African ArchaeologicalReview, Vol. 13, No. 2, 1996New Research on the Holocene Settlement andEnvironment of the Chad Basin in NigeriaPeter Breunig, 1 Katharina N e u m a n n , 1 and Wim Van Neer 2 Recent investigations of three archaeological sites in the Nigerian part of the Chad Basin during the Holocene reveal key stages in the cultural development and environmental history of that region. At Dufuna, a dugout boat was dated to around 6000 BG making it the oldest known boat in Africa and one of the oldest in the world. Boats may thus have contributed to the mobility of thepopulation of the southern edge of the Sahara 8000 years ago and, thereby, to the cultural homogeneity of this period. The pottery site at Konduga is around a thousand years younger than Dufuna but still belongs to the time ofMega-Chad. The site is on the Bama Ridge, an old shoreline. Its pottery,decorated in the Saharan tradition, belongs to the earliest ceramic phase of the West African Later Stone Age, long before the beginnings of foodproduction. Although this site was probably settled by pioneers advancing intoa largely flooded landscape along the slightly raised shoreline, the humanoccupation of the area previously covered by Mega-Chad began along a broadfront around 2000 BC. Archaeological and palaeoecological finds from twosettlement mounds at Gajiganna are described as case studies for this phase, which predates the well-known site of Daima.Les r~sultats prdsentds proviennent de recherches rgcentes sur trois sitesarchdologiques, dont chacun reprgsente un dpisode clg du d~veloppementculturel et de l'histoire environnementale du Bassin nigdrian du Tchad aucours de l'Holoc~ne. A Dufuna, la ddcouverte d'une pirogue monoxyle, datded'environ 6000 ans BC, constitue le tdmoignage le plus ancien d'uneembarcation en Afrique et l'un des plus ancien dans le monde. Ce moyen detransport indique la mobilitd des populations de la marge sud du Saharal'Holoc~ne infdrieur et moyen; ce qui a dft contribuer ~ l'homogdnditd culturellede cette pdriode. Le site d poterie de Konduga est le plus jeune d'un milldnaire~Universit/it Frankfurt, Seminar ftir Vor- und FriJhgeschichte, Arch/iologie und Arch~iobotanik Afrikas, 60325 Frankfurt am Main, Germany.2IUAP 2 8 - "Interdisciplinary Archaeology," Royal Museum of Central Africa, 3080 Tervuren, Belgium. 111 0263-0338/96/0600-0111509.50/0 9 1996 Plenum Publishing Corporation112 Breunig, Neumann, and Van Neerque Dufuna, ma& appartient ?~ l'~poque du M~ga-Tchad. Le site est install~sur une ancienne ligne de rivage, la Bama Ridge. Sa poterie, ddcorde selon latradition saharienne, appartient ~ la plus ancienne phase ~ cdramique du LaterStone Age ouest-africain, bien avant le d~but d'une production alimentaire.Vu que ce site ne devait ~tre occup~ que par des pionniers aventurds dans unpaysage largement amphibie glla faveur d'un mince cordon dmergd, lacolonisation de l'espace ant~rieurement couvert par le M~ga-Tchad, ddbute lelong d'un large front vers 2000 ans BC Les trouvailles archdologiques etpaldo~cologiques de deux tertres anthropiques ~ Gajiganna sont d~critescommes dtudes de cas illustrant cette phase, qui pr~cdde celle du c~lObre sitede Daima.KEY WORDS: Nigeria; Holocene; Later Stone Age; paleoenvironment; paleoeconomy. INTRODUCTION In contrast to the Sahara, little is known of the pre-Iron Age culturaldevelopment and Holocene environmental history of the West Africansavanna. Of necessity, overviews such as that by S. McIntosh and R.McIntosh (1983) have developed models of settlement and environmentalhistory for the Sahara, in which the savanna has then been included. Thepresent pape r deals with the southwestern Chad basin, which is noexception to these tendencies. Although the region's unusual richness inprehistoric sources has been known since Graham Connah's fieldwork inthe 1960s (1976, 1981), the settlement mound at Daima has remained theonly site shown on most archaeological maps. The lack of attention paidto archaeological research in this region stands in stark contrast to thei m p o r t a n c e of the C h a d basin. T o g e t h e r with the rest of theSahelo-Sudanic zone, it is considered to be the area of origin of the majorAfrican cultigens and thus of food production in sub-Saharan Africa(Harlan, 1971). The beg!nning of food production and its terminological treatment inthe division of African prehistory have been central concerns until recenttimes (Sinclair it al., 1993). However, from the varied use of the term"Neolithic," taken from European and Near Eastern prehistory, it is clearthat no unified view has been reached. The same applies to the importanceattributed to food production in human history in general. It has beenconsidered as anything from an epochal change to a relatively subsidiarysupplement to the foraging ecPonomy, expressed by the terminologicaltreatment of the beginning of food production as a late development in the Later Stone Age (Shaw, 1981). These varying attitudes arise to some extent from the extreme paucity of tangible facts which might transform theoreticalHolocene Settlement and Environment of the Chad Basin 113postulates into concrete statements. Results such as those from Dhar Tichittin Mauretania on subsistence strategies leading to the domestication ofPennisetum are among the much-quoted exceptions (Munson, 1976; Amblardand Pern6s, 1989). Our aim in a long-term joint research program by theUniversity of Frankfurt (Germany) and the African partner universities inOuagadougou (Burkina Faso) and Maiduguri (Nigeria) is therefore tocontribute to the insufficiently researched archaeology, archaeobotany, andarchaeozoology of the origins of animal and plant domestication .and toincorporate them into the discussion of a primary center of domestication inAfrica. We believe that progress may be made only by means of extensiveexcavation and collaboration with archaeobotanists and archaeozoologists.Every year since 1989 the project has included several months of fieldwork,working toward a typological and chronological framework, as well as therecovery and analysis of faunal and floral remains. The project has carried out research in the southeast of Burkina Faso.Various assemblages of the Later Stone Age were recorded. At Maadaga,a large rock shelter at the edge of the Chaine de Gobnangou, the LaterStone Age was stratified above a rich macrolithic complex (Breunig andWotzka, 1993). With the help of numerous thermoluminescence dates, abasic framework for stages in ceramic development reaching back 4000years was constructed (Ballouche et al., 1993). The investigation of charcoalfrom the sites shows changes from forest formations to savanna in thecourse of the last 7000 to 8000 years. These seem to have been due mainlyto human impact -- most probably the use of fire for hunting purposes --whereas climatic fluctuations were of minor importance. However, indica-tors of large-scale cultivation have not been observed in the charcoalsamples, and remains of cultivated plants are completely absent from thesites (Neumann and Ballouche, 1992). Apart from dog, the animal boneassemblage -- not securely datable -- consists entirely of remains of wildanimals. From Oursi in the north of Burkina Faso recent research has defineda sequence of vegetation history with indications of an abrupt change inthe pollen spectrum around 3000 years ago. Plants from open habitats,plowed and fallow land, increase clearly from 3000 bp, giving the first in-dications that agriculture may have been practiced here (Ballouche andNeumann, 1995). As this area is rich in prehistoric sites, it is hoped tofollow the development of cultivation back to its early stages using ar-chaeobotanical methods. The pollen diagram from Oursi clearly showshuman influence on the development of the vegetation. It thus representsa clear contrast to the widely held tendency to attribute the vegetation his-tory of the whole Holocene in the Sahel to climatic change rather than to human settlement and activity (cf. IAzine, 1989).114 Breunig, Neumann, and Van Neer I T ~ 12" 16" L l U Cmd --16' 16 NIGER S P S Lo~mmd Owt YN / 9 / -L _. 