Prehistoric Timbuktu and Its Hinterland
Park, Douglas Post, Antiquity
Timbuktu and its associated region are located in the nation of Mali in a transitional zone between the southern Sahara Desert and the Sahel, which is a long and narrow semiarid grassland or steppe that stretches across the entire continent (Grove 1978; Tucker & Nicholson 1999: 587). Timbuktu can also be viewed as lying on the southern edge of the Azawad basin (Figure 1). Although now dry and desolate, the Azawad can be inundated during periods of great precipitation with river water that travels more than 150km north through a network of wadis, aquifer-charged lakes, seasonal playas and marshes (Webb 1995; R. McIntosh 1998: 34-48).
During the colonial period, archaeological research in the Azawad was conducted by the French, mainly in areas around military posts or wells along the salt caravan, a strategy influenced by the harsh climate of the region, as well as by unfriendly relations with the Tuareg. Numerous Late Stone Age (LSA) sites were discovered along the banks of now desiccated waterbodies and palaeochannels, with finds such as stone tools, human and animal remains, harpoons and ceramics (Monod & Mauny 1957: 242-7; Mauny & Poussibet 1962: 1-5; Guitat 1972: 896-925; Raimbault 1992: 85-7; see Vernet 1993: 63-73 and R. McIntosh 1998: 38-9).
[FIGURE 1 OMITTED]
While research has been inhibited for a number of decades, due to the Sahel Drought and the Tuareg resistance movement, a few post-colonial projects were conducted in the northern Azawad. In 1964 Alain Gallay and his team of specialists produced evidence of diverse, heterogeneous sites of different types clustered together near ancient waterbodies (Gallay 1966: 167-208; Gallay & Huysecom 1993: 357-64). Such research provides information about LSA occupations that may have relevance to the greater understanding of the later prehistoric occupations, both at Timbuktu and within the greater Middle Niger. Up in the far north of the Azawad basin are the remnants of the probable LSA precursor for the largescale, urban clusters of corporate groups of the prehistoric Middle Niger.
Until now most scholarly attention at Timbuktu itself has been focused on the important Islamic history of the city (Pefontan 1922; Mauny 1952; Herbert 1980; Saad 1983; Insoll 1998, 2002, 2004; Hunwick 1999), but equally significant is Timbuktu's prehistoric phase (c. 500-400 BC to c. AD 900-1000). Archaeological survey in 1984 provided the first evidence for prehistoric urbanism at Timbuktu (R. McIntosh & S. McIntosh 1984, 1985, 1993; S. McIntosh & R. McIntosh 1986). The new research of 2008 reported here was aimed at developing a ceramic chronology and establishing a context for the prehistoric settlements in the Timbuktu region. These settlements take the form of tells ranging in size from 10-40ha (or more) in area and 3-9m in height, each of them surrounded by dozens of smaller satellite sites.
Definitions: prehistory and urbanism
Historic Timbuktu is widely believed to have been founded by the Tuareg in the twelfth century AD, an event that changed the role of the region, placing it in a larger pan-African perspective. Prehistory at Timbuktu refers to the period before this, even if in other parts of West Africa, the prehistoric period is considered to end in the eighth century AD with the rise of the Ghana Empire and the influx of Islam from the north. The definition of urbanism in this paper goes beyond the simple equation of urban societies with the emergence of states, to a broader definition that allows different degrees and types of urbanism to various parts of the world at various times (S. McIntosh 1999; Cowgill 2004; R. McIntosh 2005a; Yoffee 2005; Heckenberger et al. 2008). Prehistoric urbanism is certainly expected to be a variable phenomenon in the great swath of the Middle Niger. While there may be a common spatial pattern in the form of clustered settlements situated around a large central tell, different degrees of social organisation are expected across space and time.
The Timbuktu expedition 2008
The 2008 survey and excavation projects, in conjunction with the McIntoshs' 1984 survey, have produced two main datasets that are vital to a more intensive study of the prehistory of Timbuktu. These two main datasets lead to a preliminary ceramic chronology and dated settlement patterns.
