The artefacts found included the pick illustrated at Figure 164.4. All were unstratified except for one flake which was in situ 46 cm below the top of the lateritised gravel. The other pieces had laterite concretions adhering to them, and so were considered to have had the same provenance. Nonetheless, the artefacts themselves were all in a fresh condition, hence it was maintained that they had to ‘post-date the actual deposition of the gravel’. Clearly a great deal has happened at Jebba since the early 1960’s, notably the construction of a dam and hydroelectric power plant on the Niger, completed in 1984 (Encyclopaedia Britannica Online, 2017). Much of the construction work was undertaken by Fougerolle Nigeria Ltd., thanks to whom I was able to visit the area in December 1983. One of the company geologists, Mr Arthur Harding, had a good knowledge of the borrow pits on the northern side of the river, not far from the main road, which no doubt are very similar to the one noted by Soper. Two of these pits (‘Localities 3 and 4’) are illustrated at Figures 165-168. Artefacts can be clearly seen in the section of Locality 3, at the junction of two different lithological units. The lower unit is characterised as a coarse gravel, which in Harding’s opinion must have been formed by vigorous water action, subject after deposition to lateritic cementation. The upper unit is a fine sandy silt, with some small sharp edged quartz fragments, and in Harding’s opinion this unit will have been put in place by wind-blown (aeolian) action. Between the two, there is a marked break in deposition, and it was during this break that an occupation marked by the visible artefacts took place. What we can see are the traces of an ancient palaeosurface. No suggestion could be made as to the age of this

Figure 165. Jebba. Locality 4, December 1983.

106

The Middle Stone Age of Nigeria

occupation nor the duration of the hiatus between the two units. It is evident that this independent assessment of the situation agrees closely with Soper’s own observations. Some of the artefacts recovered from the borrow pits on this and other occasions are illustrated at Figure 169-173. On the assumption that this area has escaped further destruction or development, it would be well worthwhile undertaking a more thorough going archaeological and geological investigation of it. Southern Nigeria Relatively speaking, less is currently known about the Palaeolithic succession in this region than in the north. From the point of view of Quaternary geology, the area in the vicinity of Ibadan has however been well studied, and a number of sites have been identified and interpreted in the light of this sequence, as well as at Asejire, somewhat to the north-east.

Figure 166. Jebba. Locality 3, December 1983.

Figure 167. Jebba. Locality 3, artefact horizon indicated by Arthur Harding.

107

The Middle Stone Age of Nigeria in its West African Context

Figure 168. Jebba. Locality 3, artefact horizon.

Figure 169. Jebba. 1-platform quartzite core.

Figure 170. Jebba. sidescraper, pseudo-levallois point, two flakes.

The Bodija Formation The geology and Quaternary history of the Ibadan area is well known, thanks to what has become a standard account, established in the 1970’s by Kevin Burke and ‘Bisi Durotoye (Burke, 1970; Burke and Durotoye, 1971; Burke et al., 1971; Durotoye, 1972 and 1976). Later writers, critical or not, have relied on this framework to a large extent (Kiladejo, 1980; Andah and Ajayi, 1981; Faniran, 1982; Bagodo, 2012). 108

The Middle Stone Age of Nigeria

Figure 171. Jebba. two flake/blade cores, one round scraper.

Figure 172. Jebba. quartzite biface pick, quartz uniface pick.

Figure 173. Jebba. quartzite pebbles/choppers.

109

The Middle Stone Age of Nigeria in its West African Context

Figure 174. Geological Map of Ibadan (after Burke and Durotoye, and Kiladejo, 1980, Fig. 2).

110

The Middle Stone Age of Nigeria

Figure 175. Sites in the Ibadan area (after Bagodo, 2012, Fig. 15).

The basic geological features of the Ibadan area are shown at Figure 174, and the general geographical framework at Figure 175. The bulk of the area lies at elevations between 180 and 210 metres above sea level, but the land rises to above 225 metres on the north-east and the north-west. The most prominent features of the landscape today are elongated quartzite ridges and isolated gneissic inselbergs, both resistant to erosion. The bedrock is formed of crystalline basement rocks attributed to the Pre-Cambrian period, affected by metamorphism and folding. The folded quartzite ridges form nearly parallel highland lines, while the banded gneisses in general form low hills. The major strike of the foliation is from north-west to south-east, crossed almost at right angles by joints of varying sizes. The drainage pattern is mainly controlled by this geological framework, trellis pattern on crystalline basement and dendritic on sedimentary rocks. The major rivers, Ona, Ogunpa, and Ogbere, all trend 111

The Middle Stone Age of Nigeria in its West African Context

from north to south. The great majority of the surface area is formed of what Burke and Durotoye describe as pediments, i.e. gently inclined surfaces formed by erosion, sloping down from residual hills usually towards incised streams. They may also be described as a particular kind of piedmont slope (Tricart, 1972: 199-200, note 1). According to Durotoye (1976, Fig. 4) such pediments may comprise three levels depending on elevation and local circumstances: an upper pediment covered by a laterite crust, an intermediate level overlain either by pediment gravels or pediment wash deposits, and a lower level which may be covered by alluvium. This model of geological evolution is fundamental to their description of the Bodija Formation in the Ibadan area. It has been recognised most clearly at the type site in the Bodija Railway Cutting in the Kongi district (7° 25’ N, 3° 54’ E) (Figure 176) with a second important section along Onikoko Avenue in the Agodi district (Figure 175). The Bodija Railway Cutting drawn section is at Figure 177, and a summary of the stratigraphic column is at Table 15. Three Members have been distinguished. The uppermost Orita Member is described as a pediment wash, but has also been referred to as a so-called termite worm and wash layer (TWW). It consists of predominant sand, silt, and clay. The Agodi Member is described as a pediment gravel, uncemented, with a predominant gravel fraction, and a matrix of sand and silt, separated from the Member below by a minor erosion surface. The distinction between the Orita and Agodi Members can be clearly seen in terms of grain size in the histograms at Figure 178. The Kongi Member is described as a coarse goethite and hematite cemented pediment gravel, separated from the weathered basement complex beneath by a major unconformity. The basement in this area is composed of weathered gneiss with quartz veins. The cementation of the Kongi Member is assumed to have taken place as a result of alternate wetting and drying conditions affecting the gravel in situ after its deposition. Both of the pediment gravels themselves are assumed to have been formed in dry climatic conditions, unlike the most recent pediment wash layer which

Figure 176. Bodija Railway Cutting, 13 December 1979.

112

The Middle Stone Age of Nigeria

Figure 177. Bodija Formation type section in the Bodija Railway Cutting, Ibadan (after Burke and Durotoye, 1971, Fig. 6).

Table 15. Bodija Formation. Stratigraphic column (after Durotoye, 1976, Table 1). Deposit

Climate

Suggested age

Orita Member

Pediment wash

Thickness 0-4 m

Wet

Holocene

Agodi Member

Pediment gravel

0-2 m

Dry

Late Pleistocene

0-2 m

Dry

Middle Pleistocene

Minor unconformity Kongi Member

Cemented pediment gravel

Major unconformity 0-50 m

Weathered basement complex

Pre-Cambrian

is taken to represent a humid period. The assumed chronological equivalents for the three Members are indicated in Table 15. No specific time frame is suggested for the Kongi Member other than a broad bracket in the Middle Pleistocene, but for the Agodi Member - attributed in general to the Late Pleistocene - a more specific proposal has been made. It is suggested that it is equivalent chronologically to the Late Glacial Maximum in the northern hemisphere at approximately 20,000 years ago. In fact the authors (Burke et al., 1971, Figure 2) propose a broad correlation between the uncemented pediment gravel of the Bodija Formation and other deposits in West Africa indicative of dry conditions at that time, in particular the dunes which blocked the mouth of the Senegal river and the Great Erg of Hausaland. This scenario, as mentioned in the Introduction, looks promising, but there are no absolute dates available for the Bodija Formation in the Ibadan area. The authors called upon archaeology to support their hypothesis, since some artefacts described as flakes and pebble tools were found in the Agodi Member (as well as apparently in the Kongi Member, which does not form part of this argument), and Thurstan Shaw expressed the opinion that they should be less than 40,000 years old (Burke and Durotoye 1971: 441; Durotoye 1976: 352). Unfortunately these 113

The Middle Stone Age of Nigeria in its West African Context

Figure 178. Bodija Railway Cutting, particle size distribution for Orita and Agodi members and weathered in situ gneiss (after Durotoye 1972, and Kiladejo, 1980, Fig. 11).

114

The Middle Stone Age of Nigeria

artefacts have never been published, nor has the rationale for the dating been explained, so in the existing state of knowledge this claim cannot be substantiated. Hence for the moment the validity of the model rests upon purely geological criteria.

Figure 179. Sites in the vicinity of Ajibode (after Momin, 1995, Fig. 1, and Bagodo, 2012, Fig. 13).

115

The Middle Stone Age of Nigeria in its West African Context

Ajibode Ajibode (7° 27’ N, 3° 53’ E) is a village north of the University of Ibadan, on the other side of the Odo-Ona or Orogun river, which forms the boundary with it in the Botanic Gardens. The area was acquired by the University in 1985, and from 1992 onwards reconnaissance was conducted along the valley of the Yamoje stream, which flows into the Odo-Ona from the north, by Dr K.N. Momin (Andah and Momin, 1993; Momin, 1995). As a result, 10 surface sites were located on both banks of the stream, where scatters of lithic artefacts were found (Figures 175 and 179). They were concentrated at three different elevations, which were denominated terraces, at 225, 210, and 195 metres above sea level. In the first published account (Andah and Momin, 1993), it was claimed that the material from all three terraces could be assigned to the Early Stone Age (ESA), but in Momin’s own account (1995) it was subsequently suggested that the artefacts at the 195 metre level could more appropriately be regarded as MSA. According to the authors, three test excavations were conducted in 1994, but the results from these excavations have never been published. Work was resumed in the area by Obare Bagodo in 1996-1997 and 2002 as a PhD project, completed in 2004, and later published in WAJA (Bagodo, 2012). In particular, excavations were carried out at Mallam Umaru Gol’s Farm (UMF) on the eastern side of the stream at a point on the 195 metre contour line (Bagodo 2012, Fig. 19). An L-shaped 12 m2 trench designated D provided the essential stratigraphic record, and the bulk of the excavated material, supplemented by a certain amount more from Cutting I immediately to the north of it. The stratigraphy in the second location was said to be ‘not notably different’ from that at Trench D (Bagodo 2012: 93). Bagodo’s account provides a detailed record of the succession in D, which can be summarised as in Table 16. The sediments were analysed according to standard procedures (Wentworth, 1922; Krumbein, 1934; Folk and Ward, 1957; Folk, 1965) with a concentration on layers B, C, E, and F. Layers C and E constitute the upper and lower occupational levels. All the layers are characterised as poorly sorted and are uniformly acidic. The author draws attention to the presence of numerous pebbles and larger stones, and also to the lateritic crust, which is commonly interpreted as a product of alternating wet and dry conditions (Thomas 1974, chapter 2: 49-82). While recognising that the sequence as a whole could be regarded as belonging to the Bodija Formation, the author was reluctant to ascribe the layers to any particular member of that Formation (Bagodo 2012: 109-112). The percentage occurrence of the coarse fraction (gravel and very coarse sand above 1 mm in size) shown in Table 16 however suggests that, certainly for layers E and F, there is a good parallel to the Agodi Member, and that should not occasion any surprise. Table 16. Ajibode UMF site. Stratigraphy of Trench D. Median grain size Layer Sample A

Depth (cm)

Field description

0-12.6

Dark reddish brown loose soil.

Munsell

Φ

mm

Sediment category by median grain size

Φ= 1 mm)

5 YR 3/4

B

I

Dark yellowish brown loose sandy 12.6-33.7 silt. 10 YR 3/4 Lies disconformably on C.

0.50

0.75

Coarse sand

33.1%

C

II

Reddish brown gravelly sandy silt. 33.7-42.4 Lies disconformably on D. Upper occupational level.

5 YR 4/3

0.20

0.90

Coarse sand

23.0%

D

42.4-57.4 Yellowish red lateritic crust.

5 YR 5/6

E

III

Light reddish brown gritty loose 57.4-79.2 deposit. Lower occupational level.

5 YR 6/4

-0.80

1.90

Very coarse sand

61.9%

F

IV

79.2-93.0

5 YR 6/3

-0.50

1.50

Very coarse sand

55.5%

Light reddish brown sterile gritty deposit, overlies bedrock.

116

The Middle Stone Age of Nigeria

During the excavation, all lithic materials of whatever kind, a total of N Length Width Thickness 1822 pieces, were removed from the Modal class site. ‘On the basis of closer observation’, flakes 29 3.5-4.5 cm 1.5-2.5 cm 0.5-1.0 cm the majority of these were rejected and Mean value not subjected to further study (Bagodo 2012: 113). Combining the material cores 19 4.95 cm 3.80 cm 2.50 cm from layers C and E, from both Trench D points 20 3.98 cm 2.60 cm 0.86 cm and Cutting I, only 87 artefacts survived picks 7 7.97 cm 4.67 cm 3.47 cm the sorting process, 40 from layer C and 47 from layer E (Bagodo 2012, Tables 13 and 14). Of these, 19 were classified as cores, 29 as flakes, and 39 as tools, 53 being of quartz and 34 of quartzite. It is said that many of the cores and flakes are Levallois in character, but on the basis of the illustrations it is impossible to verify this. The 39 tools are divided by the author into the following categories: endscrapers 1, notched flakes 1, denticulates 1, sidescrapers 8, points 20, picks 7, and choppers 1. Some metrical characteristics of the flakes, cores, points, and picks are summarised in Table 17. Table 17. Ajibode UMF site. Dimensions of flakes, cores, points, and picks.