9 . ,_... / "2:2:? Dufuna --12" I - 12" 16" Fig. 1. Map of the Chad basin showing the sites discussed in the text. In Nigeria, work is concentrated in the area around Maiduguri, theadministrative center of Borno, and in another case on the region northof Potiskum (-Yobe State) (Breunig et al., 1993a, b; Breunig, 1996). Fur-thermore, in late 1993 excavations were begun to the south of Lake Chadon settlement mounds in the vicinity of Daima. These mounds, all withstratified deposits several meters in thickness, represent archaeological po-tential comparable with that of Near EaStern tells -- as shown by Connah(1976, 1981) in the case of Daima. The reasons for the formation of thetells in the southern Chad basi~) arise from the seasonal flooding of theclay plain during the rainy season. Then as now, the inhabitants protectedtheir settlements from flooding by building them on relict islands of sand.Through continuity of settlement and the use of clay for building, thickHolocene Settlement and Environment of the Chad Basin 115 Fig. 2. The Dufuna boat. A tree trunk lies directly over the boat.deposits accumulated, containing extraordinary amounts of archaeologicalmaterial. The preservation of organic material, -including that of botanicalremains, is good. We are therefore confident that we will gain knowledgeabout the history of plant cultivation and the exploitation of natural re-sources in this region. While this part of the project is still in its initial stages, other inves-tigations, on which detailed reports follow, are more advanced. Theseconcern three sites from various stages of the Holocene in northeast Ni-geria (Fig. 1). Each represents, in our opinion, a key site in the Holocenesettlement of the Chad Basin.116 Breunig, Neumann, and Van Neer D U F U N A - - A N 8000 Y E A R O L D D U G O U T B O A T At Dufuna (ca. 12~ 11~ a small village on the KomaduguGana River, outside the former extent of Mega-Chad, a dugout canoe wasdiscovered, which is dated to 7264 B.P. + 55 years (61164-6005 CalBC, KN-4683) and 7670 B.P. + 110 years (6556-6388 CalBC, KI-3587) according tothe conventional C-14 dates. This makes it the oldest boat in Africa andone of the oldest in the world. The boat was found at the edge of a seasonally flooded basin, 5 mbelow the surface. It is stratified below clays and sands -- some containingpollen -- whose alternating sequence is evidence of deposition in standingand flowing water, presumably as a result of shifts in the river channel.There are no associated archaeological finds. A few sherds and charcoalfragments were found in the layers above the boat. In a layer of gray claydirectly above the coarse sand in which the boat lies, a small, complete,undecorated pot was found. A tree trunk is embedded in the same layer,transverse to the axis of the boat. The date of the trunk is 2612 B.P. + 48years (809-786 CalBC, KN-4689). The boat is fully preserved apart from a few missing fragments fromthe tops of the sides (Fig. 2).-It has a length of 8.40 m and maximumbreadth and height of around 0.5 m. The sides are barely more than 5 cmthick. The bow. and stern are both pointed, giving the boat a notably moreelegant form than finds of similar age from Mesolithic Europe, such as thedugout from Presse in the Netherlands (Van Zeist, 1957), which is datedto 8265 B.P. + 275 years (Gro-486) (conventional date) and is the oldestboat in the.world apart from suggested Late Palaeolithic skin boats (Ell-mers, 1984; Tromnau, 1987). It is highly probable that the Dufuna boatdoes not represent the beginning of a tradition, but that it was precededby a long development and that the origins of water transport in Africareach even further back in time. The boat was made from the genus Khaya (family Meliaceae), Af-rican mahogany. The four species of mahogany occurring today in Ni-geria cannot be.separated by their wood structure. K. ivorensis and K.anthotheca are rain forest trees, K~ grandifolia grows in drier forests, andK. senegalensis extends through the savanna regions up to the southernedge of the Sahel where it is found along water courses (Keay, 1989;Maydell, 1986). All species of mahogany are suitable for boats and ca-noes as the wood keeps well under water because of the phenol contentin its heartwood (Irvine, 1961;'q~ichter, personal communication). Thewood of K. senegalensis is the heaviest of the African mahoganies andless easy to work than the other.Khaya species. Nevertheless, it is oftenHolocene Settlement and Environment of the Chad Basin 117used in the savanna regions today, and its suitability for boat buildingis well-known. Judging by the modern distribution, K. senegalensis wouldbe the only species that could have grown as far north as the Dufunaregion. However, as the climate was most probably more humid 8000years ago, it is possible that the more southerly species were carried tothis area by river. T h e b o a t will r e m a i n in the g r o u n d until the p r o b l e m s ofconservation have been solved. Accordingly, the analysis of the boat's sidesand the tool-marks on them will only be possible after the boat is lifted.However, even without the tool-marks it is likely that axe- or adze-liketools would have been necessary to hollow out the tree trunk, with orwithout controlled burning. Ground stone tools of this kind are known inWest Africa only from later archaeological contexts. The tools weretherefore probably unground core-axe-like and pick-like bifacial tools ofmacrolithic appearance, such as are known from Rim 1 (northern BurkinaFaso) at about the same latitude (Andah, 1978) and known elsewhere inWest Africa for decades (MacDonald and Allsworth-Jones, 1994). In somecases they have b e e n c o n s i d e r e d as an a d a p t a t i o n to the forestenvironment (Shaw, 1980, p. 77). The inhabitants of Dufuna probably used the boat in exploiting thefish stocks of the Komadugu Gana. Whereas the present-day river almostvanishes in the dry season, and fishing plays a relatively minor part inlocal subsistence, hydrological conditions around 6000 BC would havemade aquatic resources much more available. The boat existed at the sametime as Mega-Chad, whose surface is agreed by most geographers to havebeen 40 m higher than its present level (Bawden, 1972). Today this wouldmake it the largest lake in the world together with the Caspian Sea. Theformer shoreline of Mega-Chad indicates that Dufuna would then havelain barely 50 km from the lake. The flood plain at Diafuna may evenhave been a permanently flooded lagoon of Lake Chad, formed byinundation of the Komadugu Gana. Under these conditions fish wouldhave been in abundant supply all year round.. If these assumptions arecorrect, the makers of the dugout belonged to a population that spreadalong the southern edge of the Sahara, from north Kenya through thecentral Sudanese Nile Valley to the western Sahara, and adapted to theresources of the lakes of the early and mid-Holocene wet phase. Thedugout indicates that water transport was part of the cultures of thisperiod. The vicinity of the M e g a - C h a d inland sea stimulated them a n u f a c t u r e of boats in West Central Africa at a time for whichcomparable finds are known only from Europe.118 Breunig, Neumann, and Van Neer _ _ _ _ _ _ + r w c l 3 l L ~ f f TU +r "- t U I ' ~t KI)NDII(;A I Fig. 3. Rocker stamp-decorated pottery from Konduga (around 5500-5000 BC). K O N D U G A - THE BEGINNINGS OF POTrERY PRODUCTION Just as pre-Neolithic or epil~laeolithic pottery production is known inthe Sahara with "wavy line decorated pottery," there seems also to havebeen a ceramic phase in the wetter regions of West Africa before the be-ginnings of food production (Shawl 1981, p. 216). Pottery was manufacturedHolocene Settlement and