The survey was conducted in transects along the palaeo-wadi el-Ahmar avoiding the areas already surveyed in 1984 while also testing the area between the wadi and Timbuktu (Figure 2). In addition to the 33 sites discovered in the Timbuktu region by the 1984 survey, 159 unique sites were located during the 2008 survey giving a total of 192 sites. Of these, 18 central tells were discovered, ranging from 10-40 ha (or more) in area and 3-9m in cultural accumulation. The remainder of the sites tend to be smaller (less than 10ha in area) with very little visible cultural accumulation (less than lm). The smaller sites can be considered satellite settlements surrounding the central tells. Four settlement clusters were chosen for intensive mapping (Figure 2: SC1-4). In each cluster (except SC1) a central tell dominates its area, and is located in proximity to the palaeo-wadi el-Ahmar.
The excavation targeted the central tell within Settlement Cluster 3 (SC3). Known as Tombouze 1 (TBZ1), named after the village located only 2km to the east, this central tell presented both logistic advantages and an accumulated cultural stratigraphy of 3.4m. This provided multiple phases of occupation and a ceramic chronology, a provisional account of which is given here.
[FIGURE 2 OMITTED]
[FIGURE 3 OMITTED]
The excavation took the form of four test pits and two test trenches dug through the tell TBZ1 (Figure 3). The section shown (Figure 4) is that cut through the tell by Unit 6 (a 3 x 4m test trench), and the summary below relates to this unit.
Phase 1 ceramics (Figure 5) are similar in both style and form to central Saharan protoBerber twine-impressed and sand-tempered ware (R. McIntosh & S. McIntosh 1985: 31-2; cf. Castdli et al. 2005; cf. Gatto 2005). Provisionally, it is suggested that there may be connections between contemporary Garamantian ceramic assemblages from the central Libyan Desert (Gatto pets. comm.). Phase 1 generated significant amounts of fish and aquatic faunal material, with fewer goat/sheep and cow remains when compared to the upper levels. Evidence for iron smelting in the form of iron slag was abundant in Phase 1, along with large quantities of red and yellow ochre. The final level of Phase 1 witnesses a possible short abandonment as evidenced by a collapsed building and burnt roof beams. Preliminary dates for this phase are from c. 500-400 BC to a radiocarbon date of cal AD 120-330 (at 95% probability) (Beta-256796).
[FIGURE 4 OMITTED]
Phase 2 ceramic decorations are similar to a painted red- or black-on-white geometric style found in small numbers at Phase III Jenne Jeno (S. Mclntosh 1995: 155), although applied to possible Saharan vessel types with localised variations in style (Figure 5). Paint is often applied over horizontal channelling in cross-hatched patterns. Polychrome sherds (red, purple and white) are also common and paints are usually applied over the body of the vessel in a distinctive zonal manner, i.e. a solid purple band separating a solid red band below and a purple-on-white cross-hatching pattern above. In the cultural accumulation above the abandonment of Phase 1 (1.1m deep), large amounts of carnelian and glass beads were found along with a slightly reduced amount of iron slag. Far less aquatic fauna was discovered, and a significant increase in sheep/goat and cow remains was found. Radiocarbon dates for this phase are cal AD 120-330 (at 95% probability) to cal AD 430-640 (at 95% probability) (Beta-256795).
[FIGURE 5 OMITTED]
Phase 3 ceramics were characterised by ribbed/channelled, red slipped, rocker combed and burnished wares typical of Gao Ancien (Insoll 1996: 83-4, 2000) and, alongside the 0.8m of cultural accumulation, provide a preliminary chronological interpretation that limits the site's terminus date to c. AD 900-1000 (Figure 5). It is thus far believed that, due to the complete absence at TBZ1 of imported glazed pottery, which may have come into Gao as early as the tenth to eleventh centuries AD (Insoll 1996: 65), it is unlikely that the population at TBZ1 persisted far into the second millennium AD.