The modal classes of the flakes are as given by the author, whereas the mean dimensions of the other artefacts have been calculated on the basis of the figures given by him. The author goes so far as to say that since in his view the deposits at the UMF site cannot be dated geologically (Bagodo 2012: 111-112), it is only on the basis of archaeological evidence that the decision to call the industry at the site MSA can be justified. In that respect, it can be seen that the small tool assemblage, with its predominant points and picks, as well as sidescrapers, bears little resemblance to the MSA of Northern Nigeria. The author prefers to compare it to the ‘MSA of Southern Ghana’ identified by Oliver Davies, but since Davies himself referred to this entity disparagingly as ‘an industry of a most amorphous character’ or ‘a nondescript culture’ (Davies, 1964: 124; 1967: 133) it is difficult to have much confidence in that comparison, let alone to use it as a basis for the establishment of a so-called ‘floating chronology’ (Barker 1993, 207) for the site as a whole. If anything, the presence of admittedly small picks suggests that this small assemblage has something in common with the Sangoan, but, as the term is used in West Africa, that is also an illdefined concept. Only a personal inspection of the artefacts could lead one to feel at all confident about an appropriate classification for this material. Olude-Araromi Olude-Araromi (7° 29’ N, 3° 51’ E) is a site north-west of Ibadan on the road to Akufo, about 8 kilometres from the then northern campus gate of the Ibadan Polytechnic (Figure 175). It was investigated by I.A. Kiladejo as part of her B.Sc. project for Combined Honours Archaeology/Geology in the Department of Archaeology at the University of Ibadan in 1980, the field work aspect being supervised by myself (Kiladejo 1980: i). The results were published in WAJA in 1981 by Andah and Ajayi, but it should be noted that the Figures relating to Ibadan, and all the Tables, were lifted direct from Kiladejo’s dissertation (including misprints), so it would have been appropriate to have listed her as a co-author rather than a name in the references. The following summary relies on the information in the dissertation. At the site, by the roadside (Figure 180), three profiles were established (Kiladejo 1980, Fig. 8, here Figure 181). The heights of profiles 1-3 are respectively 2.20, 2.05, and 1.30 metres. The profiles are on something of a slope down to the Oterun stream, the distance from profile 1 to 3 being a little over 6 metres, with the slope then continuing down to the river for another 12 metres (Kiladejo 1980, Fig. 7). As indicated, the Kongi member is absent at this location, deposits identified as representing the Orita and Agodi members resting directly on weathered gneiss in situ bedrock. Single Munsell colour determinations were obtained for the successive deposits, as detailed in Table 18 (Kiladejo 1980: 34). Nonetheless, as shown in Figure 181, the Agodi member at this site was divided into a number of facies, consolidated and unconsolidated, in places with iron oxide concretions or large quartzite slabs. The 117

The Middle Stone Age of Nigeria in its West African Context

Figure 180. Olude-Araromi, general view, sections left of the road, 13 December 1979. Table 18. Olude-Araromi. Stratigraphy with major units identified.

points at which numbered samples for analysis were taken are indicated on profiles 1-3. Five of these samples (4-7) relate to the Agodi member. Top soil Brown 7.5 YR 5/2 The degree of variation within this member can Orita sand Pink 7.5 YR 7/4 be demonstrated by reference to the differing Agodi gravel Pink 5 YR 7/4 percentage occurrences of the coarse fraction Weathered gneiss Reddish yellow 7.5 YR 7/8 (gravel and very coarse sand above 1 mm or Φ = 60,000 BP (Williams 1976: 438). Current estimates are much older: Aterian 45-150,000, Middle Stone Age >100 to 250,000 BP (Clark et al. 2008: 46). It is admitted however that the dates for these earlier occurrences were arrived at (in both cases) by analogy with what was thought to have been established elsewhere. They do not reflect independent dating achieved at the site itself. Overall, the sequence established at Adrar Bous provides an interesting comparison, and for the most part a contrast, to what has been discovered so far about the archaeological succession in the Younger Granite Province of Nigeria. With the exception of the Nok Valley, as recorded by Fagg, and possibly Pingell and Rop, there is no record of the MSA occurring in an established stratigraphy above the Acheulean. The Aterian in that area is absent, and (again with the exception of Nok) there is no parallel to the later occurrence of an in situ succession of late Pleistocene and early Holocene occupations such as we see at Adrar Bous. In part this is explained by their different geomorphological situations. At Adrar Bous there was an exceptionally favourable landscape configuration whereby sediments could become trapped between the main granite massif and the surrounding remnants of the ring dyke. This allowed the successive accumulation of in situ deposits in a reconstructable stratigraphic order, in such a manner that the environment at the time of deposition could be more or less convincingly established. It is of capital importance that the succession Acheulean-MSA-Aterian has been established at this site, and that certain diagnostic distinctions between the MSA and the Aterian have been made. Although the dates for both of them at the site are presently no more than estimates based on analogy, the radiocarbon dates for the Epi-Palaeolithic and later entities at Adrar Bous demonstrate how unlikely it is that the dates for the MSA at the Nigerian sites reveal anything other than the processes that have taken place since the creation of those industries. Returning to Adrar Bous itself, it has been suggested that the sites of Bilma and (in particular) Seggédim are comparable to the MSA and Aterian at that site, and that together they form part of a distinct ‘regional Middle Palaeolithic tradition’ (Clark et al. 2008: 157). 149

The Middle Stone Age of Nigeria in its West African Context

Figure 229. Bilma stratigraphic section (after Maley et al., 1971, Fig. 5). 1, Holocene layers. 2, Calcareous diatomites. 3, Silémi terrace, including the following sequence from the top: (e) sand with calcareous nodules (d) lacustrine limestone with reed prints (Gif-1788 33,400±2500 BP) (c) coarse sandstone (in situ Levallois industry) (b) white aeolian sand (a) coarse sandstone (in situ Acheulean industry).

Bilma The site of Bilma (18°39’N, 12°56’E) was discovered in 1970 by a team from ORSTOM (Maley, Roset, and Servant, 1971). Its stratigraphy was described as in Figure 229. In the central portion of the site, the Silémi terrace 1 was divided into five layers. A Levallois dominated industry was located in layer (c), beneath layer (d), described as a lacustrine limestone with reed prints. This layer was dated to 33,400±2500 BP (Gif-1788, Delibrias et al. 1974), the layer itself being interpreted as evidence of a moist period. There were three other dates relating to the diatomite formation 2 (the oldest of which was 8480±300 BP) and one to the Holocene formation 3 (5070±110 BP). An Acheulean industry was located at the base of the Silémi terrace 1 in layer (a), beneath an intervening layer of aeolian sand (b). The Levallois dominated industry was in situ, although it was thought to have been somewhat displaced. It was said to be abundant, but unfortunately it has not been described in detail. Tillet (1983: 53-77) returned to the site three times in 1978 and 1979. He confirmed that the Acheulean was in situ, and he made an extensive surface collection, including 171 bifaces, without any trace of the Table 20. Seggédim. Artefact inventory. Levallois technique. Seggédim The open air site of Seggédim (20°12’N, 12°58’E) was excavated by Tillet (1983: 195-243) in 1976-1979. A single Aterian occupation was recognised, in three distinct areas, two smaller ones (A and B) and one larger one (C). In Tillet’s book on the subject, only area A was published in detail. Some environmental evidence was recovered, including pollen grains and particles of wood identified as tamarisk, from a red-brown occupation soil, suggesting that this was an oasis at the time of the Aterian settlement (Tillet 1983: 31-32, 198). Cyperaceae (sedges) and Typha (rushes) are indicative of such an environment, with abundant Gramineae (grasses) in the surrounding landscape. Area A, according to Tillet, contained 5616 artefacts all told, including 1445 tools and 237 cores. 137 of the cores could be classified as Levallois, and taking into account all the blanks the Levallois technological index (IL) came to 35.02. The 1445 150

Levallois flakes and points

903

Retouched Levallois points

3

Pseudo-Levallois points

8

Mousterian points

8

Sidescrapers

132

Endscrapers

21

Burins

37

Awls

29

Backed knives

26

Notches

43

Denticulates

106

Tanged points

12

Other tanged tools

36

Bifacial foliate points

47

Various

34

Total

1445

The Middle Stone Age in West Africa

tools, in an inventory based on Tillet’s count (1983: 210-211), but somewhat simplified, are listed at Table 20. The importance of Levallois flakes and points is clear, with a Levallois typological index (ILty) of 62.68. Sidescrapers and denticulates play a prominent role, but bifacial foliate points and tanged tools of various kinds are the distinctive elements that would be expected in an Aterian (rather than a generally Middle Palaeolithic) assemblage. The decisive role played by the Levallois technique is clear and is reminiscent of the situation at Adrar Bous. Mékrou Valley For the sake of completeness, it should be mentioned that relevant Palaeolithic sites have also been located along the banks of the Mékrou river, in the extreme south-west corner of Niger (Vernet, 1994). This river, a right hand tributary of the Niger river, presently forms the boundary with the Republic of Bénin to the south. On the northern side of the river, in its middle section, there is an extensive system of terraces, which were investigated by Vernet and found to contain abundant archaeological remains. He lists over 70 sites, the majority of which are neolithic or historical, but there are also 7 attributed to the Acheulean and 10 to the Middle Palaeolithic, as well as a small unspecified number regarded as later Upper Palaeolithic. The sites extend over a distance of about 30 km, between the villages of Meyoyaga and Boniongou, where the direction of the river’s flow turns from north to east (Vernet 1994, Fig. 1). This area corresponds to an approximate rectangle with boundaries 12°-12°30’N and 2°20’-2°40’E. Practically none of the material was found in situ, but from the description given it clearly merits attention. The raw material employed in the Middle Palaeolithic was almost exclusively quartzitic sandstone, in contradistinction to lithics attributed to later periods, which became more varied. The condition of the artefacts recovered was variable, depending in the author’s opinion on their degree of exposure to the elements. They included sidescrapers of various kinds, discoids, notches, and denticulates (Vernet 1994, Figs. 8 and 9), but apparently no Levallois elements, at least not in the author’s enumeration. As a whole, he considers that the valley has the potential to be no less significant than the Falémé valley in eastern Senegal. That may well be true, and a more detailed survey would no doubt produce a lot more information about it. Ghana There is a long history of Palaeolithic research in Ghana, thanks above all to the work of Oliver Davies, summarised in his two books (1964 and 1967) as well as in numerous articles, and no student of the subject can afford to ignore his work, but it does give rise to certain difficulties (cf. Maggs, 1986). I have recently published a full assessment (Allsworth-Jones, 2017), in which I emphasised both the positive and the negative aspects of this work, so it is not proposed here to repeat all those arguments. On the positive side, it is important to take note of Davies’s “Ghana Field Notes” (1970-1976) consisting of four sets of cyclostyled records with his detailed observations, including exact geographical coordinates, concerning all the sites which he visited in Ghana. This is an invaluable reference, which must be taken into account, in addition to his formally published work. Some of the major sites identified by him are shown in the map at Figure 230, with his coordinates. On the other hand, Biberson, Clark, and Fölster (1968) already drew attention to some of the drawbacks inherent in his approach. The great majority of Davies’s material consists of surface collections, and the number of good sections into which he could tie his finds was comparatively few. Comparison to the originals suggests that in many cases Davies’s typological identifications are surprising and doubtful, even misleading. Nowhere is there a rigorous definition of types. This difficulty is compounded by the style of the illustrations, some of which were carried out by Davies himself but others by Mr E. Quansah (Davies, 1964, Preface: xiii). The illustrations do convey an adequate idea of the general appearance of the objects, but it is seldom possible to discern the techniques by which they were made. There is no consistent orientation and no indication whether any individual piece is broken or not. These 151

The Middle Stone Age of Nigeria in its West African Context

Figure 230. Map of Ghana, with principal sites indicated.

152

The Middle Stone Age in West Africa

considerations mean that there is no substitute for a personal inspection of the collections, and it is this which the author was able to carry out during a visit to Ghana in 1980-1981, when I stayed at Legon, thanks to the kind invitation of Professor John Sutton. Davies was obliged to use geological features to erect a comparative chronological framework, the number of absolute dates at his disposal being very few. In doing so, he employed the dominant paradigm of his day, whereby it was assumed that broadly speaking pluvial periods in Sub-Saharan Africa corresponded to glacial episodes in the temperate northern hemisphere, as outlined in the Introduction. Thus the Gamblian or fourth pluvial corresponded to the last glacial period. It was preceded by the third interpluvial, corresponding to the last interglacial. Two wet phases of lesser intensity referred to as the first and second post-pluvials (or Makalian and Nakuran) came after the Gamblian, and were separated from it and from each other by two arid episodes. It was assumed that the Middle Stone Age in SubSaharan Africa corresponded chronologically to the Upper Palaeolithic in North Africa and Europe. Following finds attributed to the Acheulean, evidence claimed as Sangoan has a very significant presence in Ghana. The most characteristic tool of the Sangoan was said by Davies to be ‘the heavy pick, biface or less commonly uniface, made preferably on a large pebble of which the butt was left unflaked’ (Davies, 1964: 98). He considered certain artefacts, including some biface picks, to be Kalinian or Lupemban and thus separate from the Sangoan. They were said to form ‘a continuous series from before and through the Gamblian pluvial’ (Davies, 1964: 108). The evidence quoted for these entities was however never more than very slight. Davies admitted that only one piece of ‘true Kalinian type’ came from Ghana; this was from Kunkoa in the extreme north, and it apparently lay in the same lower stone-line as Sangoan picks (Davies, 1964, Fig. 34.3; Davies, 1970-1976, Part 2: 98-99). The stray Lupemban pieces found were said to be nearly all without proved association (Davies, 1964, Fig. 43). Davies recognised a number of MSA entities, including the Guinea Aterian, the MSA of Southern Ghana, and the Ultimate MSA or ‘pre-mesolithic’. The name ‘Guinea Aterian’ was coined by Davies. One site mentioned in this context is Jimam, on the river Oti (Davies 1970-1976, Part 1: 40-43). There are two radiocarbon dates here of 12,400±300 and 11,400±250 BP (I-2264 and GaK-955) on a peat horizon above a basal gravel, but the basal gravel did not in fact contain any diagnostic material (Davies 1967: 79). At New Todzi, on the west bank of the river Detoe, Davies described a sequence about 3-4 metres thick (Davies, 1964: 77, 116-117, 119-121, 134, Fig. 48; 1967: 74-75, 132-133, 138; 1968; 1970-1976, Part 1: 75-76). In red sandy earth above an upper stone-line, an ‘Ultimate MSA’ horizon was stratified above one attributed to the Guinea Aterian. Davies emphasised the microlithic character of the Ultimate MSA, and to the extent that one can judge from his illustrations the same seems to be true of the Guinea Aterian, which suggests that both should be regarded as LSA. At a number of Davies’s other sites, e.g. Kuradaso and Kpafa, his socalled tanged pieces proved to be just undiagnostic flakes, and probably the same was true in this case. At Takoradi, on the coast, a deposit of woody peat was flooded by a marine transgression (Davies, 1964: 158-159, 183; 1967: 49, 80; 1970-1976, Part 4: 202-205). A date on material from this peat was first given as 5570±70 BP (Gro-1194) but was subsequently amended to 5810±70 BP (Flight, 1968). The archaeological material from the site is very slight. An artefact described as a backed blade was found in the associated beach and was ascribed to the Ultimate MSA (Davies, 1964: 136 and Fig. 70.10). This piece is no proof of MSA, and as at Jimam the dates quoted by Davies are essentially of geological significance only. Some of the weaknesses of Davies’s system should be evident from the above. He placed an excessive reliance on isolated specimens, to which names were arbitrarily attached. The connection between dated geological events and their supposed archaeological correlates could be tenuous or non-existent. His writing also shows a strongly marked tendency to treat stone tool ‘evolution’ in biological terms, e.g. a supposed ‘fusion’ of the Late Sangoan with the Lupemban, and it is often assumed that changes were brought about by waves of invasion. In these respects, it is fair to say that Davies reflected widespread attitudes among the prehistorians of his time, and allowances should be made for that. It is also true to say that some of the sites mentioned might be worth reinvestigating, with new research goals in mind, but for now in this context, many of them have to be disregarded. Since Davies’s work still looms so large 153