Environment of the Chad Basin 119in the West African south of the Sahara for 3000 years before the first foodproduction, for which there is no firm evidence much before 3000 bp. Thelong-known observation, made in various parts of the world, that the inven-tion of strong vessels suitable for use in cooking is not necessarily relatedto food production can be further confirmed by a newly discovered, veryearly pottery site in northeast Nigeria. The site is called Konduga. It is situated on the highest point of theBama Ridge, which consists of beach sands and extends north as far asNiger and south into Cameroon (Grove, 1985, p. 39). The ridge i s u p to12 m in height and spreads out in places to a width of several kilometers.It is generally considered to represent the shoreline of Lake Chad in theearly and middle Holocene (Bawden, 1972). The site was found by H. Thiemeyer during geographical surveys. Itis in a sand quarry, from whose undisturbed sections pottery was recovered,associated with charcoal which yielded dates of 6340 B.P. + 250 years(5515-4994 CalBC, KN--4300), (Thiemeyer, 1992, 1993). Excavations to adepth of 1.2 m below the surface have revealed finds distributed throughthree horizons, of which only the lowest (1.0-1.2 m below the surface) isrelevant in the present context. This assemblage is very small, consistingof quartz fragments, one abraded flake, charcoal particles, and a few sherds.There are no faunal remains. The age of the assemblage has been con-firmed by a further C-14 date from individual charcoal fragments (6180B.P. + 60 years, 5241-5004 CalBC, UTC-2248). However, there is an unex-plained discrepancy between these dates and a thermoluminescence datefrom pottery of 4150 BC. The context of the finds is stratigraphically secure. The sherds haveunabraded edges and cannot have been transported any great distance.Their spatial distribution also suggests rapid incorporation into primary de-posits: all conjoining sherds were found less than a meter apart. Thedeposition of the charcoal found with the pottery most probably took placenot much later, and it is unlikely that older charcoal could have intrudedinto the sand ridge. It seems reasonable to assume that the available C-14dates represent the age of the pottery. The pottery is heavily tempered with quartz fragments. Its surface isextensively decorated with a rocker stamp technique (Fig. 3). Typologicalparallels may be easily found with pottery of a similar age from the Sahara.In the eastern central Sahara, for example, pottery decorated in the styleof Konduga succeeds the wavy line complex and falls into the sixth or fifthmillennium B.P. (Schuck, 1990, p. 135). Shaw (1978/1979, p. 70) has sum-marized the characteristic techniques of pottery decoration in the pre-food-producing phase of West Africa as consisting exclusively of comb decora-tion. The question of whether the rocker stamp technique known from120 Breunig, Neumann, and Van NeerKonduga is also present as an archaic element at the other sites still needsto be addressed. According to Shaw this technique, together with groove,roulette, and cord decoration, is not characteristic until the later ceramichorizons (after 3500-3000 B.P.). The importance of the site to the history of the landscape comes fromthe first direct date for the shoreline. The structure and laminated natureof the sand deposits leaves no doubt that they were laid down in water.The dated charcoal fragments are clearly rounded, which is considered tobe typical of beach sediments (Thiemeyer, 1992, p. 116). The majority of the species in the modern plant communities of Kon-duga is of Sahelian origin even though Konduga belongs to the Sudanianzone in the sense of Keay (1959). This is due mainly to the fact that therecent vegetation around the site is completely anthropogenic and consistsof millet fields and fallows which are heavily grazed. In contrast, the char-coal samples recovered from the site (Table I) indicate a Sudanianwoodland or even a forest around 5500-5000 BC. In the upper layers frag-ments of Terminalia were found which gave modern radiocarbon dates.Obviously Sudanian species were able to exist in the Konduga area up tomodern times, and most probably a combination of overexploitation anddecreasing rainfall is responsible for their disappearance in the twentiethcentury (Bailouche and Neumann, 1995). The date of the pottery is early for northeast Nigeria and breaks --as does the Dufuna dugout -- the 3000 B.P. barrier, earlier than which noevidence of settlement had previously been known in the southwesternChad Basin (Connah and Jemkur, 1982; Connah, 1984). However, it seemscertain that. the environs of the early and middle Holocene Chad basinwere inhabited and -- in view of the desiccation of the Sahara -- were evenimportant as an ecologically favorable environment. On the other side ofthe shoreline, away from Palaeo-Chad, there were probably extensive la-goons, fed by the Yedseram and other rivers, such as are known today ona smaller scale as seasonal floods at Konduga. The Bama Ridge could thushave been a narrow land bridge at the time of the early and middle Holo-cene lake, leading into a largely flooded area and playing a central part inthe later settlement of the former lake bed. GAJIGANNA THE BEGINNINGS - - OF ANIMAL DOMESTICATION AND VILLAGE SETI'LEMENT IN THE ~OUTHWESTERN CHAD BASIN The earliest phase of settlement yet documented in the area of landexposed by the retreat of Lake Chad is known from several sites near ??,o O Q Terminalia Monocoty/edonae O0 Acacia Euphorbiaceae J Grewia cf. mollis O Securinega/Hyrnenocardia e~ I'0 Parinari polyandra Ix.) Leguninosae type Parkia i'O Allophylus V P~ O~ Rubiaceae Leguninosae type 1 0 Lannea/Sclerocarya Anona senegalenais Isoberlinia doka Unknown type 0[Z'I u!se8 P e q 3 oql .IO luauluo.lI.AU~l puE luom~[lloS auaaOlOH122 Breunig, Neumann, and Van Neer Fig. 4. Map showingarchaeologicalsites in the Gajiganna region.Gajiganna, 80 km north of Konduga and just over 30 km from the oldshoreline. Whereas Konduga simply provides evidence for the presenceof a community that manufactured pottery comparable with that of theSahara, the sites described below, some 3000 years younger, provide amore detailed picture of settler~ent and environment at the end of theStone Age. In the time elapsed between Konduga and Gajiganna thelandscape changed drastically. The waters of Lake Chad, more than 40 mabove their present level in the Konduga phase (ca. 6000 B.P.), fell toHolocene Settlement and Environment of the Chad Basin 123 Fig. 5. Section through Gajiganna A.around 20 m above present levels by the time of the sites aroundGajiganna (the earliest of which, not dealt with in the present paper, dateback to around 3700 B.P. according to uncalibrated C-14 dates). Previouslysubmerged areas dried out, forming a habitable landscape that wascolonized by a population bringing with it domesticated animals andbuilding the first villages in the southwestern Chad basin. Following theconventional view of the spread of domestic animals, particularly cattle,through North Africa over several millennia (Shaw, 1981, p. 223; Smith,1984), it may be assumed that these groups came from the north, drivensouthward by the desiccation of the Sahara. However, this is difficult toprove archaeologically. According to the geographical division ofnortheast Nigeria by Tuley(1972), Gajiganna is situated in the northern part of the Bama DeltaicComplex. This complex has a very gentle, undulating relief, consisting ofslightly raised sandy areas and clay plains which predominate toward LakeChad. Delta sediments of the rivers Yedseram and Ngadda, and longitu-dinal dunes in the southeastern and northern parts of the complex, arealso present. Between the dunes are clay depressions, which combine inthe east to form a clay plain with only sporadic sandy cover. The clays arelagoon deposits from the lake. Gajiganna was situated