With an area of 40ha and an elevation of 3.4m, TBZ1 proved to have a long-term occupation spanning from around 500-400 BC to about AD 900-1000. The TBZ1 Phase 3 chronology of cal AD 430-640 (at 95% probability) ending in 900-1000 AD coincides with periods 5 and 4 from Gao Ancien (MM96[C]) (Insoll 2000: 4). It is suggested that, due to the similarities in ceramic style between TBZ1 and Gao Ancien, an informal social or economic relationship may have existed between the two sites. Other than ceramics, almost no aquatic fauna are found in this level, and the ratio between goat/sheep and cow remains suggests that bovids were the most common form of meat. Fewer beads were recovered, and an increase in smelting activity is present in amounts far greater than what was found in either Phase 1 or Phase 2.
All the central tells of SC1-4 had dense Phase 3 ceramic scatters on the surface (cal AD 430-640 [at 95% probability] to c. AD 900-1000), with probable cultural accumulation spanning back to Phase 1. The satellite sites surround the central tells in a non-uniform pattern that does not seem to be directly related to proximity to main water sources. Evidence for segregated iron smelting satellite sites is present at SC4. Surface scatters at most of the satellite sites date to ceramic Phase 2 (cal AD 120-330 [at 95% probability] to cal AD 430-640 [at 95% probability]). No Phase 1 ceramics (c. 500-400 BC to cal AD 120-330 [at 95% probability]) were found within the surface scatters at any sites, and thus far only come from the lowest layers of the excavated central tell in SC3.
The survey thus creates a picture of a dynamic landscape where urbanism grew during Phase 2, but population levels then declined by Phase 3. Eventually the satellite sites were abandoned and occupation was found solely in the central tells. Faunal subsistence patterns at the central tell of TBZ1 changed over time as aquatic protein became less available, or less important, than ruminant protein. Variations in iron production show that initial and final period occupations at TBZ1 may have been more active in blacksmithing than those of Phase 2, where the iron production may have been moved outside the main central tell to the satellite sites. During Phase 2 at TBZ1 there are indications of informal social or economic relations with Inland Niger Delta sites such as Jenne Jeno. However, it is impossible to say what kind and at what level those relations were at this point. A similar situation is observed at Phase 3 TBZ1 with Gao Ancien. At the moment there is no strong body of evidence coming from the TBZ1 excavations or regional surveys which suggests any meaningful relationship with Islam before the abandonment of the prehistoric settlements around Timbuktu.
The role of climate change
A major factor in the variation in settlement structure is likely to be the changing climate. Understanding human response to climate change has been a recent focus of many archaeologists, climatologists, geomorphologists and various international organisations, and it is believed that such an understanding can come from a multidisciplinary and diachronic study spanning numerous episodes of climate variations and of human social response to those variations (R. McIntosh et al. 2000). However, at present the partially reconstructed Holocene climate of the Middle Niger and Azawad basin does not have enough resolution to be of much application to the archaeological record beyond approximative relationships between climate and society. Recent coring at Lake Bosumtwi in Ghana suggests that a long-term cyclical pattern of wet and dry episodes on multi-decadal and century scales has been a feature of WAM for the past 3000 years (Shanahan et al. 2009: 379). However, the regional patterns recorded in the Lake Bosumtwi coring project, and also trends discerned at a lower resolution from research at Lake Chad (Maley 2000; Brunk & Gronenborn 2004), may have limited applicability to our region of study (located 1200km north of Lake Bosumtwi and even further from Lake Chad) since climate patterns are more variable in both time and space than previously assumed (Bradley 2000).
Since the modern and future implications of continuing droughts in the Sahel are of immediate concern, understanding historical processes of human response to climate change provides insight into social actions that can be implemented, or avoided, to mitigate population stresses caused by future long-term dry episodes. In other words, when trying to solve contemporary issues of sustainability in the Sahel, it may be possible to use studies of ancient experience. In particular we would address the question of why some institutions of climate response survived abrupt climate change, and why others failed. Institutions of climate response represent a society's active component of perceptions and ideologies about how to respond to climate anomalies with the goal of fostering sustainability in terms of subsistence, conflict resolution and socio-economic interaction. Conflict and collapse is a result of the failure of those cultural institutions to adapt to strenuous and abrupt climatic change, not a direct result of climate change.