The Middle Stone Age of Nigeria in its West African Context

in anything written about the Palaeolithic of Ghana, there is no alternative but to subject it to a critical review, but it still provides an indispensable starting point for any further study. Some of the principal sites mentioned by Davies which do repay detailed investigation are considered below. Asokrochona, the Ghana Nautical College, and Tema Asokrochona (5° 36’ N, 0° 03’ W) is about 6½ km northeast of Accra, near the coast, on a low ridge between the Mokwe and Sakumo lagoons, both of which according to Davies will formerly have been estuaries. A railway cutting was made through the northern end of the ridge, running in an approximately east-west direction, in 1953 (Figure 231). The exposed northern and southern faces of the railway cutting were examined by Davies in 1958, and he made section drawings of them (Figure 232). The sections and the material recovered from them provided the basis for his account of the site (Davies, 1964: 137-142). A road runs parallel with the railway line about 160 metres south of it, and excavations were conducted in the area between the two by University of Ghana staff in 1972-1974. Four test pits (I-IV) and two areas (I and II) were excavated. Nygaard and Talbot first reported their results from test pits II-IV, and later from area II, with appropriate schematic sections (Nygaard and Talbot, 1976, Fig. 2; 1984, Fig. 2). Area I was excavated by Bassey Andah and has been reported on by him (Andah, 1979). The stratigraphy as described and interpreted by Davies, and by Nygaard and Talbot, differs in certain respects. Davies’s units are numbered alphabetically from the top down, as follows. D. (1) Soft grey sandy earth. (2) Hard red sand. Some ‘mesoneolithic’ microliths were possibly associated with this deposit. C. A crust or stone line, extending over the whole area, covering both deposits A and B below. In the centre, the crust covered a band of lateritization representing the altered surface of B. Unrolled Sangoan material was incorporated in the crust directly above A, but it was mixed with Late MSA

Figure 231. Asokrochona railway cutting, 28 December 1980.

154

Figure 232. Asokrochona railway cutting (after Davies 1967, Fig.1).

The Middle Stone Age in West Africa

155

The Middle Stone Age of Nigeria in its West African Context

artefacts, which occurred on their own in the crust above B. Davies regarded the crust, and thus the MSA but not the Sangoan, as belonging to the maximum of the last glacial period. B. Mottled sandy clays, up to 3-4 metres thick in the centre of the cutting, regarded as slumped ‘Terre de Barre’ material, without artefacts. Regarded as equivalent to the early part of the last glaciation. A. Sand and beach shingle 50-70 cm thick immediately above bedrock, with rolled Late Acheulean tools. Davies regarded this as a remnant of his Beach IV, equivalent to the maximum of the last interglacial period. Hence the Sangoan, stratified above the beach and either below or mixed with the MSA, could be expected to have belonged somewhere in the early part of the last glaciation (Davies, 1976: 888). With this may be compared the succession as described by Nygaard and Talbot, shown here at Figure 233. 3b. Kpone Formation. Yellowish-red medium sand, with some potsherds and scattered flakes, LSA presumably equivalent to Davies’s ‘mesoneolithic’. Considered to be of aeolian origin, and to belong to a dry windy phase about 4500-3500 years ago. 3a. Nungua Formation. Red or mottled medium or clayey sand, often pebbly near the base. A few flakes and cores were found, and these are regarded as MSA, thus supporting Davies’s idea that there was a separate MSA horizon. This Formation is also regarded as of aeolian origin, and the authors suggested that it may have accumulated about 25-13,000 years ago during an arid windy phase. In places (though not at Asokrochona) the Nungua Formation is capped by a palaeosol, and both this and the pebbles at the base were regarded as indicative of more humid conditions.

Figure 233. Asokrochona generalised section (after Nygaard and Talbot 1976, Fig.2).

156

The Middle Stone Age in West Africa

2.

1.

Asokrochona Formation. Ironstone gravel of well-rounded laterite fragments resting on a deeply gullied surface, suggested to have been produced in situ by exposure and erosion of the lateritic profile beneath under semi-arid climatic conditions. Presumably, Davies’s crust corresponds approximately to this Formation, which produced, according to Nygaard and Talbot, ‘the great majority’ of the archaeological material, i.e. Davies’s Sangoan. They emphasise however that most of the artefacts were found on top of the Asokrochona Formation; only a few pieces occurred within the uppermost 5 cm and none were found at the base of the gravel (Nygaard and Talbot, 1976: 1415; 1984: 31). Hence the human occupation is likely to have post-dated the formation of the gravel, and they confessed that at the time of writing they could not actually suggest an appropriate timebracket for it (Nygaard and Talbot, 1984: 35). Tema Formation. Highly weathered sand, with a conglomeratic base and nodular laterite at the top, regarded as of Tertiary age, above bedrock (gneiss cut by quartz veins up to 25 cm thick). Contrary to Davies’s account, Nygaard and Talbot specifically state that these sediments contain no features that are in any way diagnostic of beaches. They also failed to find any traces of Davies’s Late Acheulean industry. The sole material evidence for the latter therefore is that published by Davies (1964, Fig. 65) and, in my opinion, this small collection cannot be taken as satisfactory evidence for the existence of any occupation horizon earlier than the main one at the site.

Apart from Asokrochona, Nygaard and Talbot recorded the presence of MSA stratified above Sangoan artefacts at the Ghana Nautical College, on Nungua Bay, west of Asokrochona (5° 35’ N, 0° 04’ W) (Figure 234) and at Tema West, east of the site, where the Asokrochona Formation is also prominently in evidence (5° 37’ N, 0° 00’ W) (Nygaard and Talbot, 1984, Figs. 1 and 2). Tema West was originally investigated by Davies (1964: 125-128; 1970-1976, Part 4: 212-218). As he said, extensive works to create a new town and harbour in this location had uncovered numerous sites, and he found that the MSA often occurred on a lateritic crust above red sandy clays (as with units C and B at Asokrochona). He distinguished three main find areas. Since that time, as Nygaard and Talbot pointed

Figure 234. Ghana Nautical College, 28 December 1980.

157

The Middle Stone Age of Nigeria in its West African Context

out, large amounts of sediment have been removed in connection with the building of the harbour, and they were able to study new sections and new sets of artefacts which had not been available before. They investigated a 200 metre long cutting exposed by the harbour works, and they opened two test pits: I to the east and II to the west. Whereas I produced material similar to that from Asokrochona in the same stratigraphic context, II produced a somewhat different collection at the base of the Nungua Formation. The artefacts were said to be ‘obviously related’ to the main Asokrochona industry, but there were some distinctions in terms of raw material and typology, and the stratigraphic context presumably means that they should be somewhat younger than the main occurrence. The artefacts present in the main occupation horizon at Asokrochona have been separately reported by Nygaard and Talbot and by Andah (Nygaard and Talbot 1976: test pits II-IV, 1984: area II; Andah 1979: area I). None of these authors found the task of classification particularly easy. Andah experimented with an approach based on attribute analysis, concluding that only three broad functional classes were present (Andah 1979: 82 and Fig. 6). Nonetheless, he also presented a more ‘traditional’ type list (Andah 1979: Table 3 and Fig. 4a, 4-9) and this allows for comparison between his results and those of Nygaard and Talbot. These results were already reported and assessed in an earlier article (Allsworth-Jones, 1987) so those arguments will be summarised here as briefly as possible, with the statistics reproduced in Tables 21 and 22. The artefact totals, in terms of primary components, are presented in Table 21. The figures given by the different excavators are reasonably consistent, and they confirm their general conclusion that this was certainly a factory site. It is much more difficult coming to an acceptable list of the tool classes Table 21. Asokrochona artefact totals Artefacts

Test pits II-IV

Tools

173

Cores Worked debris

Area II 5.6%

297

891

7.3%

22

0.2%

-

14

1589

63.4%

-

-

Manuports

586

23.4%

Total

2507

Hammer/ grinding stones Flakes Chips & chunks

6.9%

692

137

5.5%

22

0.9%

-

Area I

Total 5.5%

1162

371

6.8%

1399

6.9%

42

0.8%

86

0.4%

0.1%

25

0.5%

39

0.2%

7125

58.1%

2944

54.1%

11,638

57.7%

1731

14.1%

516

9.5%

2247

11.1%

1791

14.6%

1248

22.9%

3625

17.9%

12,266

5443

5.7%

20,216

Table 22. Asokrochona tool classes Artefacts LCT

HD

LD

Handaxes

Test pits II-IV 1

Area II

0.6%

Area I

8

1.2%

-

-

Knives

2

1.2%

7

1.0%

-

-

Picks

13

7.5%

35

5.1%

6

2.0%

Core-axes

10

5.8%

5

0.7%

-

-

Choppers

17

9.8%

101

14.6%

121

40.7%

Core-scrapers

27

15.6%

-

-

-

-

Spheroids

6

3.5%

46

6.6%

71

23.9%

Other large tools

8

4.6%

-

-

-

-

Scrapers

76

43.9%

404

58.4%

82

27.6%

Other small tools

13

7.5%

86

12.4%

17

5.7%

Total

173

LCT

3

1.8%

15

2.2%

0

-

HD

81

46.8%

187

27.0%

198

66.6%

LD

89

51.4%

490

70.8%

99

33.3%

692

158

297

The Middle Stone Age in West Africa

present as shown in Table 22. The broad division is into Large Cutting Tools (LCT) Heavy Duty Tools (HD) and Light Duty Tools (LD) according to the best approximation this author was able to come to on the basis of the information supplied by the excavators. A number of points of difficulty need to be specified. In their first account, for test pits II-IV, Nygaard and Talbot separated core-scrapers from other scrapers, but in their second account, for area II, this was not done. Hence the apparent differences between the two in terms of HD vs LD may be artificially exaggerated. The difference between the first two lists and Andah’s list is obviously more considerable. He illustrates only 6 bifaces, which may be more appropriately classed as picks rather than handaxes, and he makes no mention of core-axes. Both choppers and spheroids figure prominently in his list, which means that the HD total far outweighs that for LD. Andah admits the contrast between areas I and II, which in his view might reflect chronological factors (Andah 1979: 83), although one suspects that the ‘large dose of subjectivity’ to which he refers (Andah 1979: 82) may also have a lot to do with it. There is nothing in the site stratigraphy to indicate any chronological divide between the two areas, although, since this is a factory site, there could well have been differences between its various parts, without affecting its homogeneity. Clearly the ideal solution would be for a fresh study to be undertaken of the entire material to come to a unified view of it. In the absence of that, one has to take note of the various suggestions that have been made so far as to the appropriate classification of this industry. Davies, accepting earlier criticism of his approach, in the end suggested that the name Sangoan should be restricted to Central Africa, and that the Ghanaian occurrences (including Asokrochona) should be re-named the ‘Awudome industry’, after the ridge of that name east of the river Volta (Davies, 1976). Andah (1979) also suggested the use of a different term, the ‘Sakumo industry’, after the adjacent lagoon of that name. Nygaard and Talbot (1984) favoured the expression ‘Asokrochona industry’, but they conceded that it did still form part of the ‘larger Sangoan Industrial Complex’ which occurs in other parts of Africa as well. It seems to me that this is correct, and that, as I said already, it would be appropriate to regard this manifestation of the Sangoan as a distinctive cultural entity ‘closely allied to or possibly derivative from the Late Acheulean’ (Allsworth-Jones, in ed. Sutton, 1982).

Figure 235. Asokrochona [Oliver Davies]. 1, 2 and 4, disc cores. 3, Levallois flake/blade.

Figure 236. Asokrochona [Oliver Davies]. 1, core-axe. 2, blade. 3, convex sidescraper. [QCG Fig. 67.18, 21, 3].

159

The Middle Stone Age of Nigeria in its West African Context

Figure 237. Asokrochona [Oliver Davies]. 1, handaxe. [CQG Fig. 66.1]. 2, Levallois flake/point.

Figure 238. Asokrochona. 1, initial core. 2, disc core. 3, 2-platform flake/blade core. 4, 1-platform flake/ blade core. [1, test pits I-IV, 2-4 Area II].

Figure 239. Asokrochona. 1, 3, 5 and 6, disc cores. 2, notch. 4, sidescraper on Levallois flake. [1 and 3-6, Area II, 4 Area I].

Figure 240. Asokrochona. 1 and 5, bifaces in course of manufacture. 2, point/convergent sidescraper. 3, disc core. 4, convex sidescraper. 6, convergent sidescraper. 7, awl. [Area II].