on the edge of thislagoon (Fig. 1).124 Breunig, Neumann, and Van Neer Fig. 6. Section through Gajiganna B. Stratigraphy and Dating Both of the sites at Gajiganna presented here (Gajiganna A and Ga-jiganna B) consist of low mounds, arouhd 2 m in height and 150 m indiameter, situated directly next to each other. They are surrounded by clayflats, which in places retain rain,,~ater on the surface well into the dry sea-son. Surface prospection and tri~il excavations in the environs of GajigannaA and B have located numerous similar sites with comparable finds (Fig. 4).The two excavated mounds are thus representative of a phase of relativelydense settlement of this region, dated to the centuries before and afterHolocene Settlement and Environment of the Chad Basin 1253000 B.P. However, they do not represent the earliest settlement phase,which, according to recent excavations of sites nearby, goes back as far as3700 B.P. (uncalibrated) and is characterized by a pottery facies quite dif-ferent from that of Gajiganna A and B (Wendt, 1995). The northeastern sides of both mounds are eroded. The strong windsof the dry season have carried away the finer sediment, leaving concentra-tions of archaeological material on the surface, which led to the sites'discovery during geographical surveys by H. Thiemeyer. In both cases, 5x 5-m trenches dug at the highest points of the mounds revealed similarstratigraphy and almost identical finds. Settlement mound A consists of about 2 m of deposits (Fig. 5). Theupper part of the meter-thick main cultural layer (layers 3, 4, and 5) isdated to 2960 B.P. + 50 years (1286-1022 CalBC, UTC-2795). The sedi-ment in these layers consists largely of clay. As there is no explanationfor the natural deposition of clay on a mound situated on sand, we as-sume that the clay was brought onto the site by the inhabitants in theform of mud-bricks or, more probably, daub for houses. However, apartfrom small daub fragments, we have no evidence to support this assump-tion. The structural remains have probably been destroyed by the shrink-ing and swelling of the clayey soil. Below the upper cultural layer thereis a sterile sand horizon (layer 6), formed during a phase of increasedaeolian deposition which may have been the result of a drier climate orof human influence on the environment. The lower cultural layer, onlysome 20 cm thick, is dated to 2930 B.P. + 60 years (1236-1009 CalBc,UTC-2329) and is, thus, not clearly separated from the stratigraphicallylater phase. The stratigraphy of mound B at Gajiganna is very similar (Fig. 6),again with an upper and a lower cultural layer. The difference in the C-14dates is clearer than for mound A. The upper .cultural la3'er (layer 3) has a date of 2740 B.P. + 50 years (967-813 CalBC, UtC-2331~), and the lowercultural layer (layer 4) is dated to. 3150 B.P. + 70 years (1490-1318 CalBC, UTC-2330), 3140 B.P. + 110 years (1520-1310 CalBC, UTC-2332) and 2880 B.P. + 165 years (1304-834 CalBC, KN-4675). A pit, dated at its base to 3040 B.P. + 120 years (1429-1052 CalBC, KI-3605), reaches from the lower cultural layer to a depth of 4.5 m below surface. The upper cultural layer was also identified in a small test excavation at the foot of the mound, where it yielded dates of 2730 B.P. + 50 years (966-811 CalBC, UTC-2796) and 2750 B.P. + 70 years (971-811 CalBC, UTC-2797). The calibrated C-14 dates suggest an approximate duration of settlement at Gajiganna A and B from 1200 to 900 BC for the upper cultural layer and from 1500 to 1200 BC for the lower.126 Breunig, N e u m a n n , a n d Van N e e r Table II. Faunal List of Gajiganna A and B (Figures Indicate Fragment Counts) Gajiganna A Gajiganna B Mollusks Large bivalve (of, Aspatharia) 4 -- Pila cf, wernei 50 36 Limicolaria sp, 3 -- Fish Polypterus sp, 1 28 Heterotis niloticus I 36 Gymnarchus niloticus .2 -- Clatias spp. 385 747 Reptiles Monitor lizard (Varanus sp,) -- 1 Python (Python sp,) -- 22 Terrapin (Pelusios sp,) 15 7 Unidentified turtle 18 8 Birds Ardeidae Heron (Ardea sp.) Ciconiidae 9 Marabou (Leptoptilus crumeniferus) 1 Ciconiidae indet. 1 Anatidae Spur-Winged goose (Plectropterus gambensis) 2 Knob-billed goose (Sarkidiornis rnelanota) 1 Egyptian goose (Alopochen aegyptiaca) 3 Pintail (Anas acuta) -- 1 Accipitridae Large vulture (Aegypinae indet.) Phasianidae Francolin (Francolinus sp.) 1 D Guinea-fowl (Numida meleagris) 3 1 Rallidae Moorhen ( Galinulla chloropus) Otididae ) Arabian bustard (Ardeotis arabY) Unidentified birds 11 ContinuedHolocene Settlement and Environment of the Chad Basin 127 Table II. Continued G~igannaA G~iganna B Mammals Giant rat ( Cricetomys gambiunus) 1 D Unstriped grass rat (Arvicanthis niloticus) 2 Small rodents 4 D Patas or guenon (Cercopithecus sp.) 4 Canid (Canis sp.) 6 3 Small viverrid (size Mungos mungo) 3 1 Medium viverrid (size Ichneumia albicauda) 7 African wild cat (Felis libyca) 1 Serval (Felis serval) 4 Warthog (Phacochoerus aethiopicus) 18 3 Bohor reedbuck (Redunca redunca) 31 6 Forest duiker {Cephalophus sp.) 3 Grimm's duiker (Sylvicapru gnmmia) 2 Oribi ( Ourebia ourebia) 7 Small and medium-sized bovids 289 52 Sheep (Ovis ammon f. aries) 1 1 Goat (Capra aegagrus f. hircus) 73 14 Sheep/goat 172 [6 Cattle (Bus primigenius f. taurus) 855 433 Total identified 1,979 1,433 Unidentified 10,338 4,172 Several observations indicate that the now rather flat landscape hada more varied relief at the time of the prehistoric settlements. The lowestcultural layers are below the modern surface of the seasona!!y flooded clayplain surrounding the mounds. The surface of the day must once havebeen at a lower level; otherwise, the settlements would have been built inwater. Palaeoecology and Palaeoeconomy Evidence for the environment and economy is provided by weU-pre-served organic material: pollen, animal bones, and charcoal. However,attempts to recover fruit and seed remains by large-scale wet-sieving andflotation were unsuccessful. Only the kernels of Ziziphus sp. (carbonized)and Celtis sp. (uncarbonized) from Gajigarma A provide evidence for the128 Breunig, Neumann, and Van Neergathering of wild fruit. Uncarbonized modem seeds, particularly of grasseswhich grow today in the area around Gajiganna, were found at depths ofup to 330 cm, indicating severe bioturbation, which may have affected stra-tigraphic integrity. Although the extent of any disturbance is difficult toestimate, the total mixing of deposits may be discounted in view of thelogical sequence of charcoal dates and the existence of some intact contexts(burials). Analysis of pollen from the sands below the cultural layer at Gaji-ganna B revealed a predominance of the pollen type Eugenia, which,together with other hygrophilous taxa, points to dense vegetation at theedge of water. The later settlement phases, on the other hand, are repre-sented by plants of the open savannah, together with indicators ofdisturbance (A. Ballouche, personal communication). Further information on the environment and economy comes fromthe abundant faunal remains. Approximately 18,000 faunal remains werestudied, of which almost 2000 were identifiable at site Gajiganna A andmore than 1400 at site Gajiganna B (see Table II). The inhabitants used a wide variety of animals to meet their proteinneeds. Mollusks were collected, especially Aspatharia, which is a largeaquatic bivalve, and Pila, a large terrestrial gastropod. During its estivationin drying swampy areas, the latter mollusk can be easily collected in largequantities. Fishing was practiced for a limited number of species: Polypterussp., Heterotis niloticus, Gymnarchus niloticus and especially Clarias sp. Allthese fish can live in shallow waters, which would have made exploitationeasy. The other aquatic animals that were captured were terrapins, whichare ffeshw~iter turtles that can survive in drying waters by burrowing intothe mud. Besides fishing and harvesting of mollusks and aquatic turtles,hunting played a role in subsistence. At least 12 bird species were hunted,the majority of which live on or near water. Among the exploited wild mam-mals the following species occur: giant rat, guenon monkey, viverrids, wildcat, serval, warthog, bohor reedbuck, duikers, and oribi. The majority ofthe faunal remains, however, comes from domestic animals: goats, sheep,and cattle. Sheep were rare in comparison to goat, which indicates thatthe available pasture may have been limited and used primarily for cattle(Redding, 1984). Whether this was due to climatic conditions, to overgraz-ing as a result of the large size of the flocks, or to a combination of bothfactors is uncertain. Gajiganna is one of the e~liest sites with domestic animals in WestAfrica south of the Sahara. The radiocarbon age of the site fits nicelywith the theory of the spread of earliest pastoralists in Africa as shownin the isochronic diagram presented by Shaw (1981, p. 224) and recentlyupdated by Krzyzaniak (1992). Southward migrations of pastoralists from "0 ????? Acacia Acacia cf. nilotica tan .1~ ~ to Balanites aegyptiaca CI" Cehis integrifolia Combretum .