Sustainability and resilience are central components in either the success or failure of institutions of climate response. Sustainability, defined as '... maintaining, or fostering the development of the systemic contexts that produce the goods, services, and amenities that people need or value, at an acceptable cost, far as long as they are needed or valued' (Allen et al. 2003: 26), can only be achieved in high-risk environments, such as the Sahel, if institutions of climate response are highly resilient. Resilience, defined as '... the ability of a system to adjust its configuration and function under disturbance' (R. McIntosh & Tainter 2004: 4), is directly related to the development-anthropological literature of risk management (Raynaut 1997). Resilience activities related to risk management during periods of abrupt climate change include: mobility, diversification, intensification, specialisation, storage, exchange and reciprocity (Redman 2005). Each of these activities is archaeologically detectable, and a chronological study of how they changed over time in response to problems with sustainability, as caused by climate change, should provide information about the development of institutions of climate response.
Reconstructing haw the prehistoric urbanism at Timbuktu arose and persisted in such strenuous climatic conditions is the goal of our future research. Reconstructing the local climate in the southern Sahara will also help to generate an accurate picture of possible migration routes into the Sahel. Irreparable damage to the Saharan interior results from high amplitude and temporal fluctuations, thus desertification and the silting of watercourses eats away at the fragile desert landscape (Mainguet 1991). As a result, people in the Azawad may have slowly moved south entering the Middle Niger and, along the way, also settled at Timbuktu. Research in the Mema region, some 300km west-southwest of Timbuktu, provides evidence for initial LSA occupation by about 2000 BC (MacDonald 1994: 273-5), which is at the hypothesised terminus of LSA migrations out of the Sahara that may have, in part, gone through Timbuktu (R. McIntosh 1998: 60). Therefore, the earliest ephemeral occupation levels around Timbuktu may prove to be at least contemporary with the Mema. However, LSA populations around Timbuktu are hypothesised to be mostly hidden by shifting and recently mobile dunes and are therefore difficult to detect (R. McIntosh & S. McIntosh 1984: 22), although every village in the area has a collection of LSA stone tools that are found eroding out onto the dry palaeo-wadi channels crisscrossing the desert floor. Targeting LSA populations at Timbuktu will take a separate, dedicated and highly specialised survey approach that includes satellite remote sensing and predictive ArcGIS geographic spatial modelling.
Discussion and a working hypothesis
Making an initial comparison between Timbuktu and Middle Niger urban settlements such as Jenne Jeno (S. McIntosh 1995; R. McIntosh 1998: 208), Dia (R. McIntosh & S. McIntosh 1987; Bedaux et al. 2001) and the Mema (Togola 1993, 1996; MacDonald 1994; R. McIntosh 2005b), we observe a diverse social and physical landscape. This observation offers a correction to the assumption that complex societies necessitate centralised hierarchies and homogeneous cultural traits (Stahl 2004:151). Not limited to terms such as hierarchy, the concept of urbanism becomes broader, including large and heterogeneous places that provide services and specialist manufactures to a broader hinterland (Trigger 1972: 577). Such a definition includes multi-centric and clustered settlement patterns; it does not require places to be of a particular size or population, and there is no specification as to the minimal number or ranking of corporate groups. Here, corporate groups are defined as cohesive social entities that are reinforced both by common possessions and specialised skills, along with access to corporate property (Cochrane 1971; R. McIntosh 1993:188). What is important about this definition of urbanism is that it focuses on the exchange of human products between the urban cluster and the hinterland. Human products are considered to be: goods, services, information and ideology. Without the flow of human products from the urban centre to the hinterland, urbanism could not persist.
The settlement patterns at Timbuktu are believed to represent an expansive urbanism possibly characterised by heterarchical power relationships. While it remains to be tested, the rise and fall of the prehistoric urbanism at Timbuktu was probably due, in part, to a longterm process of human response to variable, and largely unpredictable, climatic conditions. Enough information has been gathered to start asking questions and formulating hypotheses about the process of urbanisation. It is proposed here that high amplitude and high temporal climatic variations, along with a general aridification trend beginning at the end of the second Saharan Holocene pluvial period (c. 2500 BC), acted as a forcing mechanism that brought various corporate groups from the northern Saharan Azawad region down to the banks of the Niger Bend where, over time, separated but articulated corporate groups formed urban clusters which carried out relations of reciprocity, and which resisted centralised power gathering, resulting in a complex and heterarchical society.