160

The Middle Stone Age in West Africa

Some of my drawings of artefacts from Asokrochona are at Figures 235-243 and 244, 1-3. Drawings from Tema are at Figures 244, 4-7 and 245, and two artefacts from the Ghana Nautical College are at Figure 250, 1-2. All of these drawings were executed during my visit to Legon. As noted, many of these pieces come from the Oliver Davies collection, and in some instances they correspond to items published by him in ‘The Quaternary in the Coastlands of Guinea’, in which case the reference is indicated (QCG, 1964). Other pieces come from the collections of Nygaard (Asokrochona area II and test pits I-IV, Tema West I and II, Nautical College) and Andah (Asokrochona area I). Figure 236 includes items specifically stated by Davies to have come from his MSA level at Asokrochona, and Figure 237,1 is specifically stated to be Sangoan. Otherwise there is no precise indication as to whether the artefacts are to be attributed to the MSA or to the Sangoan at these sites. On a typological basis, some pieces (in addition to the handaxe at 237,1), can be regarded as characteristically Sangoan, especially those from Asokrochona area II (Figures 240-243) which include bifaces in course of manufacture, chopping tools, choppers, handaxes, and picks. Interestingly enough, Davies attributed one of his core-axes (Figures 236,1) to the MSA level at Asokrochona and not to the Sangoan. The quartz bifaces photographed at Figure 246 are certainly characteristic for the Sangoan, which must account for the great majority of the pieces. Nonetheless, there are numerous artefacts in the collections studied which could be regarded as characteristically MSA, such as flakes with facetted striking platforms, including Levallois flakes and flake/blades, as well as retouched tools made on these blanks, and quite a large number of disc cores. In the absence of stratigraphic information, it is not possible to determine whether these pieces come from MSA layers as such or whether they represent the emergence of a new technology within the Sangoan. Ambiguity in the existing state of affairs cannot be avoided, and could be resolved only by a more detailed study of the available records, or by careful new excavations, with plotting of the individual artefacts.

Figure 241. Asokrochona. 1, convergent sidescraper. 2, denticulate. 3 and 5, chopping tools. 4, chopper. [Area II].

Figure 242. Asokrochona. 1, pick. 2, transverse convex sidescraper. [Area II].

161

The Middle Stone Age of Nigeria in its West African Context

Figure 243. Asokrochona. 1, handaxe. 2, point. [Area II]. Figure 244. Asokrochona. 1, point. 2, concave sidescraper/notch. 3, awl. [Area II]. Tema [Oliver Davies]. 4, Levallois flake. 5, blade. 6, disc core. [QCG Fig. 57.7, 56.9, 56.7]. 7, disc/bipolar core [Tema West I 1976 Nygaard].

Figure 246. Asokrochona quartz bifaces.

Figure 245. Tema [Oliver Davies]. 1, handaxe/ pick. 2, Levallois flake. 5, pseudo-levallois point. [QCG Fig. 56.16 and 17]. 3, convex sidescraper. 4, chopper. 6, chopping tool. [Tema West II 1976 Nygaard].

162

The Middle Stone Age in West Africa

Other sites in Ghana The same apparent conjunction of Sangoan and MSA forms can be noted at several of Davies’s other sites, of which mention might be made here in particular of three, Legon Botanic Gardens, Hohoe, and Chawenu. Davies himself conducted excavations at Legon Botanic Gardens (5° 39’ N, 0° 12’ W) in 1958, by means of a number of pits, some of which reached a depth of about 2 metres (Davies, 1964: 76-77, 124-125, 135-136, 149, 155, and Fig. 54; 1967: 74-75 and Fig. 3; 1968). According to Davies, there were a few heavily weathered Sangoan pieces at the base with some large stones interpreted as a lower, pre-Gamblian, stone-line. This was surmounted by 25 cm of red earth interpreted as an eroded Gamblian soil, and then by an upper stone-line. In and just below this stone-line was the ‘lower Main MSA’ occurrence (alternatively referred to as Late MSA). 50 cm of red sandy earth above the upper stone-line contained two artefact horizons, the first considered to be ‘upper Main MSA’ and the second ‘Ultimate MSA’. This unit was surmounted by a thin layer of rolled lateritic nodules. A ‘mesoneolithic’ zone occurred 15-20 cm above this layer, followed by ‘true’ neolithic and iron age finds. While therefore on paper the situation looks to be quite clear-cut, the reality in terms of the available artefacts is not. Davies himself stated that ‘the Sangoan was not separated stratigraphically from the MSA; it was possible to distinguish Sangoan material only on the basis of typology and patination’, and 38 pieces were selected on that basis (Davies, 1968: 1155). With that limitation, the chosen Sangoan artefacts seem to be of the usual kind (Davies, 1968, Figs. 5 and 6). With regard to the MSA, Davies admitted that it was ‘not easy to describe or sort; some names and forms are not very certain, and colleagues might hesitate to call them more than utilised flakes’ (Davies, 1968: 1159, 1161). My examination of the material at Legon confirmed the accuracy of that observation. Davies listed 1964 artefacts as belonging to the MSA, but the majority are indeed no more than waste products. Moreover, in my opinion, many of the pieces which could be accepted as tools or cores would fit far better into an LSA context. That is disturbing in terms of the general classification of the material. There are only a few pieces which, in terms of technology or typology, could be regarded as MSA. They include some flakes with facetted striking platforms (Davies, 1968, Fig. 4.10), pseudo-levallois points (Davies, 1968, Fig. 4.2 and 9), possibly Levallois flake/blades, and disc cores (Davies, 1968, Fig. 4.3 and 23). Hence it seems that Davies included in his MSA much material which can in fact be better classified as LSA. That is not to say that (as Davies claimed) there may not have been genuine Sangoan and MSA horizons at the site, but it would probably require new excavations to satisfactorily clarify the situation. A not altogether dissimilar scenario appears to have existed at two sites on the coast in the vicinity of Accra, Manprobi (5° 31’ N, 0° 15’ W) (Davies, 1964: 32, 136, 156-158, 174, 183, Figs. 71 and 72; 1967: 49, 135136, 139, Fig. 32; 1970-1976: 151-152) and Labadi (5° 32’ N, 0° 10’ W) (Davies, 1964: 30, 136, 159-160, Fig. 73; 1967: 139; 1970-1976: 135-138). Here again, Sangoan, MSA, and ‘mesoneolithic’ occurrences were tied in to geological phenomena, although the instability of the sections may be illustrated by the fact that Davies’s ‘sketch-profile’ at Manprobi ‘was taken on one day and had disappeared a week later’. The Late MSA illustrated pieces from the two sites again consist almost entirely of unretouched flakes, although some of them do have facetted striking platforms (Davies, 1967, Fig. 72.7 and 73.4). Hohoe and Chawenu, both inland sites in what is now eastern Ghana, provide in my view rather clearer instances of the joint occurrence of Sangoan and MSA at certain favoured localities. Hohoe (7° 10’ N, 0° 26’ E) is on the west bank of the river Dayi (Davies, 1964: 55-56, 67, 73; 1967: 54, 57; 1970-1976, Part 1: 36-37). According to Davies, who examined various exposures, the Sangoan was situated on or above a 12-13 metre terrace gravel and in the laterite which overlay it (Davies, 1967, Plate 14). The MSA was in a stone-line above the laterite (Davies, 1967, Plate 26). Davies referred to ‘several unrolled Early Sangoan picks’, and one of them is illustrated here at Figures 247 and 249,3. Technologically and typologically it is quite comparable to artefacts from Asokrochona. The MSA material is not abundant, but does include a genuine Levallois element. Chawenu (6° 34’ N, 0° 19’ E) is on the west bank of the river Tsawe (Davies, 1964: 57-58, 134-135; 1967: 65, 85, 139, 180; 1970-1976, Part 4: 87-88). At different localities in the same area, Davies described an 163

The Middle Stone Age of Nigeria in its West African Context

Figure 248. Chawenu [Oliver Davies]. 1, Levallois flake/ blade core. 2, Levallois flake. 3, biface in course of manufacture. 4, pseudo-levallois point. 5, endscraper. 6, chopping tool. Figure 247. Hohoe handaxe/pick.

8-10 metre terrace gravel with Sangoan, and an MSA beneath a surface of lateritic nodules; but at all the find spots listed, both Sangoan and MSA type artefacts appear in the collections kept at Legon. The Sangoan type pieces appear to be in every way comparable to those from Asokrochona; their state of preservation does not differ from those which can be assigned to the MSA. Davies’s classification of the MSA type tools is largely illusory, but what one does find are convincing examples of flakes with facetted striking platforms, including Levallois flakes, pseudo-levallois points, disc and Levallois flake/blade cores (Figures 248 and 249, 1 and 2). Both these sites therefore certainly deserve a reexamination. So far, we have noted the occurrence of material which typologically, or more particularly technologically, is of a generally middle palaeolithic character, either in somewhat ambiguous association with Davies’s Sangoan or, as a small component, in assemblages which look for the most part more like LSA. But there are a few instances, on the basis of the material 164

Figure 249. Chawenu [Oliver Davies]. 1 and 2, disc cores. Hohoe [Oliver Davies]. 3, handaxe/pick.

The Middle Stone Age in West Africa

collected by Davies, where there are traces of an independent MSA horizon. Narago (10° 15’ N, 2° 45’ W), on the river Kule close to its confluence with the Black Volta, in the extreme northwest of the country, seems interesting in this regard (Davies, 1964: 6263, 133-134, Figs. 16.10 and 11, and 62; 1967: 66, 139-140, Fig. 35; 1970-1976, Part 2: 121-122). Davies identified two sites. Upstream he located what he termed rolled Guinea Aterian artefacts in basal gravels. Downstream, on the south bank of the river, there were small erosion patches with a few quartz microliths, regarded as derived from a higher layer of lateritic nodules, but also with a more abundant industry from a rubble horizon which he labelled Ultimate MSA. He listed a number of ‘classifiable’ artefacts (Davies, 1970-1976, Part 2: 122), and some were illustrated (Davies, 1964, Fig. 62; 1967, Fig. 35), but it is remarkable just how wrong an impression can be gained from his classification. The collection is interesting from a raw material point of view, since many different kinds were used: greenstone, chert, quartzite, granite, lava, and shale. There is a multi-platformed flake/blade core (Davies, 1964, Fig. 62.1) and some flakes with facetted or dihedral striking platforms, including Levallois flakes (Davies, 1964, Fig. 62.5). Similar pieces are illustrated here at Figure 250, 3-6, the drawings by Davies (1964, Figs. 62. 6 and 8) being given quite different names.

Figure 250. Ghana Nautical College. 1, disc core. 2, double sidescraper. [Nygaard 1976]. Narago [Oliver Davies]. 3, pseudo-levallois point. 4, retouched Levallois point. 5, Levallois flake. 6, retouched Levallois flake. [QCG Fig. 62.6, 62.8].

Davies’s Figure 62.4 is not a burin but probably a biface in course of manufacture, as is the ‘small handaxe’ which he refers to; both are in very flaky shale which makes interpretation difficult. In short, this little assemblage does seem to constitute a simple strongly Levallois-oriented Middle Palaeolithic, without admixture; it is actually quite similar to material of this kind from Nigeria, particularly Mai Lumba, and the MSA at Chawenu is not unlike it. A completely new direction, and an escape from old paradigms, has been provided by the discovery and analysis of the site of Birimi (10° 34’ N, 0° 25’ W) in Northern Ghana (Hawkins et al. 1996; Casey et al. 1997; Quickert et al. 2003). The site is on the south side of the Gambaga escarpment, where erosion has cut into a terrace beside the seasonal Birimi Kuliga stream, as shown in the map at Figure 251. The principal objective of the research work was the examination of a Kintampo culture (LSA) settlement, but in 1996 in situ MSA material was also found in a buff-yellow sediment horizon about one metre beneath the surface. This small assemblage (mainly on mudstone which must have been brought in from some kilometres away) shows clear signs of the use of the Levallois technique and has no heavy duty component. The two principal sections indicated on the map are KPP (Kintampo Pit Profile) and MSAP (MSA Profile) on the west and east sides of the main gully, respectively. The MSA Profile is reproduced at Figure 252. It clearly shows three different sedimentological units, all of which were apparently laid down by water action, and not by aeolian processes. A series of OSL dates have been obtained for the site (Quickert et al. 2003, Table 2) of which four came directly from the MSA Profile. The date of 40,800±11,400 years ago (BRSD3) was specifically associated with a Levallois blade. The date of 58,400±15,300 years ago (BRSD4) comes from the base of the same deposit but was not associated with any artefacts. The two younger dates (20.4±3.4 and 4.13±0.55 kyr) are post-MSA. The fifth date superimposed on the diagram, 28,400±4,100 years ago (BRSD6), comes from undisturbed sediments in 165

The Middle Stone Age of Nigeria in its West African Context

Figure 251. Birimi location and topographic map (after Quickert et al. 2003, Fig. 1).

166

The Middle Stone Age in West Africa

Figure 252. Birimi MSA profile east wall of main gully (after Quickert et al. 2003, Fig. 2).

the adjacent Kintampo Pit Profile, well below the Kintampo Pit itself and consistent with the other results, but not associated with any artefacts (Quickert et al. 2003, Fig. 3). A sixth date, not from either of the profiles, is reported to be associated with MSA artefacts at 23,600±2,900 years ago (BRSD55) in unit 25S 81E of the excavated area (Quickert et al. 2003, Fig. 4B). Since the excavators were not able to trace a single MSA horizon linking the two dated occurrences, they state that there is ‘no a priori reason to suggest that they reflect the same occupation event’. If so, the site may bear witness to repeated MSA occupations at around 40,800 and 23,600 years ago, respectively. By contrast, both radiocarbon and TL dates suggest that the Kintampo occupation of the site dates to around 3,500 to 3,800 years ago. Obviously the results from Birimi put the dating of the MSA in Ghana onto a completely new footing. They should provide a stimulus to go back and search for other such sites, including those mapped by Davies, using new techniques, to see whether these results are replicated elsewhere. Ivory Coast Quite apart from the two countries’ propinquity, it is logical to move on from Ghana to the Ivory Coast since the three most significant sites in that country - Attinguié, Anyama, and Gouabuo - have themselves been identified as Sangoan. The Sangoan industry from Anyama has been published according to one scheme of study, unlike Asokrochona, and to that extent at the moment it could be said to represent a more secure general frame of reference, whether one agrees with all the details of that scheme or not. Anyama also contains an MSA layer, well stratified above the Sangoan, and there are some other MSA sites as well, of which Sablière is the best documented for now. Attinguié and Anyama are each about 25 km north of Abidjan, whereas Gouabuo and Sablière are in the south west of the country, in the vicinity of Issia and Bouyo respectively, as shown on the maps at Figures 253 and 254.

