nyame akuma issue 039

8/10/2019 Nyame Akuma Issue 039

1/80

NY ME AKUMA No. 39 June 1993

It appears likely that a substantialnumber and variety of conferences onAfrican archaeology will have beencelebrated in the three-year span betweenSeptember 1992 and September 1995. Thisencouraging development is illustrated inthe present issue of Nyame Akuma, where thereader will find announcements of sixconferences, two of which deserve special

mention. First and foremost, of course, arethe forthcoming biennial meetings of SAfA,the parent organization of this bulletin.

The SAfA conference is scheduled to beheld at Indiana University, April 28 to May1,1994, under the organizational guidanceof Kathy Schick and Nick Toth. In additionto the rich scientific program that hasgenerally characterized these meetings, theagenda will undoubtedly also include aseries of important issues to be considered inthe business session, such as the possibilityof a European venue for the 19% conference.For more detailed information on the 1994SAfA meetings, please s the entry in the

Meetings section of this issue.The same section also contains the first

announcement of the longawaited 10th PanAfrican Congress of Prehistory and RelatedStudies, which is scheduled to take place inHarare during September, 1995. A strongvote of thanks for undertaking the congressis owed to the Department of History at theUniversity of Zimbabwe, particularly Prof.Pwiti, to whom the bulletin extends anopen-ended offer to assist by any means atour disposal in the process of organizing thecongress.

On a completely different subject,readers may be interested to know about anew graduate program in archaeology at theUniversity of Minnesota. The program w ll

offer M.A., M.S. and Ph.D. degrees inInterdisciplinary Archaeological Studies andwill require the completion of a curriculumin archaeology plus one other related field,such as geology, ecology, ancient history,etc.. accordine to the student's interests.

Within the archaeology curriculum, therewill be two distinct emphases, one onarchaeological saence and the other oncultural resource management. These

emphases, together with the absence of anyrequirement in anthropology, may renderthe IAS program especially interesting toarcheologists n frican countries.

Students may enroll in the IAS programat either the win Cities (Minneapolis, SaintPaul) or Duluth campus of the University ofMinnesota. The program will be initiated inautumn, 1993, and applicaticms are currentlybeing accepted. Inquiries should beaddressed to

Prof. George RappDirector Archaeometry LaboratoryThe University of MinnesotaDuluth, N 55812-2496U. S.A.Fax: 218-7264181 /8275

Finally, an erratum the copyright date of1990 was inadvertently omitted from thereview of John Sutton's A housand Years ofEast frica in Nyame Akurna 38


2/80

NY ME AKUMA No. 39 June 1993

CAMEROON

From Clay to Pottery withStyle: 1990-1 992 Fieldworkin Cameroon

Olivier P, GosselainMusde Royal de I Afrique Centrale8 3080 ervuren, Belgique

Despite a continuously renewedethnoarchaeological interest in pottery,studies concerned with technique remainoccasional or lack a systematic approach anda geographical dimension. Comparingtechniques at a regional level, or even at anethnic level, is often difficult due to thenumber and dispersion of potters studiedfor an exception s Barreteau and Delneuf

1990, Delneuf 1991). Also, factors linkingtechniques and culture are rarely exploredin the ethnoarchaeology of pottery.Technical processes are still considered to begoverned by environmental and functionalconstraints Arnold 1985, Rice 1987) andartisans are too often assumed to beobsessed with the physical and chemicalnature of their pots.

Using the isochrestic definition of styleproposed by Sacket 1982, 1990), andattempting to evaluate the degree of culturalintegration of techniques Lemonnier 1986,1990, 19911, I have made a stylisticcomparison of pottery technical processes

c h i n e s opiratoire) among several ethnicgroups in Cameroon. This research focusedboth on the diversity of technical optionschosen by the artisans and on explaining thereasons underlying their choices.