With further excavation, survey, environmental and climate studies, the working hypothesis can be tested along with an exploration of the diachronic sociopolitical transformations during the prehistory of the Timbuktu region. Future research at Timbuktu will also add to the growing discussion of alternative forms of urbanism while contributing interpretations on how humans respond to climate change in a long-term historical framework. Furthermore, such research will contribute to the understanding of an important, yet under-researched, region of the world.
Research was conducted at Timbuktu during the months of June and November 2008. I would like to thank Dr Roderick McIntosh (Yale University), Dr Susan Keech McIntosh (Rice University), Dr Klessigue Sanogo (Directeur National de la Direction Nationale des Arts et de la Culture, Mali), Dr Klena Sanogo (Directeur de Hnstitut des Sciences Humaines), Dr Ali Ould Sidi (Chef de la Mission Culturelle de Tombouctou), Mamadou Cisse (Rice University), Dr Maria Gatto (Yale University), Peter Coutros (Yale University) and the entire staff at the Direction Nationale du Patrimonie Culturel (DNPC, Mali).
Received: 3 July 2009; Accepted: 24 September 2009; Revised: 14 October 2009
ALLEN, T.F.H., J.A. TAINTER, T.W. HOEKSTRA (ed.). 2003. Supply-side sustainability. New York: Columbia University Press.
BEDAUX, R., A. SCHMIDT, K. MACDONALD, A. PERSON, J. POLET, K. SANOGO & S. SIDIBE. 2001. The Dia archaeological project: rescuing cultural heritage in the Inland Niger Delta (Mali). Antiquity 75: 837-48.
BRADLEY, R. 2000. 1000 years of climate change. Science 288: 1353-5.
BRUNK, K. & D. GRONEBORN. 2004. Floods, droughts and migrations: the effect of late Holocene lake level oscillations and climate fluctuations on the settlement and political history in the Chad Basin, in M. Krings & E. Platte (ed.) Living with the lake: perspectives on history, culture and economics of Lake Chad (Studien zur Kulturkunde 121): 101-32. Cologne: Rudiger Koppe.
CASTELLI, R., M. CREMASCHI, M. GATTO, M. LIVERANI & L. MORI. 2005. A preliminary report of excavations in Fewet, Libyan Sahara. African Journal of Archaeology 3(1): 69-102.
COCHRANE, G. 1971. Use of the concept of the 'corporation': a choice between colloquialism or distortion. American Anthropologist 73: 1144-55.
COWGILL, G.L. 2004. Origins and development of urbanism: archaeological perspectives. Annual Review of Anthropology 33: 525-49.
GALLAY, A. 1966. Quelques gisements neolithiques du Sahara malien. Bulletin de la Societe des Africanistes 36(2): 167-208.
GALLAY, A. & E. HUYSECOM. 1993. Un site neolithique de l'Adrar Tabarbarout (Sahara malien oriental). Bulletin de la Societe Prehistorique Francais 90(5): 357-64.
GATTO, M. 2005. The local pottery, in M. Liverani (ed.) Aghram Nadharif: the Barket Oasis in Garamantian times (Sha Abiya of Ghat, Libyan Sahara) (The Archaeology of Libyan Sahara 2; Arid Zone Archaeology Monographs 5): 201-240. Florence: All'Insegna del Giglio.
GROVE, A.T. 1978. Geographical introduction to the Sahel. The Geographical Journal 144(3): 407-415.
GUITAT, R. 1972. Carte et repertoire des sites neolithiques du Mali et de la Haute-Voha. Bulletin de l'Institut Fondamental d'Afrique Noire, serie B 34(4): 896-925.