167

The Middle Stone Age of Nigeria in its West African Context

Figure 253. Abidjan area, with Attinguié and Anyama.

Figure 254. South-west Ivory Coast, with Gouabuo and Sablière.

168

The Middle Stone Age in West Africa

Attinguié and Anyama Attinguié was the first of the sites to be published (Paradis, 1980). Good sections were observed by Paradis in an old quarry by the autoroute leading from Abidjan to Yamoussoukro (Paradis 1980, Fig. 1, A-D). Beneath superficial deposits, a clayey-sandy formation up to 6 metres thick was attributed to the ‘Terre de Barre’, beneath which were three lines of pebbles, above bedrock. The uppermost of the basal pebble lines produced two artefacts identified as a cleaver and a pebble tool. A more abundant industry was discovered at depths of 1 and 2.3 metres in the ‘Terre de Barre’, including according to Paradis bifaces, spheroids, pebble tools, and a few flakes, mostly of quartz. The site produced no direct dating evidence, but Paradis (partly on the basis of the archaeological finds) suggested that the ‘Terre de Barre’ at this location was certainly of Quaternary age, perhaps as recent as 40,000 years ago, and not ‘Neogene’, i.e. Miocene or Pliocene, which would be the case if it was regarded as part of the ‘Continental Terminal’. The stratigraphic position of the finds at Anyama is comparable to that at Attinguié, and the same questions of attribution and nomenclature arise, so it might be as well to take a closer look at this terminology before considering Anyama in detail. The basic geological formation throughout the greater part of the Ivory Coast consists of Precambrian bedrock. There is an important sedimentary basin along the seashore between Fresco and Axim, stretching inland for 50 km at the most (Bagarre and Tagini, 1965). The approximate boundary between the two formations in the Abidjan area is indicated on the map at Figure 253. Commonly the sedimentary basin has been said to consist of two parts, Quaternary deposits to the south of the Ebrié lagoon, and ‘Continental Terminal’ deposits between the lagoon and the junction with the Precambrian rocks to the north (Leneuf, 1968). The classification ‘Continental Terminal’ was originally proposed by Conrad Kilian for Saharan continental sediments of ‘Neogene’ age (Kilian, 1931; Taquet, 2007). It was subsequently used rather more loosely, but Kogbe (1978) cautions that it should strictly be confined to its original meaning. According to Tastet (1977, 1981) the sedimentary basin in the Ivory Coast is similar to those in Togo and Bénin. In his view, the deposits south of the Ebrié lagoon consist of two elements, ‘low plateaux’ attributed to an ‘Ogolian’ dry period (40-10,000 years ago) and two or three generations of sandy ridges formed during the Holocene. North of the lagoon, the ‘high plateaux’ do belong to the ‘Continental Terminal’ sensu stricto, but they are mostly covered by the ‘Terre de Barre’, attributed to the Quaternary and not to the ‘Neogene’. As noted above, this was the situation claimed by Paradis at Attinguié, and it also obtains at Anyama. ‘Terre de Barre’ itself is a curious phrase, coming from the Portuguese ‘terra barro’ meaning clayey soil (Houessou and Lang, 1979). It constitutes a poorly sorted reddish to ochre clayey sand where there are normally few or no pebbles >1 cm in diameter, and where >50% of the deposits are 39,000 BP), an intermediate MSA horizon (with an average radiocarbon date of 5575±65 BP), and an overlying occupation attributed to the Nok culture. Stratigraphically, the three different components were associated with a 221

The Middle Stone Age of Nigeria in its West African Context

lower, a middle, and an upper tin-wash, but the bases for this succession (not in itself unreasonable) were far weaker than at Adrar Bous. A further sequence concerns the MSA and the Sangoan. The very good stratified succession at Anyama reveals three components, one above the other, B (geological layer 13) microlithic LSA, C (geological layer 12) MSA, and D (geological layers 11-9) Sangoan. There is a TL date of 254,000±51,000 BP in geological layer 9, 1.5 to 2 metres below the lowest occurrence of Sangoan tools, which provides a terminus post quem for the entire sequence. The same succession is suggested for Asokrochona, with LSA in the Kpone Formation, MSA in the Nungua Formation, and Sangoan in the upper part and on the surface of the Asokrochona Formation, above the Tema Formation (attributed to the Tertiary) at the base. In all these cases, it would be desirable - if possible - to carry out further work, to supplement the OSL dating record already available at Ounjougou and elsewhere. Since they also form a prominent part of the record, it is not possible to ignore the radiocarbon dates currently available, whatever their limitations might be. In northern Nigeria, the date of 5440±100 BP from Zenabi has long been known. It was obtained on a large log near the base of the sequence at the site. Mention has already been made of the analogous date from Nok. In addition, in the course of this work, four more dates were obtained on wood from the alluvial tin workings at Banke. There are three from the upper tin wash in the range from 4710 to 4130 BP, and one from above that point of 2400±50 BP. The archaeological material from Banke is heterogeneous, but that is not the case at Zenabi, Mai Lumba, Yelwa, and Tibchi. The MSA in the light of the evidence summarised above could not be expected to be as young as that. The dates could simply be mistaken, or they could reflect real processes which have redeposited the material at a date subsequent to its manufacture and use. The author has taken the view that the latter interpretation is correct. So what were those processes? M.B. Thorp, working in the central part of the Liruei complex, identified two successive alluvial sequences of erosion, each commencing with a cassiterite bearing gravel, which he correlated with the lower and upper tin washes outside the ring dyke at Zenabi. In his view, these erosional phases will have occurred in the late Pleistocene and the Holocene in unstable ‘morphodynamic’ episodes, characterised by aridity, vegetation reduction, and increased surface run-off. He compared this scenario to the one identified by Burke and Durotoye in the Ibadan area, but he did not suggest any more specific date for these occurrences. A similar kind of explanation has been offered for the formation of the underbank gravels (graviers sous berge) which are a recurring feature in West Africa. According to Tricart, these deposits were formed in a semi-arid climate by the action of unstable channels or braided streams in a sometimes quite violent manner. The industry at Mayo Louti in northern Cameroun, comparable in many ways to those north of the Jos plateau, occurs in just such a deposit. It is surmounted by a yellowish sandy formation with calcareous nodules, which provided the material for radiocarbon dating, in the range from 6,210 to 17,860 BP, and this is considered to provide a terminus ante quem for the underbank gravels themselves. Whether by coincidence or not, calcareous concretions in the clayey sands (limons) above a ferruginous gravel horizon in the Dakar peninsula have also produced a maximum age of 16,900±250 BP, significant because (at least in Descamps’ interpretation) that horizon corresponds to the one at Tiémassas which yielded an MSA-oriented industry of that name. Not that all underbank gravels were necessarily of the same age. At Anyama in the Ivory Coast they have been dated to 8460±80 BP, in this case surmounted by silts (with polished stone axes) which are younger. In Sierra Leone, in the ‘broad shallow valleys’ examined by them, Thomas and Thorp detected a period of ‘fluvial scour’ between about 12,430 and 10,500 BP, although in ‘buried swamp’ deposits there were also six dates much older than this, all without archaeological correlates. In the existing state of knowledge, it would be idle to try to propose some over-arching scheme to explain all these features, which occur over a vast area, with differing local conditions. The erosional activity detected, and the climatic conditions which brought it about, is a subject which cries out for further study from a geological point of view, and is worth pursuing on its own account. Presumably however it will not succeed in throwing light on the time when the artefacts in such deposits were actually made, and for this purpose attention in the future must be directed to in situ occurrences such as were found at Ounjougou. Exactly the same kind of process has taken place with regard to the Nok culture in Nigeria, where the emphasis has shifted from the alluvial contexts in which the artefacts were first found, to more informative non-alluvial sites (Breunig, 2013). 222

West Africa: regional summary

Any problems of provenance notwithstanding, the archaeological assemblages attributed to the MSA in West Africa do not lack intrinsic worth and distinctive variability. Attention has been drawn to some rather idiosyncratic elements which characterise the assemblages from Zenabi and Tibchi, and the extent to which they differ from the industry discovered nearby at Mai Lumba. These differences can be expressed in a shorthand way by comparing their Levallois typological and broad facetting indices and their sidescraper percentages, as follows.

Tibchi

ILty

IR

IFl

15.22

54.35

19.81

Zenabi

14.13

54.35

18.65

Mai Lumba

55.80

27.90

31.80

It was suggested previously, on the basis of figures provided by Marliac, that there was a close similarity in terms of ILty between Mayo Louti, in northern Cameroun, and Mai Lumba. If Digara’s much increased totals are to be believed, the resemblance between the two in terms of this index is less than hitherto calculated, but there is still a general resemblance between this site and those north of the Jos plateau, such that one can speak of them as forming a single ‘province’. Digara pointed out that 55.6% of the artefacts from Mayo Louti were very worn, similar to Mai Lumba, but not Zenabi. It is noticeable that at Saminaka, somewhat removed from the other sites north of the Jos plateau, non-quartz material (principally rhyolite) was mainly used for the tools and cores, even though quartz accounted for the majority of the artefacts as a whole. Their relatively small size is attributed to greater intensity of use, which in turn reflects their distance from the raw material sources, which were evidently very attractive for all the inhabitants of this area. The small published assemblage from Ajibode, in southern Nigeria, is exclusively on quartz and quartzite and (while its classification is uncertain) it bears little resemblance to the sites in the north. At Adrar Bous, J.D. Clark also considered that a distinct ‘regional Middle Palaeolithic tradition’ could be detected, further evidence for which could be found at Bilma and Seggédim. At the Lost Valley site, a clearly stratified Late Acheulean occupation was located which showed definite evidence for the use of the Levallois technique, hence Clark concluded that it was ‘anticipatory of, or even transitional to, the Early Middle Stone Age’. The MSA assemblages in the three stratified locations referred to above, and in the Main Wadi, showed strong Levallois characteristics and employed exclusively local raw materials, mainly rhyolite and grey vitric tuff. The most representative artefacts characteristic of the Aterian came from Lookout Hill and from Yellowstone Hill between the Main Wadi and the Valley of the Lake, which also contained stratified material. Clark emphasised the presence of bifacial foliate points and core-axes, as well as a few tanged points, and the prevalence of the Levallois technique. Typologically the Aterian was distinct from the preceding MSA, and (no doubt because of the presence of the first two elements) Clark suggested that it had affinities to the Central African Lupemban. There was also a significant shift in raw material usage, with a new reliance on green fine-grained silicified vitric tuff, which must have come from a considerable distance. Nonetheless, in Clark’s opinion, the Aterian at the site was undoubtedly ‘a direct descendant from the Middle Palaeolithic Mousterian’. Bifacial foliate points (as well as tanged points) recur in the Aterian at Seggédim, and according to Tillet there is a considerable Levallois component as well, with a Levallois technical index (IL) of 35.02 and a Levallois typological index (ILty) of 62.68. All of this does hark back to Adrar Bous, but Clark’s over-arching scheme of continuity through time and wide ranging spatial connections no doubt will require further testing. The record from Senegal, as presently known, displays considerable variability. Conventionally, the MSA in this country has been divided into two facies, the ‘Mousteroid’ and the ‘Tiémassassian’. The former includes such sites as Bargny, Lguer, Bafoulabé, and Takoutala, geographically quite far from each other. Most of the material inventoried is from the surface, and the suggestion is that the first two may well have functioned as workshop sites. The assemblages in general are quite simple, with a limited use of the Levallois technique, as at Takoutala which has a Levallois typological index (ILty) of 12.31 and a 223

The Middle Stone Age of Nigeria in its West African Context

sidescraper index (IR) of 34.62. Descamps’ excavations at Tiémassas in 1969 established the fact that the industry did originate in a layer of ferruginous gravel but the quantity of material recovered was insufficient to permit a convincing description of the industry as a whole. The published inventories therefore rely on the surface material, either that collected by R. Guillot or that deposited at IFAN. It is entirely clear that this material is to some extent mixed, hardly surprising since it comes from a very large area, but the bulk of it can be regarded as definitive of the ‘Tiémassassian’. Typologically the presence of bifacial and tanged points is considered to be decisive, and there is a considerable Levallois component. The Levallois typological index (ILty) in the Guillot collection (as recalculated) comes to 23.66, and the percentage of Levallois cores in the IFAN collection (as recently restudied) comes to 48.91. Apart from Tiémassas itself, sites comparable to it include Sébikotane, and a number of smaller localities on the Dakar peninsula, which are important stratigraphically. The artefacts from Sébikotane are very finely made, making good use of the flint which is characteristic for the area. Insofar as reliance can be placed on ‘type fossils’, the question clearly arises as to the relationship between the ‘Tiémassassian’ and the Aterian, with the possibility that the former is no more than a local variant of the latter. These occurrences do not exhaust the variability so far recorded in Senegal. It has been claimed that an ‘Evolved Palaeolithic’ exists in the south and east of the country, at Badoye and at various localities along the Falémé river. Badoye is seemingly quite idiosyncratic, characterised by so-called ‘rabots nucléiformes’, and a more thorough going investigation would be needed to clarify its position. Camara and Duboscq earlier suggested that several sites on the Falémé (including Diboli, Guédékou, Doundé, Oubol, Marsa, and Kayes) should be classified as ‘Evolved Palaeolithic’ at a time which they thought could be between about 27,000 and 13,500 BP. The pieces they illustrated are mostly bifacial points. It is noteworthy that the work now under way in the Falémé river valley has also recovered a number of bifacial points, not necessarily in this time bracket, so this will be an ongoing subject of research. At Toumboura III a stratified assemblage with over 4000 artefacts has been recovered, so it is the detailed study of sites such as this that will enable the characteristics of the MSA in Senegal to be placed upon a firmer footing. The same goes for West Africa in general. As shown above, the region is not lacking in characteristic assemblages, and there are clear indications of significant variability, but the elucidation of what these differences might mean depends upon finding sites in acceptable stratigraphic context. If questions of terminology are not particularly vexing in relation to the MSA in West Africa, the same cannot be said of the Sangoan. Attention has been drawn to some of these difficulties in relation to the main sites in Ghana, the Ivory Coast, and Cameroun, as well as in Nigeria. The differences in terms of analysis between Oliver Davies, Nygaard and Talbot, and Andah at Asokrochona are all too evident, and no doubt reflect, as Andah said, ‘a large dose of subjectivity’ on the part of those who tried to classify the material. This is despite the fact that, as indicated in the Introduction, they were all to a considerable degree influenced by the system previously devised for the Sangoan by J.D. Clark and M.R. Kleindienst (1974). At Anyama, Liubin and Guédé specifically rejected ‘core-axes’ in favour of ‘picks’, of which they suggested there was more than one variety. Picks and bifaces also loom large in the descriptions given of localities attributed to the Sangoan in Cameroun and Nigeria. It would be desirable for a single team to reexamine all these collections, applying a uniform system of analysis, to come to some kind of consistent judgement as to what they represent, and what the differences of emphasis are between them. This is certainly an easier procedure to suggest than to carry out. In the meantime, further developments have taken place elsewhere which have modified our picture of the Sangoan, both in terms of the artefacts which are characteristic of it, and its chronology. Kalambo Falls has always played a key role in our understanding of the African Palaeolithic, including the Sangoan. The main bulk of J.D. Clark’s fieldwork was carried out in 1953-1966, but it was only in 2001 that his final volume was published. For the first time, it contained a detailed account of the Middle and Earlier Stone Age occurrences, as well as a reconsideration of the site as a whole (Clark, 2001). For a long time, it had been supposed that two radiocarbon dates of 60,300±750 and 61,200±1280 BP for the Upper Acheulean provided an acceptable starting date for the sequence. Now, as Clark said, it is clear that ‘most or all of the Pleistocene radiocarbon dates’ at the site are ‘almost certainly much too young’ 224