= Ficus Grewia Z Lannea/Sclerocarya ~oo~ Leguminosae type 2 Piliostigrna to Prosopis afncana ~t.,a Rubiaceae type Mitragyna -.,J~...,ta, a Terminalia Ximenia americana taak, a Ziziphus6~I u!sefl p e q ~ ~q! jo luomuoa.~U~l pue luom~lll~ S au~3olo H130 Breunig, Neumann, and Van Neerthe Sahara started with the onset of arid conditions roughly 5000 to 4000years ago (Gautier, 1987). Around 3000 B.P. (according to uncalibratedradiocarbon dates), probably even half a millennium earlier, herdersarrived in the West African savanna. Among the few earliest find-spotsin the savanna (Daima and Gajiganna in Nigeria, Douentza in Mali,Kintampo and Ntereso in Ghana), Gajiganna provides one of the largestand most secure samples. Further research is presently being conductedto try to reconstruct the possible seasonal occupation of the site. The cattleand goat remains comprise several juvenile and neonate specimens whichshould allow us to determine approximately the season during which theydied. Combined with data on the present-day movements of herds in thearea (Brauk~imper, 1996) an attempt will be made to obtain a model forthe past. T o d a y the area of Gajiganna is a rather harsh environment,belonging to the southernmost part of the Sahel. It is visited only duringthe rainy season by herders. Isolated bodies of water of limited extent arepresent. They are fed by the annual rains. The animal remains found atthe sites allow a reconstruction of the former aquatic and terrestrialenvironment. The fish fauna comprise four taxa, of which the clariidcatfish are the most common, These fish prefer shallow habitats and cansurvive in strongly deoxygenated waters with elevated temperatures and ahigh salinity. Atmospheric oxygen can be taken up by an accessorybreathing organ. Both Heterotis and Gymnarchus are most common inmarshy waters with abundant aquatic vegetation (especially Echinochloastagnina). They also have accessory breathing organs allowing survival inoxygen-depleted waters. This is also the case for Polypterus, a taxon thatprefers sandy bottoms and that occurs in considerable numbers in dryingpools. Today the isolated pools near Gajiganna still contain Clarias. It canbe considered a relic of an originally much richer fish fauna, which survivedthe last arid period until the present day. A natural enrichment of such anisolated pool is possible only through a connection with another body ofwater where more species occur. Establishing whether the waters near Ga-jiganna were already isolated from the rest of the Chad basin at the timeof occupation is difficult on the basis of the fish remains alone. In anycase, the ichthyofauna was richer than the present-day relic faunas in iso-lated waters of the Sahara, where clariids, cyprinids, and cichlids usuallypredominate (Van Neer, 1989). This could mean either that the water wasstill connected, even if only seasonally, with the Chad basin, or that thedisconnection and the resulting decline in the number of species had hap-pened only shortly before.Holocene Settlement and Environment of the Chad Basin 131 Regarding the number of species identified, it should be emphasizedthat it is quite possible that the ichthyofauna 3000 years ago was richerthan indicated by the remains. Fishing techniques or dietary preferencesmay account for this. Geomorphological observations in the area also seemto indicate that the backwaters of a once larger Lake Chad formed a lagoonwhich extended as far as Gajiganna. On the basis of the vertical distribution of the identified species, anattempt was made to identify an evolution in the local environment, pos-sibly toward more arid conditions. The only trend that can be seen for theterrestrial fauna is an increase in wild carnivores in the upper cultural layerof site Gajiganna A. This may be an indication of increased aridity andless phytomass, which resulted in a concentration of herbivores around thewater. Carnivores preying on these animals followed and, as a result, be-came more vulnerable to predation by man. The chronological distribution of aquatic fauna was also analyzed. Itmight be expected that increased aridity and a possible decrease in theavailable water would be reflected in a reduction of the number of species,with the most resistant ones remaining numerous. It was also interestingto look at the distribution of the sizes of fish through time as adverse hy-drologic conditions and overfishing may result in a decrease in size overtime. However, no such trends are seen. Fifteen taxa have been identified in the charcoal samples from theupper and the lower layer of Gajiganna B (Table III), six of which arepresent in more recent vegetation (Ballouche and Neumann, 1995). In thelower layer the number of species is higher than in the upper one, but thisis probably due to the larger number of identified fragments. As for thesmall number of charcoal fragments, the two layers are treated as a singleunit covering a time span of several hundred years. Firewood was collectedfrom the dunes as well as in the inundated depressions. The wooded grasslandmight have formed the grazing area for the cattle, whereas shrub was suf-ficiently available to feed the goats. Taxa such as Ziziphus sp'. and Balanitesaegyptiaca indicate that grazing and browsing occurred, but the impact onthe vegetation probably was not as severe as can be observed today. Char-acteristic elements of the modern degraded pastures, such as Guierasenegalensis, Boscia senegalensis, or Calotropis procera, are missing in thesamples. Nevertheless, the low number of sheep may indicate that therewas already some overgrazing. Goats have a different diet and competeless with cattle for the same fodder. Typical forest plants, as have been found in Konduga, are notpresent in the charcoal samples. However, the bones of the forest duikerCephalophus testify that there must have been some denser woodland oreven patches of forest, probably at the edge of the water. In general, the ~J t~ a, o i Li f ~3C~ Z 0. w~ Z c~ Q ~ 3 t-. \o oO ~3C~ ~ ~L~ ~ J134 Breunig, Neumann, and Van Neer I 2 3 4 $ 6 7 8 ~ ~ ~ 9 10 II 12 13 14 if 15 17 16 j : " Ir ~ / / 0 ~cm 19 ~:~ '3 v.~- ~. ,~r -,.'... 