HECKENBERGER, M.J., J.C. RUSSELL, C. FAUSTO, J.R. TONEY, M.J. SCHMIDT, E. PEREIRA, B. FRANCHETTO & A. KULKURO. 2008. Pre-Columbian urbanism, anthropogenic landscapes, and the future of the Amazon. Science 321: 1214-17.
HERBERT, H. 1980. Timbuktu: a case study of the rule of legend in history, in B.K. Swarz & R.E. Dumett (ed.) West African culture dynamics: 431-54. The Hague: Mouton.
HUNWICK, J. 1999. Timbuktu and the Songhay Empire. Leiden: Brill.
INSOLE, T. 1996. Islam, archaeology and history: Gao Region (Mali) ca. AD 900-1250 (Cambridge Monographs in African Archaeology 39; British Archaeological Reports International series 647). Oxford: Tempus Reparatum.
--1998. Archaeological research in Timbuktu, Mali. Antiquity 72: 413-17.
--2000. Urbanism, archaeology and trade: further observations on the Gao Region (Mali). The 1996 fieldseason results (British Archaeological Reports International series 829). Oxford: J. & E. Hedges.
--2002. The archaeology of post medieval Timbuktu. Sahara 13: 7-22.
--2004. Timbuktu the less Mysterious?, in P. Mitchell, A. Haour & J. Hobar (ed.) Researching Africa's past: new contributions from British archaeologists: 81-8. Oxford: Oxbow.
MACDONALD, K.C. 1994. Socio-economic diversity and the origin of cultural complexity along the Middle Niger (2000 BC to AD 333). Unpublished PhD dissertation, University of Cambridge.
MAINGUET, M. 1991. Les actions du vent au Sahara: un example de systeme eolien ouvert. Sahara 4: 7-12.
MALEY, J. 2000. Last glacial maximum lacustrine and fluviatile formations in the Tibesti and other Saharan mountains, and large-scale climatic teleconnections linked to the activity of the subtropical jet stream. Global and Planetary Change 26(1): 121-36.
MAUNY, R. 1952. Notes d'archeologie sur Tombouctou. Bulletin de I'Institut Francais d'Afrique Noire (B) 14: 899-918.
MAUNY, R. & F. POUSSIBET. 1962. Nouveaux sites harpons et faune subfossile de l'Azawad (Sahara malien). Notes Africaines 93: 1-5.
MCINTOSH, R.J. 1993. The pulse theory: genesis and accommodation of specialization in the Middle Niger. Journal of African History 32(2): 181-220.
--1998. The peoples of the Middle Niger. Oxford: Blackwell.
--2000. Social memory in Mande, in R.J. McIntosh, J.A. Tainter & S.K. McIntosh (ed.) The way the wind blows: climate history and human action: 141-80. New York: Columbia University Press.
--2005a. Ancient Middle Niger: urbanism and the self-organizing landscape. Cambridge: Cambridge University Press.
--2005b. Pulse of the Mema: deep-time risk strategies, in E. de Vries, B. Makaske, J.A. Tainter & R.J. McIntosh (ed.) Geomorphology and human palaeoecology of the Mema, Mali (Alterra-Report 1244): 67-83. Wageningen: Alterra.
MCINTOSH, R.J. & S.K. MCINTOSH. 1984. Prospection archeologique aux alentours de Tombouctou, 1984. Rapport preliminaire. Nyame Akuma 24/25: 5-12.
--1985. Archaeological reconnaissance in Timbuktu, Mali. Unpublished Final Report to National Geographic.
--1987. Prospection archeologique aux alentours de Dia, Mali, 1986/1987. Nyame Akuma 29: 42-5.
--1993. La prospection d'apres les photos aeriennes: regions de Jenne et Tombouctou, in J. Devisse & J. Polet (ed.) Vallees du Niger: catalogue scientifique: 234-47. Paris: Reunion des Musees Nationaux.
MCINTOSH, R.J. & J.A. TAINTER. 2004. Paleoclimates and the Mande. Mande Studies 6: 1-10.
MCINTOSH, R.J., J.A. TAINTER & S.K. MCINTOSH. 2000. Climate, history and human action, in R.J. McIntosh, J.A. Tainter & S.K. McIntosh (ed.) The way the wind blows: climate history and human action: 1-42. New York: Columbia University Press.