West Africa: regional summary

and should be rejected. Curtis McKinney (in Clark 2001: 665-674) obtained uranium series dates on wood of 76,000±10,000 BP for the Sangoan and two dates of 182,000 ± 10/16,000 BP for the Upper and Lower Acheulean; but Clark considered these dates also to be ‘much too young’. Relying on sites elsewhere in Africa dated by K-Ar, uranium series, and electron spin resonance (ESR), he suggested that the Sangoan might more appropriately be dated to about 250,000 BP and the end of the Acheulean to about 300,000 BP. Recent surveys (Barham and Mitchell, 2008; Basell, 2010) confirm the plausibility of this suggestion. At Anyama, the TL date obtained of 254,000±51,000 BP, serving as a terminus post quem for the Sangoan, is obviously compatible. In the opinion of the Liubin and Guédé, this industry, although it has some local characteristics, does constitute part of the broader Sangoan Industrial Complex as known elsewhere in Africa. The same conclusion was reached by this author in regard to the industry from Asokrochona. Pending the conclusion of any new comprehensive study, it would be appropriate to regard the West African Sangoan as a whole as a distinctive cultural entity ‘closely allied to or possibly derivative from the Late Acheulean’ (Allsworth-Jones, in ed. Sutton, 1982). The final details clearly remain to be worked out. A particularly interesting site, which has recently been discovered, is that of Sai Island, in Northern Sudan (Van Peer et al., 2003, 2004; Rots and Van Peer, 2006; Rots et al., 2011). In a Middle Pleistocene depression south of Jebel Adu - which functioned as a sediment trap - five well stratified archaeological units are superimposed, attributed to the Nubian Complex, the Lupemban, the Middle and Lower Sangoan, and the Acheulean respectively. There are three OSL dates for the site: 152,000±10,000 BP at the level of the Lupemban, 182,000±20,000 BP just above the Middle Sangoan, and 223,000±19,000 BP just above the Acheulean (Van Peer et al., 2003, Fig. 2). The archaeological material is said to be similar to that from Kalambo Falls in several respects. In other respects, however, this is an unusual site. In the Lower Sangoan occupation horizon, there are many yellow and red ochre lumps which could have acted as colouring material, as well as stones used for grinding them up. Phytoliths and starch granules detected on some cobbles in the Middle Sangoan suggest that they were used to process plant tissues. ‘Core-axes’ were much in evidence in the Lupemban and Sangoan horizons, and detailed use-wear analyses carried out by Veerle Rots have enabled us to get a much better idea of how these hitherto rather enigmatic artefacts - so crucial for the definition of the Sangoan - were made and used. Initially, 48 core-axes were identified in the Middle Sangoan, but of these the majority never came into use, mostly due to damage during manufacture. On the basis of 14 which did have an extended life-cycle, Rots was able to determine that hafting wear was present on both left and right edges, demonstrating that these pieces had been fixed in wooden shafts, transverse to the main axis, thus acting as percussion instruments not projectiles (Rots et al., 2011, Fig. 2). The butts were undamaged, but the distal ends had signs of use-wear. What kind of use could not be determined, but it may very well not be the woodworking with which these tools have traditionally been associated. Similar results were obtained for the Lower Sangoan and the Lupemban. There are good reasons for thinking that this was a special kind of workshop, which bears witness to a complicated provisioning network. Hence, these studies have not only pushed back the chronological boundaries of the Sangoan and the MSA, they have also shown the degree of complexity existing in the societies of that time, and demonstrated the kind of techniques that are needed to research it convincingly. References Barham, L., Mitchell, P. 2008. The First Africans. Cambridge: Cambridge University Press. Basell, L. 2010. Middle Stone Age Sangoan-Lupemban Lithic Assemblages of Africa. In ed. P. AllsworthJones, West African Archaeology: New developments, New perspectives: 15-27. Oxford: BAR International Series, 2164. Breunig, P. (ed.). 2013. Nok: Ein Ursprung Afrikanischer Skulptur. Frankfurt: Africa Magna Verlag. Clark, J.D., Kleindienst, M.R. 1974. The Stone Age cultural sequence: terminology, typology and raw material. In J.D. Clark (ed.), Kalambo Falls Prehistoric Site II, The late prehistoric cultures: 71-106. Cambridge: Cambridge University Press. Clark, J.D. 2001. Kalambo Falls Prehistoric Site, volume III. Cambridge: Cambridge University Press. Rots, V., Van Peer, P. 2006. Early evidence of complexity in lithic economy: core-axe production, hafting and use at Late Middle Palaeolithic site 8-B-11, Sai Island (Sudan). Journal of Archaeological Science, 33: 360-371. 225

The Middle Stone Age of Nigeria in its West African Context

Rots, V., Van Peer, P., Vermeersch, P.M. 2011. Aspects of tool production, use, and hafting in Palaeolithic assemblages from Northeast Africa. Journal of Human Evolution, 60: 637-664. Van Peer, P., Fullagar, R., Stokes, S., Bailey, R.M., Moeyersons, J., Steenhoudt, F., Geerts, A., Vanderbeken, T., De Dappere, M., Geus, F. 2003. The Early to Middle Stone Age Transition and the Emergence of Modern Human Behaviour at site 8-B-11, Sai Island, Sudan. Journal of Human Evolution, 45: 187-193. Van Peer, P., Rots, V., Vroomans, J.M. 2004. A story of colourful diggers and grinders: the Sangoan and Lupemban at Site 8-B-11, Sai Island, Northern Sudan. Before Farming, 3: 139-166.

226

Chapter 5

A wider perspective The last few decades have witnessed an unprecedented amount of work devoted to the environmental, archaeological, and evolutionary record as revealed by research on the African continent. The West African story, relatively modest as it is, has to be seen in the light of these broader achievements. Attention can be drawn only to some of the publications summarising these developments. Climate In the first place, one may note the comprehensive reconstructions advanced for the changes in climate over the period in question, changes recorded above all in East Africa (Maslin and Trauth 2009, deMenocal 2011, Maslin et al. 2014, 2015), though they reflect the global variations in the earth’s axis and orbit referred to in the Introduction. Maslin and his colleagues have detected a long term drying and cooling trend over the course of the Pleistocene, punctuated by shorter alternating episodes of extreme humidity and aridity, against a background of no less significant tectonic activity. They have named their interpretation a ‘pulsed climate variability hypothesis’. The humid phases are marked above all by the presence of large deep lakes in East Africa, lakes which periodically may have dried up completely. In particular, three lake high stands have been detected at approximately 2.7-2.5, 1.9-1.7, and 1.1-0.9 million years ago. The whole model has been pictured in a ‘snapshot’ provided by deMenocal (Figure 315). In this diagram, the various proxies employed to trace the changes over time are indicated, as are the three lake high stands referred to. Also shown are changes in the types of hominin present (and some archaeological markers) during the same period, the authors’ contention being that hominin evolution was decisively affected by the changing environment, although they caution that their hypothesis does not actually provide a mechanism through which the evolutionary process may have occurred.

Figure 315. African evolutionary and paleoclimatic changes (after de Menocal, 2011).

227

The Middle Stone Age of Nigeria in its West African Context

The greater part of the ‘snapshot’ at Figure 315 relates to a period before that with which we are mainly concerned, but there is no doubt that the trends detected earlier continued in being. Reference has already been made in the Introduction to the African Humid Period (AHP), the principal part of which may be dated to c. 9,000 - 5,500 years ago. Research now suggests that a ‘green Sahara’, marked by an extensive interconnected hydrological system, may have existed at about 120,000 years ago as well (Osborne et al. 2008, Drake et al. 2011). This time corresponds to Marine Isotope Stage (MIS) 5e, otherwise known as the Last Interglacial. Osborne and his colleagues suggest that at that time there was an influx of fresh water into the Mediterranean from North Africa, as measured by oxygen isotope data (δ18O) on marine planktonic foraminifera and on neodymium (Nd) on freshwater molluscs from ancient river courses linked to the Fezzan basin in Libya and ultimately to lake Megachad (Osborne et al. 2008, Fig. 1). OSL dates obtained on beach ridges in the Chad basin of 114±14 and 125±12 thousand years ago support this hypothesis, and it is suggested that apart from this ‘eastern’ route there also have been another one further west (Drake et al. 2011, Figs.1 and 4). These more favourable climatic conditions, it is suggested, may account for the presence of MSA and Aterian artefacts in areas which are now desertic, and more broadly have provided a possible corridor for the migration of early modern humans ‘out of Africa’, hence they are of direct relevance for the interpretation of the archaeological and human evolutionary record. Archaeology So far as the archaeological and evolutionary records are concerned, they are closely intertwined. As I pointed out some time ago (Allsworth-Jones, 1993: 21), ‘Whereas in Europe the transition from Middle to Upper Palaeolithic and the replacement of Neanderthal by anatomically modern humans appear to be synchronous events, in Africa this is not the case. Neanderthals as such were not present in Africa, and if the ‘Out of Africa’ model is correct, the ancestors of anatomically modern humans must have made their appearance in a Middle Stone Age context before 100,000 years ago’. This chronological estimate was based among other things on the evidence from Klasies River Mouth and Border Cave in South Africa (Allsworth-Jones, 1993, Fig. 2). These arguments were carried a great deal further and in more detail by McBrearty and Brooks in their influential article (2000) concerning the ‘revolution that wasn’t’ in Africa. Since then, more substantial advances have been made, only the most striking of which can be mentioned here. For convenience sake, the more strictly archaeological evidence will be summarised first, followed by that concerned primarily with human evolution, in a roughly geographical manner. We may begin the story in South Africa (Lombard 2012, Wadley 2015, Ziegler et al. 2013). As Wadley says, radiometric measurements (OSL, TL, U-series, and AAR, excluding radiocarbon dating) demonstrate that at least 13 MSA sites in this part of the continent have ages older than 100,000 years (Wadley 2015, Table 1). They include Blombos, Border Cave, Diepkloof, Florisbad, Klasies River Mouth, Pinnacle Point, Wonderwerk, Wonderkrater, and Ysterfontein, among others. In fact it is considered that the MSA here may have commenced some 300,000 years ago, coming to an end at about 30,000 BP. It has been ascertained that the Stillbay and Howieson’s Poort variants of the MSA (two of the best known in South Africa) consistently follow each other, the Stillbay variant being dated to c. 75.5-67.8 and the Howieson’s Poort to c. 65-60 thousand years ago. During this time, as is now well known, a number of striking innovations can be traced in the record, notably ochre processing, with several thousand pieces including incised pieces, found at Blombos and Sibudu. Its use was probably practical, as an adhesive, apart from its decorative function. Other innovative traits include ostrich eggshell beads and perforated marine shells, and it is thought that heat treatment may have been applied in the manufacture of some stone tools, which may also have been hafted. All these traits are regarded as indicative of ‘complex cognition’ if not of ‘cultural modernity’. Since the idea is often advanced that changes in the environment facilitated or stimulated changes in material culture, a study by Ziegler et al. (2013) of a marine core off the coast of Natal near the mouth of the Great Kei river is relevant. According to these authors, the ratio of iron to potassium (Fe/K) in sediments derived from the river reflects dry (low) or wet (high) conditions respectively. On that assumption, it is considered that both the Stillbay and the Howieson’s Poort variants coincided with periods of greater humidity, which may have provided a favourable environment for their development (Ziegler et al. 2013, Fig. 4). In view of the fact that specific aspects of climatic change have 228