2O P Fig. 8. Stone artifacts from Gajiganna.Holocene Settlement and Environment of the Chad Basin 135reconstructed vegetation is of a Sahelo-Sudanian savanna with a densertree cover than today. This fits with the mammalian fauna whose moderndistribution is in the present-day Guinea and Sudanian vegetation zonesof Nigeria (Happold, 1987). Elements typical of Sahelian savanna, suchas gazelles, are completely absent. Ten of fifteen taxa in the charcoal samples are no longer present inthe recent vegetation. The modern phytogeographical distribution of thesetaxa includes the Sahelo-Sudanian zone, and the mean annual rainfall inthe Gajiganna area (as it was measured up to the 1970s) would be sufficientto support all of them. Species such as Terminalia and Prosopis africanaare known for their excellent qualities as firewood, as charcoal, and forconstruction purposes (von Maydell, 1986). For that reason it seems rea-sonable to suppose that overexploitation in the course of the past 3000years is responsible for the impoverishment of the vegetation. Small-scaleclimatic fluctuations such as the great drought of the 1970s might well haveaccelerated the extinction of those species that reach the northern limit oftheir distribution in the Gajiganna area. Burials Although human bone was found in almost every stratigraphical unit,graves are confined to the upper cultural layer of Gajiganna A. The deadwere buried within, or at least in the immediate vicinity of, the settled area.Grave pits could not be identified, and there were no grave-goods. All theskeletons were in the crouched position, lying on the side with arms flexedand with the hands in front of the face. There was no uniform orientation.Apart from one burial where the head was missing, the skeletons are com-plete and undisturbed, although poorly preserved and very fragmented.According to anthropological analysis by F. R6.sing (Unix~ersity of Ulm),two of the total of four burials are of young adult or middle-aged males,178 + 4 and 169 + 4 cm in height. At least one of them shows negroidcharacteristics. There is also the burial of a 7 to 10 year-old-girl, and ajuvenile female 158 + 8 cm tall. Artifacts Pottery is by far the most abundant category of artifact. From bothexcavations, covering a total of just 50 m 2, around 180,000 sherds, weighing750 kg altogether, were recovered. Based on the amount excavated, Gaji-ganna mound A alone must contain an estimated 200 tons of pottery.136 Breunig, Neumann, and Van Neer 0 s i0 o I 2 3 4 5 6 o 9 J~ 3 3 a II 16 7 12 13 14 15 ! z 17 18 Fig. 9a. Bone artifacts from Gajiganna. )Considering the duration of the settlement of a few centuries at most, sucha volume of material and the size of the mounds suggest the long-termexistence of a large social group and of a village settlement. o O" i m I @ .< ~o i .t " "" " ""~o00C~ w ,o o ~ Ii . . . . r- - - J Table IV. List of Radiocarbon Dates ,~ oo Konv. Calibrated date age a Lab.Site (B.P.) (CalBc) code Material DescriptionDufuna 7264 + 55 6164-6005 KN-4683 Wood BoatDufuna 7670 _+ 110 6556-6388 KI-3587 Wood BoatDufuna 2612 _+ 48 809-786 KN-4689 Wood Tree trunk above boatKonduga 6340 -+ 250 5515--4994 KN-4300 Charcoal Pottery site on top of the Bama Ridge, basal layerKonduga 6180 +- 60 5241-5004 UTC-2248 Charcoal Pottery site on top of the Bama Ridge, basal layerGajiganna A 2960 +_ 50 1286-1022 UTC-2795 Charcoal Settlement mound, upper layer in trench on top of the moundGajiganna A ~" 2930 _+ 60 1236-1009 UTC-2329 Charcoal Settlement mound, basal layer in trench on top of the moundGajiganna A 2141 + 66 344-53 KN-4674 Bone (distal femur of cattle) Settlement mound, basal layer in trench on top of the mound. Discrepancy with charcoal datesGajiganna B 2740 __. 50 967-813 UTC-2331 Charcoal Settlement mound, upper layer in trench at the foot of the moundGajiganna B 2730 -+ 50 966-811 UTC-2796 Charcoal Settlement mound, upper layer in trench at the foot of the moundGajiganna B 2750 _+ 70 971-811 UTC-2797 Charcoal Settlement mound, upper layer in trench at the foot of the moundGajiganna B 2880 _ 165 1300-834 KN-4675 Bone (calcaneus of cattle) Settlement mound, lower layer in trench on top of the mound zGajiganna B 3040 +_ 120 1560-920 K1-3605 Charcoal Settlement mound, trench on top of the mound, basal filling of m pit belonging to the lower layerGajiganna B 3150 _+ 70 1514-1328 UTC-2330 Charcoal Settlement mound, lower layer in trench on top of the mound gl.Gajiganna B 3140 __. 110 1520-1310 UTC-2332 Charcoal Settlement mound, lower layer in trench on top of the mound" O n e - s i g m a standard deviation. Calibration according to Calibration Program 1993 Rev. 3.0 (Quaternary Isotope Laboratory, University of Z Washington). o Bama Ridge Gajiganna Area mm ~ firki area ~ ~ B z 1000 1000 calAD ~m~muni~es 0 calAD calBC 0 = calBC villa~li~ -1 <. ~ O a~ Z m a - , - - , - - p i . , , m . wgmm ~ments z 1000 = 3 2000 ~ a ql~-- jiga lie c r (3 2 . ~-: ~;--. - prod~'tion, 2O0O p. m c el mo~ ol o 3OOO 400~ 4OOO 50OO 5000 = 6000 m 6OOO -= m pre- = O~'una ce~mlr ~.. 700O 7OOO ; dug,,~Fig. 10. Nigeria, Chad basin: radiocarbon chronology of Later Prehistory. Radiocarbon dates are shown for sites in the various study areas (cali- bration of radiocarbon dates: Radiocarbon Calibration Program 1993 Rev. 3.0. Quaternary). U,140 Breunig, Neumann, and Van Neer The typological and technological study of the pottery will form thebasis for the internal chronology of all sites investigated in the study area.By tying in the pottery chronology from current excavations immediatelysouth of Lake Chad (Ngala-Daima region) and studying the temporal andspatial distribution of the key types, we hope to gain an outline of the courseof settlement in the southwestern Chad basin during the Holocene. In thiscontext, the settlement mounds Gajiganna A and B represent an early stage,but not the earliest. On the basis of the pottery, two phases can be distin-guished in both mounds, identical to the stratigraphical separation betweenthe upper and the lower cultural layers. The most conspicuous distinguishingfeature is the tempering of the clay: whereas the clay of the later potteryphase is so heavily tempered with plant material that the surface is some-times covered with plant impressions (analysis in progress), vegetable temperis completely absent in the earlier phase. Instead, only mineral componentsare used. Both groups have in common rounded vessel bases, hemisphericalor globular bodies, and a common range of forms (dishes, jars, bowls, andglobular vases) which may be subdivided further. The decoration is charac-terized by painting on the upper part of the vessel and mat impressions onthe lower part and, further, by its clear, uniform structure and various combimpressions arranged in horizontal bands (Figs. 7a and b). An older phase(between ca. 3700 and 3300 B.P.) is not present at the sites descibed herebut is found at Konduga (the assemblage described above) and in the Ga-jiganna region (Wendt, 1995). Compared to the pottery, lithic artefacts are sparsely represented (ca. 18kg from both excavations). The reason for this is undoubtedly the distance be-tween Gajiganna and the nearest sources of raw material (Garba, 1993). Exceptfor Hadjer el Hamis 180 km to the north, a small range of hills consisting ofrhyolite at the southern edge of Lake Chad, all lithic sources are well over 100km south or southwest of Gajiganna. According to analyses carried out by S.Freeth (University College of Swansea, Great Britain) on the Gajiganna lithics,meta-vulcanites from Maroua (Cameroon), 200 km away, represent the mostdistant raw material source. Raw material from Hadjer el Hamis was abundantin Daima (Connah and Freeth, 1989). The fact that this material is absent fromGajiganna may suggest that access to this area was blocked by the lagoon, atthe edge of which Gajiganna was situated. The most abundant stone tool group, consists of fragments of grinding stones made of plutonites or sedimentary rocks. This is probably due to the importance that plant foods -- wild or cultivated -- must have had. As the de-gree of fragmentation is very high,.,~nd most of the pieces are only a few cen- timeters in size, it is hardly possible to reconstruct forms. Only one type, with a pointed base and dished sides, is definitely known at Gajiganna (Fig. 8: 20); it has already reached the British Museum in London as the "Sao-hat" of theHolocene Settlement and Environment of the Chad Basin 141Chad basin (Connah, 1981, p. 49). Ground stone axes, adzes, and chisels wereprobably also important. Complete examples in this