MCINTOSH, S.K. 1995. Excavations at Jenne Jeno, Hambarketotlo, and Kaniana: the 1981 season. Berkley (CA): University of California Press.
MCINTOSH, S.K. (ed.) 1999. Beyond chiefdoms: pathways to complexity in Africa. Cambridge: Cambridge University Press.
MCINTOSH, S.K. & R.J. MCINTOSH. 1986. Archaeological reconnaissance in the region of Timbuktu, Mali. National Geographic Research 2(3): 302-19.
MONOD, T. & R. MAUNY. 1957. Decouverte de nouveaux instruments en os dans l'ouest africain, in J.D. Clark (ed.) Proceedings of the third pan-African congress on prehistory, Livingstone: 242-7. London: Chatto & Windus.
PEFONTAN, L. 1992. Histoire de Tombouctou de sa fondation a l'occupation francaise (XII siecle-1893). Bulletin du Comite d'Etudes Historiques et Scientifiques de l'Afrique Occidentale Francaise: 81-113.
RAIMBAULT, M. 1992. Les facies neolithique du Sahara Malien avant l'aridification. Acres du 116e Congres National des Socieres Savantes: 85-7.
RAYNAUT, C. (ed.) 1997. Societies and nature in the Sahel. London: Routledge.
REDMAN, C. 2005. Resilience theory in archaeology. American Anthropologist 107: 70-77.
SAAD, E.N. 1983. Social history of Timbuktu: the role of Muslim scholars and notables 1400-1900. Cambridge: Cambridge University Press.
SHANAHAN, T.M., J.T. OVERPECK, K.J. ANCHUKAITIS, J.W.. BECK, J.E. COLE, D.L. DETTMAN, J.A. PECK, C.A. SCHOLZ & J.W. KING. 2009. Atlantic forcing of persistent drought in West Africa. Science 324: 377-80.
STAHL, A.B. 2004. Political economic mosaics: archaeology of the last two millennia in tropical sub-Saharan Africa. Annual Review of Anthropology 33: 145-72.
SWART, W. 1998. Stable oxygen isotopic composition of corals from the Gulf of Guinea as indicators of periods of extreme precipitation conditions in the sub-Sahara. Journal of Geophysical Research 103(12): 27 885-91.
TOGOLA, T. 1993. Archaeological investigations of Iron Age sites in the Mema (Mali). Unpublished PhD dissertation, Rice University.
--1996. Iron Age occupation in the Mema region, Mali. African Archaeological Review 13: 37-62.
TRIGGER, B. 1972. Determinants of growth in pre-industrial societies, in P. Ucko, R. Tringham & D. Dimbleby (ed.) Man, settlement and urbanism: 575-99. London: Duckworth.
TUCKER, C. & S. NICHOLSON. 1999. Variations in the size of the Sahara desert from 1980 to 1997. Ambio 28(7): 587-91.
VERNET, R. 1993. Le Neolithique recent dans le sud-ouest du Sahara, in L. Krzyaniak, M. Kobusiewicz & J. Alexander (ed.) Environmental change and human culture in the Nile Basin and northern Africa until the second millennium BC: 63-73. Poznan: Poznan Archaeological Museum.
WEBB, J. 1995. Desert frontier: ecological and economic change along the western Sahel, 1600-1850. Madison (WI): University of Wisconsin Press.
YOFFEE, N. 2005. Myths of the archaic state. Cambridge: Cambridge University Press.
Douglas Post Park, Department of Anthropology, Yale University, 10 Sachem Street, New Haven, CT 06511, USA (Email: firstname.lastname@example.org)…
Questia, a part of Gale, Cengage Learning. www.questia.com
Publication information: Article title: Prehistoric Timbuktu and Its Hinterland. Contributors: Park, Douglas Post - Author. Journal title: Antiquity. Volume: 84. Issue: 326 Publication date: December 2010. Page number: 1076+. © 2008 Antiquity Publications, Ltd. COPYRIGHT 2010 Gale Group.