A wider perspective

also been suggested as a determinant for the development of the MSA in East Africa (Owen et al. 2018), this suggestion is certainly worth bearing in mind. In East Africa, new evidence from Olorgesailie provides a convincing demonstration that the MSA at that location began certainly before 305,000, and probably at not less than 320,000, years ago (Brooks et al. 2018, Deino et al. 2018, Potts et al. 2018, Gibbons 2018). Five sites have been identified in the Olkesiteti Member of the Oltulelei Formation, which is separated by an erosional unconformity from the underlying Olorgesailie Formation. 40Ar/39Ar dated tuffs above site BOK-1E provide the minimum age mentioned, and these age determinations are supported by others. The Acheulean in the Olorgesailie Formation came to an end at about 499,000 years ago, hence it is separated from the MSA by an erosional hiatus lasting some 180,000 years. An analysis of the MSA environment reveals that this was a ‘dynamic palaeolandscape’, with much channel cutting and filling, in an open arid setting. There was practically no faunal overlap with the preceding Acheulean. Evidence for climatic change in the period intervening between the Acheulean and the MSA is provided by the results of a coring programme conducted on the sediments of Lake Magadi, also in Kenya (Owen et al. 2018). The trend indicated is one of progressive aridification, and the suggestion is that this together with considerable variability may have led to the emergence of the MSA. Be that as it may, the MSA is certainly different from the Acheulean, not only in the tool types represented, but also in terms of the raw materials employed. A proportion of the artefacts are made of obsidian, which must have come from a distance estimated at c. 25-50 km, and this is taken to indicate the existence of significant exchange network. The early dates from Olorgesailie are backed up by some others, notably from Kapthurin and Gademotta. So far as the later occurrences of the MSA in East Africa are concerned, Tryon and Faith (2013) emphasise their variability and state that there is some evidence for change over time, but no unique archaeological signal to use as a marker for an ‘Out of Africa’ scenario. In their view, such variability as there is reflects local factors such as site function and raw material. In the Nile Valley on the other hand, the prevailing narrative suggests that there was considerable movement over time, and that there may indeed be indications of links elsewhere (Wurz and Van Peer 2012). Leaving aside the Sangoan and Lupemban claimed to exist at Sai Island for the moment, attention focuses on the Nubian Complex (defined among other things by the special cores referred to in the Introduction) which is said to have developed at least by the Last Interglacial. An OSL date of 118±8 thousand years from Sodmein cave supports this attribution, whereas the Nubian Complex (sensu stricto) is said to continue through much of the ensuing period, as shown at Activity Phase III at Taramsa 1 where there is an OSL date of 78±5.6 years ago. Activity Phases IV and V at Taramsa are associated respectively with an industry termed Taramsan and a so-called Lower Valley Nile Complex (LNVC). The former is described as having a volumetric production system, which leads on to fully Upper Palaeolithic modes of expression as shown in Activity Phase VI at Taramsa, associated with OSL dates of 33,200 BP (Barham and Mitchell 2008: 236). Some details of what happened next in Central Africa are provided by Cornelissen (2002). In her view, both the technological traditions associated with the Sangoan and the Lupemban, and with a microlithic LSA, ran parallel during the period from 40,000 to 12,000 BP, i.e., both before and after the Late Glacial Maximum (LGM) (Cornelissen 2002, Table 1 and Fig. 9). Despite the problems of post-depositional bioturbation which have been identified at some of the sites, she is inclined to give credence to the relevant radiocarbon dating evidence from such Lupemban sites as Maboué-5, Dimba, Hinda, and Loukoko II, which span this time period. They contrast with LSA sites which do the same, as 230

A wider perspective

witnessed by early dates from Matupi (>40,700 and 33,000±1900 BP) and Shum Laka (31,700±750 BP). A possible explanation for this is provided in the sense that, whereas in other parts of Africa there were drastic alterations in the environment, in Central Africa the undoubted changes which took place in the later Pleistocene, affecting the extent of forest cover, still allowed the inhabitants access to abundant resources. Whether this argument could be extended to West Africa remains to be seen. Human Evolution As stated already, the archaeological and human evolutionary records in Africa are closely intertwined, and as a result of recent discoveries that is more true than ever before. In the first place, mention must be made of the site of Jebel Irhoud, in Morocco (Hublin et al. 2017, Richter et al. 2017, Stringer and Galway-Witham 2017). Originally excavated in 1961-1969, it has now been re-excavated since 2004, with the result that it has produced more human skeletal remains in a context considerably older than originally estimated. Traces of six individuals were recovered during the earlier excavations, but that total has now risen to 22. Most of these are fragmentary, but they include the well known Irhoud individuals 1, 2, and 3 from the earlier excavations, and 10 and 11 (a cranium and a mandible) from the new excavations. The cranium is said to reveal a ‘mosaic’ of features, with an essentially modern facial morphology but an elongated (rather than globular) braincase. As a whole, this specimen (and the others from the site) can definitely be classified as an early Homo sapiens, perhaps to be bracketed together with Florisbad in South Africa as ‘pre-modern’, rather than ‘fully modern’. It has been dated to 315±34 thousand years ago, compared to Florisbad at 259±35 thousand years ago. The date from Irhoud is a TL date obtained on fire heated flint artefacts found in association with the human remains in layer 7 at the site. The artefacts are classified as MSA on the basis of a dominant Levallois technology, with no Acheulean or Aterian admixture. As the authors say, this discovery means that the emergence of the MSA and of our species appear to be two events which are close together in time, whereas previously the former seemed to be a good deal older than the latter. Thus, ‘fully modern’ Homo sapiens specimens are somewhat later than Irhoud, coming from such sites as Omo Kibish at 195±5 and Herto in the range from 160 to 154 thousand years ago. In addition, one might mention the newly discovered site of Halibee, north of Herto, also in Ethiopia (Morgan et al. 2009, Brasil and Hlusko 2018). Here artefacts made from freshly extruded obsidian have been dated by 40Ar/39Ar to 106±20 thousand years ago. They are part of an industry classified as a typical MSA, and are underlain by a glass shard tuff also dated to 148±34 thousand years ago. From the same MSA layer, unusually, a partial post-cranial skeleton attributed to a young male adult has been found, and is regarded as representative of a population ‘ancestral or closely related to’ anatomically modern humans. Clearly, the dated specimens of early Homo sapiens in Africa are relevant to discussion concerning the ‘Out of Africa’ model of human evolution. Hitherto, it is Skhul and Qafzeh (dated to a period from 90 to 120 thousand years ago) which have provided the best candidates for a related (and presumably descendant) population in the Near East. Now there is another candidate from Misliya, on the western slopes of Mount Carmel, which is considerably older than that (Hershkovitz et al. 2018, Stringer and Galway-Witham 2018). Unit 6 at the site has been dated by a variety of techniques (TL, U-Th, and US-ESR) to a time span of 194-177 thousand years ago. The archaeological material in this unit is characterised as Levallois dominated, comparable to Tabun D. A human maxilla and dentition from this location is said to be Homo sapiens in character, certainly not Neanderthal, and generally comparable to Jebel Irhoud. The fact that there are at least two different age ranges now for Homo sapiens fossils in the Near East which can be linked to Africa raises the question as to whether there was a single dispersal or multiple dispersals from the continent. This question was considered by Reyes-Centeno et al. (2015) in an article published before the latest discoveries. Their study was based on a craniometric data set which included Herto, Skhul, Qafzeh, and Nazlet Khater as well as >2000 recent human crania from a wide geographical area. They concluded that there were likely to have been two waves of migration (a multiple dispersal with isolation model) where the native Australians, Papuans, and Melanesians alone represent traces of the first wave. While this model may still be valid, it will presumably require modification in the light of the latest available information (cf. Stringer and Galway-Witham 2018, Figure ‘Earliest modern human migrations from Africa’). 231

The Middle Stone Age of Nigeria in its West African Context

Genetics has also been part of the story, ever since the annunciation of the ‘Mitochondrial Eve’ at a conference in Cambridge in 1987 (Stoneking and Cann, 1989). The authors came to the conclusion that the common ancestor of all modern human populations lived in sub-Saharan Africa somewhere between 50,000 and 500,000 years ago. Things have since moved on beyond mitochondrial DNA alone, but the basic premise advanced by Stoneking and Cann has remained the same. In general, the results of research have tended to be produced in the form of tree diagrams, attempting to show divergence estimates for the various strands of humanity as presently constituted. Two such studies may be mentioned. Campbell and Tishkoff (2010) remarked on the fact that indigenous Africans today are characterized by high levels of genetic diversity. On the basis of genome-wide polymorphic markers, they suggested that there were no less than 14 genetically distinct ancestral population clusters in the continent, with the implication that anatomically modern humans evolved in Africa around 200,000 years ago, and a population divergence took place around 150,000 years ago (Campbell and Tishkoff 2010, Fig. 2). A much longer time scale is suggested by Schlebusch et al. (2017). Their study concentrated upon the remains of seven deceased individuals from KwaZulu-Natal, three Khoesan hunter-gatherers estimated to be 2000 years old and four Bantu farmers estimated to be about 300-500 years old. Their conclusion is that the Khoesan population ‘harbours the greatest level of diversity due to capturing the deepest split among humans’, the split-time (and hence the estimated latest time for the emergence of Homo sapiens) being put at approximately 350-260 thousand years ago (Schlebusch et al. 2017, Fig. 3). As the authors say, this estimate is consistent with the fossil evidence as currently known (and as summarized above). It should be pointed out however that the methodology, and hence the conclusions, of these authors have been challenged by the participants at a workshop held in Oxford in 2016 (Scerri et al. 2018). A starting point for the discussion, in Chris Stringer’s phrase, was that a version of ‘multiregionalism’ may have been operative in the continent at the time in question. This point was already made by Richter et al. (2017) in relation to Jebel Irhoud, when they cautioned against ‘favouring one region over another’ when considering modern human origins. From the genetic point of view, the criticism advanced is that tree diagrams such as mentioned above operate on the assumption that there was a single panmictic population, i.e., one where there was a random pattern of mating within that population. That might not be the case if in fact there were multiple sub-populations, isolated from each other but connected in part by migration (Scerri et al. 2018, Fig. 3d). All these avenues no doubt will be further explored. In the meantime, one might note a cautionary observation made at the ESHE meeting in Portugal in 2018 that there may yet be surprises in store, since repetitive exploration in areas where fossils have been found (particularly East Africa) may be distorting the picture, just as (following Karl Popper) ‘counting more white swans does not increase the validity of the hypothesis that all swans are white’ (Bons et al. 2018). A summary ‘snapshot’ of the current picture in Africa as a whole, as revealed by the recent research summarized above, has been provided by Gibbons (2018) (Figure 316). Conclusion It is evident that, as things stand at present, West Africa would find it difficult to compete with some of these achievements, as mentioned by other recent writers on the subject (Taylor 2014, Scerri 2017b, Chevrier et al. 2018), but it should by no means be ignored. The artefactual evidence for the MSA is abundant, even if on occasions it lacks good context, and as illustrated by J.D. Clark (1992) it fits well into the general pattern of regional variation in the continent (Figure 317). To the north, it seems that a rather sharp frontier marks it off from the Aterian, the southernmost outliers of which examined here being Tiémassas, Adrar Bous, and Seggédim. Adrar Bous provides important evidence of the stratigraphic relationship between the two, the MSA being consistently prior to the Aterian. The relationship between the Sangoan of West Africa and the Sangoan-Lupemban of Central Africa is far less certain. In West Africa assemblages called Sangoan have been defined by different authors in different ways, and the degree to which they are really comparable to their equivalents in Central Africa is unsure. Nonetheless at Anyama, and less certainly at Asokrochona, the Sangoan has been found beneath stratified MSA occurrences, the opposite situation to that prevailing with the Aterian. Satisfactory dating of the MSA in West Africa has long been elusive, in part because of contextual difficulties, in part because of the inherent limitations of 232

A wider perspective

Figure 316. The roots of modern human behaviour (after Gibbons, 2018).

Figure 317. Middle Palaeolithic-Middle Stone Age regional variants in Africa (after Clark, 1992, Fig. 1).

233

The Middle Stone Age of Nigeria in its West African Context

radiocarbon dating. Newer methods, notably OSL and TL, are helping to remedy the situation, and the antiquity of the Sangoan and MSA has been confirmed at several sites, including Anyama, Birimi, Bilma, Ounjougou, and the Falémé river. A full and convincing palaeoclimatic record is now available from Lake Bosumtwi, covering much of the Quaternary, as good as that anywhere else on the continent, though as yet unlinked to archaeological occurrences. No human remains (and virtually no fauna) have been recovered from MSA contexts in West Africa, but it is not out of place here to mention the site of Iwo Eleru (Allsworth-Jones et al. 2010, Harvati et al. 2011). Excavated by Thurstan Shaw in 1965, the site produced remains of a human burial, dated on the basis of associated charcoal to 11,200±200 BP. The archaeological context is exclusively LSA, but initial results showed that the cranium had some surprisingly ‘archaic’ features. A U-series analysis now done on one of the long bones suggests that the age of the specimen might be in the range from 11.7±1.7 to 16.3±0.5 thousand years ago, so definitely Late Pleistocene. A new morphometric analysis of the cranium [as revised] confirms the earlier results, showing that Iwo Eleru’s nearest neighbours (in terms of Mahalanobis D2 distance) are the skulls from Qafzeh and Skhul. Moreover, Iwo Eleru is not alone, since remains in some ways comparable have been located at Ishango in the DRC (Crevecoeur et al. 2016). This site has produced fragmentary evidence of at least 12 individuals with LSA artefacts in a find horizon with six radiocarbon dates in the range from 19,540 to 25,570 BP. The skeletal remains are classified as modern human, but they have ‘unusual morphometric characteristics’, including a conformation of the inner ear which is closer to Qafzeh and Skhul than to recent examples. The precise meaning of these occurrences remains to be elucidated, but Harvati et al. (2011) concluded that so far as Iwo Eleru is concerned, the results are ‘in accordance with suggestions of deep population substructure in Africa and a complex evolutionary process for the origin of modern humans’. It would be hard to find a more resounding endorsement of the importance and relevance of West Africa than this. References Climate deMenocal, P.B. 2011. Climate and Human Evolution. Science, 331 (540): 540-542. Drake, N.A., Blench, R.G., Armitage, S.J., Bristow, C.S., White, K.H. 2011. Ancient watercourses and biogeography of the Sahara explain the peopling of the desert. Proceedings of the National Academy of Sciences of the United States of America, 108: 458-462. Maslin, M.A., Trauth, M.H. 2009. Plio-Pleistocene East African Pulsed Climate Variability and Its Influence on Early Human Evolution. In (ed.) F.E. Grine, J.G. Fleagle, R.E. Leakey. The First Humans - Origin and Early Evolution of the Genus Homo, 151-158. Springer Science + Business Media B.V. Maslin, M.A., Brierley, C.M., Milner, A.M., Shultz, S., Trauth, M.H., Wilson, K.E. 2014. East African climate pulses and early human evolution. Quaternary Science Reviews, 101: 1-17. Maslin, M.A., Shultz, S., Trauth, M.H. 2015. A synthesis of the theories and concepts of early human evolution. Philosophical Transactions of the Royal Society, B 370: 20140064. Osborne, A.H., Vance, D., Rohling, E.J., Barton, N., Rogerson, M., Fello, N. 2008. A humid corridor across the Sahara for the migration of early modern humans out of Africa 120,000 years ago. Proceedings of the National Academy of Sciences of the United States of America, 105 (43): 16444–16447. Archaeology Barham, L., Mitchell, P. 2008. The First Africans. Cambridge: Cambridge University Press. Barton, R.N.E., Bouzouggar, A., Hogue, J.T., Lee, S., Collcutt, S.N., Ditchfield, P. 2013. Origins of the Iberomaurusian in NW Africa: New AMS radiocarbon dating of the Middle and Later Stone Age deposits at Taforalt Cave, Morocco. Journal of Human Evolution, 65: 266-281. Barton, R.N.E., Lane, C.S., Albert, P.G., White, D., Collcutt, S.N., Bouzouggar, A., Ditchfield, P., Farr, L., Oh, A., Ottolini, L., Smith, V.C., Van Peer, P., Kindermann, K. 2015. The role of cryptotephra in refining the chronology of Late Pleistocene human evolution and cultural change in North Africa. Quaternary Science Reviews, 118: 151-169. Brooks, A.S., Yellen, J.E., Potts, R., Behrensmeyer, A.K., Deino, A.L., Leslie, D.E., Ambrose, S.H., Ferguson, J.R., d’Errico, F., Zipkin, A.M., Whittaker, S., Post, J., Veatch, E.G., Foecke, K., Clark, J.B. 2018. Longdistance stone transport and pigment use in the earliest Middle Stone Age. Science, 360 (6384): 90-94. 234