category range from 20 tojust over 2 cm in size (Fig. 8: 1-8, 18, 19). They are made from vulcanites ofvarying textures. Flakes of this material occur in the excavated areas and allover the surface of the site. Almost all have ground surfaces, suggesting thatthey were detached during tool use. A few flakes were made into arrowheads,which, together with stone beads, represent further categories of lithic artifact(Fig. 8: 9-17). The arrowheads have biracial invasive retouch and concave bases.This is a Saharan type which occurs only rarely in the West African savannah.The few parallels include sites of the Kintampo culture in Ghana (Davies, 1980,p. 215; Flight, 1976, p. 214) and Dutsen Kongba (York, 1978) and Shilma (Con-nah, 1981, p. 62) in Nigeria. The relatively high number of bone artifacts (80 from the excavationsof both mounds) suggests that bone as a raw material probably compen-sated for the shortage of stone. Tool forms include unilateral harpoons,indicating the exploitation of aquatic resources (see, however, Connah,1981, p. 117). Bone points of various forms may be classed as weapons.Chisel-like tools are common. Rings made from long bones were probablyfor ornament (Figs. 9a and b). Art is represented by various anthropomorphic and zoomorphic clayfigurines, all fired (Breunig, 1994). The anthropomorphic figures includesome with clearly defined anatomical details such as ears, nose, eyes, andmouth and highly stylized forms on which only the hairline is shown. Theanimal figures include zoologically unclassifiable forms, on the one hand,and numerous conical fragments, on the other, probably representing thehorns of boviform figurines. Boviform figurines are also predominant severalcenturies later in the Daima I phase south of Lake Chad, where anthropo-morphic figures are absent (Connah, 1981, pp. 134ff). As far as Gajigannais concerned, the great importance of cattle in the economy seems to haveinfluenced the art. The art is simple but is among the earliest in Borno andcomes, as in Daima, from a period before the beginnings of iron production.Considering the well-known figurative art of the Nok style, it seems reason-able to suggest that art in this region has its roots in the Stone Age. SUMMARY Fieldwork carried out by the German-Nigerian project since 1991 hasled to the discovery of sites which have yielded key data on the Holocenesettlement and environment of northeast Nigeria. The present radiocarbondates are listed in Table IV and summarized in Fig. 10, which representthe current state of research.142 Breunig, Neumann, and Van Neer The Dufuna find shows that the earliest boats known in Africa werebuilt in the Chad basin, independent of contemporary developments inEurope. This, the earliest form of transport known from this region andperiod, shows the similarity of the early Holocene cultures at the southernedge of the Sahara in a new light. The dugout canoe may also have beenone of the means of transport with which, around 7000 years ago, groupswith pottery decorated in a Saharan style moved south along the shore ofMega-Chad, leaving their traces at Konduga. The sites in the Gajiganna area document the beginnings of animaldomestication and village settlement in the region previously submerged byMega-Chad. The inhabitants lived on low settlement mounds, close to claydepressions which were probably flooded all year round at the shifting edgeof a lagoon connected to Lake Chad. In the course of at most 600 years(according to uncalibrated C-14 dates) of occupation, either the climate be-came drier or human influence produced similar effects on the environment.All usable fauna of the region were hunted, including fish and other aquaticanimals. Judging by the bone remains, however, wild animals were less im-portant than domesticates, among which cattle were predominant. Thesewere the earliest known domesticated animals in the region, and their ageof around 3000 years fits with th e general pattern of the appearance of earlydomesticated cattle in Africa. The importance of plant foods in Gajigannais more difficult to estimate. Despite intensive searches, no remains of cul-tigens have been found, so that we initially assume the existence of a nona-grarian, pastoral economy, with the exploitation of wild plants. Reasons for the abandonment of the settlements are not known. The mostlikely are palaeoecological changes, which are known to have taken place duringor after the occupation of Gajiganna. The lagoon dried out, and Lake Chadmust have retreated to approximately its present extent. The waters returnedonly in the rainy season, and environmental conditions probably came to re-semble those of today. The population, whose economy was at least partiallybased on aquatic resources, may have found conditions no longer suitable forliving. Shortly after Gajiganna, settlement of the clay plains to the East began,which are only seasonally free of water. One of the important questions ad-dressed by our current investigations concerns the supposition that the settle-ment of this region spread from the former lagoon shore, including Gajiganna. ACKNOWLEDGMENTS A. Garba (Centre for Trans Saharan Studies, University of Maiduguri,Nigeria) is involved in the project on the Nigerian side. We thank him for hiscollaboration and for the continuation of the Dufuna project. We also thankHolocene Settlement and Environment of the Chad Basin 143the National Commission for Museums and Monuments, especially Mrs. A. K.Fatunsin, Dr. L. I. Izuakor, and M. Hambolu, for their reliable collaboration.D. Gronenborn and P. Wendt (Universit~it Frankfurt) are involved in the field-work and interpretation. F. R6sing (Universit~it Ulm) kindly carried out theanalysis of the human skeletal remains. S. Freeth (University College of Swan-sea, Great Britain) and A. Garba (University of Maiduguri) identified the lithicraw materials. The radiocarbon dates were provided by K. van der Borg (Uni-versiteit Utrecht), J. Freundlich, B. Weniger (Universit~t K61n), and H..Wilk-omm (Universit/~t Kiel). We are grateful to them, as we are to C. Goedickeof the Rathgen Research Laboratory for the thermoluminescence dating of thepottery from Konduga. We thank G. Connah, F. Hassan, and K. MacDonaldfor valuable advice. The project is financed by the Deutsche Forschungsgeme-inschaft, the University of Frankfurt, and the Geseiischaft ~ r TechnischeZusammenarbeit. We are grateful to Canon and Agfa for their generous sup-port. The text concerning the faunal remains presents research results of theBelgian program on Interuniversity Poles of attraction initiated by the BelgianState, Prime Minister's Office, Science Policy Programming. We thank D. Un-derwood for translating and checking the English text, Aziz Ballouch for theFrench summary, and M. Heckner, B. Voss, and H. Schluse for the illustrationsof the finds. The excavations would not have been possible without the com-mitment and enthusiasm of the workers from Gajiganna and Dufuna. REFERENCES CITEDAmblard, S., and Pern6s, J. (1989). The identification of cultivated pearl millet (Pennisetum) amongst plant impressions on pottery from Oued Chebbi (Dhar Oualata, Mauritania). African Archaeological Review 7:117-126.Andah, B. W. (1978). Excavations at Rim, Upper Volta. West African Jo,o71al of Archaeology 8: 75-138.Ballouche, A., and Neumann. K. (1995). Pollen from Oursi/Burkina Faso and charcoal from NE Nigeria: A contribution to the Holocene vegetation history of the.West African Sahel. Vegetation History and Archaeobotany 4( 1): 31-39.Ballouche, A., Kfippers, K., Neumann, K., and Wotzka, H.-P. (1993). Aspects de l'occupation humaine et de I'histoire de la vegetation au cours de-I'holoc6ne dans la region de la Chaine de Gobnangou, S.E. Burkina Faso. Berichte des Sonderforschungsbereichs 268, VoL 1, Frankfurt a.M., pp. 13-31.Bawden, M. G. (1972). Geomorphology. In Tuley, P. (ed.), The Land Resources of North East Nigeria, VoL 1, Land Resource Division, Surbiton, pp. 60-70.Brauk~imper, U. (1996). Transhumance and the use of grazing resources of the Shuwa Arabs (northeastern Nigeria). Presented at International Conference on Research for Development in the Arid Zone of Nigeria (Maiduguri, 1994) (in press).Breunig, P. (1994). Early prehistoric art in Borno (NE Nigeria). Sahara 6: 98-102.Breunig, P, (1996). Gajiganna und Konduga -- Zur frtihen Besiedlung des Tschadbeckens in Nigeria. Bericht fiber die Ausgrabungen des Frankfurter Sonderforschungsbereichs 268 ("Westafrikanische Savanne") in Borno, Nordost-Nigeria. Beitriige zur Allgemeinen und Vergleichenden Archiiologie 14 (in press).144 Breunig, Neumann, and Van NeerBreunig, P., and Wotzka, H.-P. (1993). Arch~ologische Forschungen im Stidosten Burkina Fasos 1989/90: Vorbericht tiber die erste Grabungskampagne des Frankfurter Sonderforschungsbereiches 268 "Westafrikanische Savanne." Beitrdge zur Allgemeinen und Vergleichenden Archiiologie 11: 145-187.Breunig, P., Garba, A., and Waziri, 1. (1992). Recent archaeological surveys in Borno, northeast Nigeria. Nyame Akuma 37: 10-17.Breunig, P., Ballouche, A., Neumann, K., Rfsing, F. W., Thiemeyer, H., Wendt, K.-P., and Van Neer, W. (1993a). Gajiganna -- New data on early settlement and environment in the Chad Basin. Berichte des Sonderforschungsbereichs 268, Vol. 2, Frankfurt a.M., pp. 51-74.Breunig, P., Garba, A., Gronenborn, D., Van Neer, W., and Wendt, P. (1993b). Report on excavations at Gajiganna, Borno State, northeast Nigeria. Nyame Akuma 40: 30-41.Connah, G. (1976). The Daima sequence and the prehistoric chronology of the Lake Chad region of Nigeria. Journal of Afn'can History 17(3): 321-352.Connah, G. (1981). Three Thousand Years in Afn'ca: Man and his Environment in the Lake Chad Region of Nigeria, Cambridge University Press, Cambridge.Connah, G. (1984). An archaeological exploration in southern Borno. Afncan Archaeological Review 2: 153-171.Connah, G. and Freeth, S. J. (1989). A commodity problem in prehistoric Borno. Sahara 2: 7-20.Connah, G., and Jemkur, J. (1982). Prospecting the 3000 B.P. barrier. Borno 1981. Nyame Akuma 20: 35-43.Davies, O. (1980). The Ntereso culture in Ghana. In Swartz, B. K., Jr., and Dumett, R. A. (eds.), West African Culture Dynamics."Archaeological and Historical Perspectives, Mouton, The Hague, pp. 205-225.Ellmers, D. (1984). The earliest evidence for skinboats in late-palaeolithic Europe. In McGrail, S. (ed.), Maritime Archaeology and Ethnology, Trustees of the National Maritime Museum, London, pp. 41-55.Flight, C. (1976). The Kintampo Culture and its place in the economic prehistory of West Africa. In Harlan, J. R., de Wet, J. M. J., and Stemler, A. B. L. (eds.), Origins of African Plant Domestication, Mouton, The Hague, pp. 211-221.Garba, A. (1993). The origin of the stone raw materials from Gajiganna. Berichte des Sonderforschungsbereichs 268, Vol. 2, Frankfurt a.M., pp. 133-144.Gautier, A. (1987). Prehistoric men and cattle in North Africa: A dearth of data and a surfeit of models. In Close, A. E. (ed.), Prehistory of Arid North Africa, Southern Methodist University Press, Dallas, pp. 163-187.Grove, A. T. (1970). Rise and fall of Lake Chad. The Geographical Magazine March: 432-439.Grove, A. T. (1985). The physical evolution of the river basins. In Grove, A. T. (ed.), The Niger and its Neighbours. Environmental History and Hydrobiology, Human Use and Health Hazards of the Major West African Rivers, Balkema, Rotterdam, Boston, pp. 21-60.Happold, D. C. D. (1987). The Mammals of Nigeria, Clarendon Press, Oxford.Harlan, J. R. (1971). Agricultural origins: Centers and noncenters. Science 174: 468-474.Irvine, F. R. (1961). Woody Plants of Ghana, Oxford University Press, London.Keay, R. W. J. (1959). Trees of Nigeria, Clarendon Press, Oxford.Krzyzaniak, L. (1992). The later prehistory of the upper (main) Nile: Comments on the current state of the research. In Klees, F., and Kuper, R. (eds.), New Light on the Northeast African Past, Africa Praehistorica 5, Heinrich-Barth-Institut, KOln, pp. 239-248.L6zine, A.-M. (1989). Late Quaternary vegetation and climate of the Sahel. Quaternao, Research 32: 317-334.MacDonald, K. C., and Allsworth-Jones, P. (1994).. A reconsideration of the West African macrolithic conundrum: New factory sites and an associated settlement in the Vallfe du Serpent, Mali. African Archaeological Review 12: 73-104.Mclntosh, S. K., and Mclntosh, R. Ji (1983). Current Directions in West African Prehistory. Annual Re~iew~ of Anthropology 12: 2.,1)5-258.Munson, P. J. (1976). Archaeological data on the origins of cultivation in the southwestern Sahara and their implications for West Africa. In Harlan, J. R., de Wet, J. M. J., and Stemler, A. B. L. (eds.), Origins of African Plant Domestication, Mouton, The Hague, pp. 187-210.Holocene Settlement and Environment of the Chad Basin 145Neumann, K., and Ballouche, A. (1992). Die Chaine de Gobnangou in SE Burkina Faso -- Ein Beitrag zur Vegetationsgeschichte des Sudanzone W-Afrikas. Geobotanische Kolloquien 8: 53-68.Redding, R. W. (1984). Theoretical determinants of a herder's decisions: Modeling variation in the sheep/goat ratio. In Clutton-Brock, J., and Grigson, C. (eds.), Animals and Archaeology: 3. Early Herders and Their FlOcks, British Archaeological Reports, International Series 202, Oxford, pp. 161-170.Schuck, W. (1990). Chronologische Aspekte zu prahistorischem Fundmaterial aus der 6stlichen Zentralsahara. In Gabriel, B. (ed.), Forschungen in ariden Gebieten, Berliner geographische Studien 30; Institiit for Geographic der Technischen Universit~it, Berlin, pp. 123-142.Shaw, T. (1978-1979). Holocene adaptations in West Africa: The Late Stone Age. Early Man News 3/4:51-82.Shaw, T. (1980). Hunters, gatherers and first farmers in West Africa. In Megaw, J. V. S. (ed.), Hunters, Gatherers and First Farmers Beyond Europe, Leicester University Press, Leicester, 2rid impression, pp. 69-125.Shaw, T. (1981). The Late Stone Age in West Africa and the beginnings of African food production. In Roubet, C., Hugot, H.-J., and Souville, G. (eds.), Pr~histoire Africaine. Mdlanges offerts au doyen Lionel Balout. Recherches sur les grandes civilisations, Synth~se no. 6, Edition ADPF, Paris, pp. 213-235.Sinclair, P. J. J., Shaw, T., and Andah, B. (1993). Introduction. In Shaw, T., Sinclair, P., Andah, B., and Okpoko, A. (eds.), The Archaeology of A]h'ca. Food, Metals and Towns, One World Archaeology 20, Routledge, London & New York, pp. 1-31.Smith, A. B. (1984). Origins of the Neolithic in the Sahara. In Clark, J. D., and Brandt, S. A. (eds.), From Hunters to Farmers. The Causes and Consequences of Food Production hi Africa, University of California Press, Berkeley & Los Angeles, pp. 84-92.Thiemeyer, H. (1992). On the age of the Bama Ridge -- A new 14C-record from the Konduga area, Borno State, NE-Nigeria. Zeitschrift fi]r Geomorphologie N.F. 36( 1): 113-118.Thiemeyer, H. (1993). A new 14C-record from the Bama Ridge near Konduga, Borne State, NE-Nigeria. Annals of Bomo 8/9 (1991-1992): 239-242.Tromnau, G. (1987). Late palaeolithic reindeer-hunting and the use of boats. Polska Akaderni,t NAUK 5: 94-104.Tuley, P. (ed.). (1972). The Land Resources of North East Nigeria, Vol. 1, Land Resource Division, Surbiton.Van Neer, W. (1989). Recent and fossil fish from the Sahara and their palaeohydrological meaning. Palaeoecology of Africa 20: 1-18.Van Zeist, W. (1957). De mesolithische boot van Pesse. Nieuwe Drente Volksalmanach 75: 4-11.yon Maydell, H.-J. (1986). Trees and Shrubs of the SaheL Margraf, We~ersheim.Wendt, P. (1995). Magerung und O b e r f l ~ i c h e n b e h a n d l u n g - - Z u r chronologischen Interpretation technischer Merkmale in der Keramikentwicklung in .NE-Nigeria. Berichte des Sonderforschungsbereichs 268, VoL 5, Frankfurt a.M., pp. 41-47.York, R. N. (1978). Excavations at Dutsen Kongba, Plateau State, Nigeria. West African Journal of Archaeology 8: 139-163.