A wider perspective

Chevrier, B., Huysecom, E., Soriano, S., Rasse, M., Lespez, L., Lebrun, B., Tribolo, C. 2018. Between continuity and discontinuity: An overview of the West African Palaeolithic over the last 200,000 years. Quaternary International, 466: 3-22. Clark, J.D. 1992. African and Asian perspectives on the origins of modern humans. Philosophical Transactions of the Royal Society of London, B, 337: 201-215. Cornelissen, E. 2002. Human Responses to Changing Environments in Central Africa Between 40,000 and 12,000 B.P. Journal of World Prehistory, 16: 197-235. Deino, A.L., Behrensmeyer, A.K., Brooks, A.S., Yellen, J.E., Sharp, W.D., Potts, R. 2018. Chronology of the Acheulean to Middle Stone Age transition in eastern Africa Science, 360 (6384): 95-98. Douka, K., Jacobs, Z., Lane, C., Grün, R., Farr, L., Hunt, C., Inglis, R.H., Reynolds, T., Albert, P., Aubert, M., Cullen, V., Hill, E., Kinsley, L., Roberts, R.G., Tomlinson, E.L., Wulf, S., Barker, G. 2014. The chronostratigraphy of the Haua Fteah cave (Cyrenaica, northeast Libya). Journal of Human Evolution, 66: 39-63. Foley, R.A., Maíllo-Fernández, J.M., Lahr, M.M. 2013. The Middle Stone Age of the Central Sahara: Biogeographical opportunities and technological strategies in later human evolution. Quaternary International, 300: 153-170. Gibbons, A. 2018. Complex behavior arose at dawn of humans. Science, 359 (6381): 1200-1201. Lombard, M. 2012. Thinking through the Middle Stone Age of sub-Saharan Africa. Quaternary International, 270: 140-155. McBrearty, S., Brooks, A.S. 2000. The revolution that wasn’t: a new interpretation of the origin of modern human behavior. Journal of Human Evolution, 39: 453-563. Owen, R.B., Muiruri, V.M., Lowenstein, T.K., Renaut, R.W., Rabideaux, N., Luo, S., Deino, A.L., Sier, M.J., Dupont-Nivet, G., McNulty, E.P., Leet, K., Cohen, A., Campisano, C., Deocampo, D., Shen, C.C., Billingsley, A., Mbuthia, A. 2018. Progressive aridification in East Africa over the last half million years and implications for human evolution. Proceedings of the National Academy of the United States of America, https://doi.org/10.1073/ pnas.1801357115 Potts, R., Behrensmeyer, A.K., Faith, J.T., Tryon, C.A., Brooks, A.S., Yellen, J.E., Deino, A.L., Kinyanjui, R., Clark, J.B., Haradon, C.M., Levin, N.E., Meijer, H.J.M., Veatch, E.G., Owen, R.B., Renaut, R.W. 2018. Environmental dynamics during the onset of the Middle Stone Age in eastern Africa. Science, 360 (6384): 86-90. Scerri, E.M.L. 2013a. The Aterian and its place in the North African Middle Stone Age. Quaternary International, 300: 111-130. Scerri, E.M.L. 2013b. On the spatial and technological organisation of hafting modifications in the North African Middle Stone Age. Journal of Archaeological Science, 40: 4234-4248. Scerri, E.M.L. 2017a. The North African Middle Stone Age and its place in recent human evolution. Evolutionary Anthropology, 26: 119–135. Scerri, E.M.L 2017b. The Stone Age Archaeology of West Africa. Oxford Research Encyclopedia of African History: 1-37. USA, Oxford University Press. Taylor, N. 2014. Central and West African Middle Stone Age: Geography and Culture. In Smith, C. (ed.), Encyclopedia of Global Archaeology: 1208-1227. Dordrecht: Springer. Tryon, C.A., Faith, J.T. 2013. Variability in the Middle Stone Age of Eastern Africa. Current Anthropology, 54, Supplement 8: S234-S254. Wadley, L. 2015. Those marvellous millennia: the Middle Stone Age of Southern Africa, Azania: Archaeological Research in Africa, 50:2: 155-226. Wurz, S., Van Peer, P. 2012. Out of Africa, the Nile Valley and the Northern Route. South African Archaeological Bulletin, 67 (196): 168-179. Ziegler, M., Simon, M.H., Hall, I.R., Barker, S., Stringer, C., Zahn, R. 2013. Development of Middle Stone Age innovation linked to rapid climate change. Nature Communications, 4: 1-9. Human Evolution Allsworth-Jones, P. 1993. The Archaeology of Archaic and Early Modern Homo sapiens: An African perspective. 1993. The Cambridge Archaeological Journal, 3: 21-39. Allsworth-Jones, P., Harvati, K., Stringer, C. 2010. The archaeological context of the Iwo Eleru cranium from Nigeria and preliminary results of new morphometric studies. In ed. P. Allsworth-Jones, West African Archaeology: New developments, New perspectives: 29-42. Oxford: BAR International Series, 2164. 235

The Middle Stone Age of Nigeria in its West African Context

Bons, P., Bauer, C.C., Bocherens, H., Menéndez, L., Uhl, A., Wissing, C. 2018. Pitfalls and opportunities in pinpointing the origin of modern humans – a numerical study. Proceedings of the European Society for the Study of Human Evolution, 7: 23. 8th Annual ESHE Meeting, Faro, Portugal. Brasil, M., Hlusko, L. 2018. Early Homo sapiens postcranial fossils from Middle Awash, Ethiopia. Proceedings of the European Society for the Study of Human Evolution, 7: 27. 8th Annual ESHE Meeting, Faro, Portugal. Campbell, M.C., Tishkoff, S.A. 2010. The Evolution of Human Genetic and Phenotypic Variation in Africa. Current Biology, 20: R166–R173. Crevecoeur, I., Brooks, A., Ribot, I., Cornelissen, E., Semal, P. 2016. Late Stone Age human remains from Ishango (Democratic Republic of Congo): New insights on Late Pleistocene modern human diversity in Africa. Journal of Human Evolution, 96: 35-57. Galway-Witham, J., Stringer, C. 2018. How did Homo sapiens evolve? Science, 360 (6395): 1296-1298. Harvati, K., Stringer, C., Grün, R., Aubert, M., Allsworth-Jones, P., Folorunso, C.A.. 2011. The Later Stone Age Calvaria from Iwo Eleru, Nigeria: Morphology and Chronology. PLoS One, 6 (9), e24024: 1-8. [with correction]. Hershkovitz, I., Weber G.W., Quam, R., Duval, M., Grün, R., Kinsley, L., Ayalon, A., Bar-Matthews, M., Valladas, H., Mercier, N., Arsuaga, J.L., Martinón-Torres, M., de Castro, J.M.B., Fornai, C., MartínFrancés, L., Sarig, R., May, H., Krenn, V.A., Slon, V., Rodríguez, L., García,, R., Lorenzo, C., Carretero, J.M., Frumkin, A., Shahack-Gross, R., Bar-Yosef Mayer, D.E., Cui, Y., Wu, X., Peled, N., Groman-Yaroslavski, I., Weissbrod, L., Yeshurun, R., Tsatskin, A., Zaidner, Y., Weinstein-Evron, M. 2018. The earliest modern humans outside Africa. Science, 359 (6374): 456-459. Hublin, J.J., Ben-Ncer, A., Bailey, S.E., Freidline, S.E., Neubauer, S., Skinner, M.M., Bergmann, I., Le Cabec, A., Benazzi, S., Harvati, K., Gunz, P. 2017. New fossils from Jebel Irhoud, Morocco and the pan-African origin of Homo sapiens. Nature, 546: 289-292. Morgan, L.E., Renne, P.R., Taylor, R.E., WoldeGabriel, G. 2009. Archaeological age constraints from extrusion ages of obsidian: examples from the Middle Awash, Ethiopia. Quaternary Geochronology, 4: 193-203. Reyes-Centeno, H., Hubbe, M., Hanihara, T., Stringer, C., Harvati, K. 2015. Testing modern human outof-Africa dispersal models and implications for modern human origins. Journal of Human Evolution, 87: 95–106. Richter, D., Grün, R., Joannes-Boyau, R., Steele, T.E., Amani, F., Rué, M., Fernandes, P., Raynal, J.P., Geraads, G., Ben-Ncer, A., Hublin, J.J., McPherron, S.P. 2017. The age of the hominin fossils from Jebel Irhoud, Morocco, and the origins of the Middle Stone Age. Nature, 546: 293-296. Scerri, E.M.L., Thomas, M.G., Manica, A., Gunz, P., Stock, J.T., Stringer, C., Grove, M., Groucutt, H.S., Timmermann, A., Rightmire, G.P., d’Errico, F., Tryon, C.A., Drake, N.A., Brooks, A.S., Dennell, R.W., Durbin, R., Henn, B.M., Lee-Thorp, J., deMenocal, P., Petraglia, M.D., Thompson, J.C., Scally, A., Chikhi, L. 2018. Did Our Species Evolve in Subdivided Populations across Africa, and Why Does It Matter? Trends in Ecology and Evolution, 33(8): 582-594. Schlebusch, C.M., Malmström, H., Günther, T., Sjödin, P., Countinho, A., Edlund, H., Munters, A.R., Vicente, M., Steyn, M., Soodyall, H., Lombard, M., Jakobsson, M. 2017. Southern African ancient genomes estimate modern human divergence to 350,000 to 260,00 years ago. Science, 358: 652-655. Stoneking, M., Cann, R.L. 1989. African Origin of Human Mitochrondrial DNA. in ed. P. Mellars and C. Stringer, The Human Revolution: 17-30. Edinburgh: Edinburgh University Press. Stringer. C., Galway-Witham, J. 2017. On the origin of our species. Nature, 546: 212-214. Stringer, C., Galway-Witham, J. 2018. When did modern humans leave Africa? Science, 359 (6374): 389390.

236

Envoi In this work, the present state of affairs with regard to the Middle Stone Age in Nigeria and West Africa has been set out to the best of my ability. This subject has been under study one way and another for the last 80 years or so. During this time, not a little has been achieved, but as the preceding account makes clear, what we know is still woefully inadequate. So what is portrayed here is no more than an interim picture, a picture it will be the task of people now working in the field to improve upon. They will have at their disposal an array of methods that were not available, or were only just becoming available, when the work described here was carried out. Above all, new dating methods, more precise methods of detection and analysis, and prompt publication, with adequate institutional backing, can be expected to make a big difference. It is possible - who knows? - that West Africa will still lack the over-abundant evidence available in other parts of the continent. But it is also obvious that there is plenty to find out, and surprises may well be in store, so there is every incentive to undertake this work. Whatever happens, it should not be the case that, in the broad scheme of things, West Africa should continue to be ignored. If this book helps to achieve that goal, it will have achieved its purpose.

237

238

Truth becomes fiction when the fiction’s true Real becomes not-real where the unreal’s real Cao Xueqin (1715-1763) The Story of the Stone

239

This book provides a full up to date account of the evidence relating to the Middle Stone Age in Nigeria and the other countries of West Africa. It relies upon the author’s own fieldwork and extensive personal knowledge of the region and its archaeology. It is abundantly illustrated with maps, photographs, and drawings. The emphasis is on stratigraphy, chronology, site situation, and artefact characteristics, with such general background information about the countries concerned as is required. A summary account is also provided of the current situation in relation to this topic (covering climate, archaeology, and human evolution) in the African continent as a whole, so that a judgement can made as to how the West African evidence fits in with the rest. In general accounts of the African palaeolithic record up to now, West Africa tends to be neglected, so this book goes a long way to fill a gap in the available literature. Philip Allsworth-Jones completed his PhD on the European Palaeolithic in Cambridge, before serving as a staff member of the Department of Archaeology in Ibadan for a number of years, during which the otherdescribed countriesinof West It relies the author’s own fieldwork and time Nigeria much ofand the fieldwork this workAfrica. was carried out. upon He subsequently held an analogous appointment at the University of the West Indies in Jamaica, but he has continued to take an interest in West Africa up to this day. He is currently an Honorary Research Fellow in the Department of Archaeology of the University of Sheffield, and while holding this appointment he has in recent years published a number of works on Nigeria, including the results of his excavations at the rock shelter sites of Itaakpa and Kariya Wuro. This is however his first attempt to tell the story of the judgement can as to how West evidence fits in with the rest. In general accounts Middle Stone Age in made West Africa as a the whole in African book form. long way to fill a gap in the available literature.

time much of the fieldwork described in this work was carried out. He subsequently held an analogous in West Africa up to this day. He is currently an Honorary Research Fellow in the Department of Archaeology of the University of Sheffield, and while holding this appointment he has in recent shelter sites of Itaakpa and Kariya Wuro. This is however his first attempt to tell the story of the

Archaeopress Archaeology www.archaeopress.com