The preliminary results show thattraditional pottery techniques actuallydepend very little on functional andenvironmental constraints. Each technical

process is almost completely inter-changeable between societies and ecologicalcontexts. Technical choices are especiallydictated by the social environment of the

artisan modes of transmitting knowledge,relationships with other technical systemsand the symbolism of technical actions). Thenotion of technological style Lechtman1977, Lechtman and Steinberg 1979, Childs1986), or more precisely, of technical stylesee Ingold 1990), seems therefore

particularly well adapted, and this permitsus to develop new directions in archaeologyChilds 1991,1992; Gosselain 1992a, c).

Fieldwork

Cameroon offers a excellent opportunityfor a comparative study of techniques due toits ethnic and ecological diversity. I chose towork in the southern part of the countrybecause the pottery making has been littledocumented as opposed to the northernpart: Barreteau and Delneuf 1990, David andHennig 1972, Delneuf 1991, Sterner 1989)and because it has a clear archaeological and

linguistic interest: it served as the point ofdeparture of the Bantu expansion, it has agreat variety of ethnic groups, and it is thecrossroad of several linguistic groups.

During summer of 1990 and spring of1991 and 1992, I observed a number ofpotters in the Central, Western,Northwestern, Littoral, Adamawa andEastern provinces. In total, 82 artisans werestudied in 44 villages Fig. 1 and Table 1).These artisans belong to 21 ethnic groups,representing 7 different linguistic groupsDieu and Renaud 1983): Narrow Bantu,

Ring Grassfields, Eastern Grassfields,Mambiloid, Tikar, Oubanguian andAdamawa the first five groups are part ofthe Wide Bantu) see Table 1). My intentionwas to cover the entire area occupied byeach of these groups. Despite a survey tourof more than 2,500 km he geographicaldistribution of the artisans under studyremains unequal especially for groups suchas the Basaa, Bamum, and Gbaya).

This is explained by the disappearanceof pottery production among these groups,the inaccessibility of certain areas, and fieldseason deadlines. The study area, however,


3/80


4/80

NY ME KUM No. 39 June 1993

Table 1 Data on potters studied in 1990 1991 and 1992. Nbr: number of potters.Letters in parentheses following a village name indicate that the potterlives in a village of another ethnic group: Ma-Mambila Bf-Bafeuk SA-Sanaga Ya-Yambassa Kw-Kwakum Ke-Kepere G b G b a y a Vu-Vute.

Ethnic Groups

Nsei

Yamba

Bamum

Barnileke

Mambila

Vute

Banen

Basaa

Bafeuk

Djanti

Balom

Baf ia

Sanaga

YambasaEton

Kwakum

Pol

Gba ya

Kepere

Languages

kanswei nsei

yamba

shtt pamam

fe'fe'

mambila

w e

tik ri

tunm

basaa

bafak

a

lafa'

rkpa'

tLlki

yambasaeton

kwakum

pori

kak)

gbaya

gt' E

Linguistic Gmups

Ring Grassfields

Eastern Grassfields

Eastern Grassfields

Eastern Grassfields

Mambiloi'd

Mambiloi'd

Narrow Bantu

Narrow Bantu

Narrow Bantu

Narrow Bantu

Narrow Bantu

Narrow Bantu

Narrow Bantu

NarrowBantu

Narrow Bantu

Narrow Bantu

Narrow Bantu

Narrow Bantu

Oubanguaian

Adamawa

Nbr1

4

3

4

4

2

2

9

I3

2

3

2

Villages

Mbui, Mbem,Sarki Mbaka (Ma)Mamarom

Babwantou

Atta, Ngonnkor,

KouroumYangba, Nguila(Bf), Mangay,D o m e

Mambioko,Akouen, Kong,Massaroum

Ndikinimeki,Logndeng

Likound, Mbay

Yassem, Ndimi,Nguila

Ngora Sa)

Mouzi, Beandong

Goufan I, Kiki,Biamo, Mouko,Bouraka (Ya)

Ngoro, Nguette

KellendeEmana, Kela ( I,Kouse Sa)

Hona

Hona (Kw), Dengdeng (Ke)

Dai'guene (Gb),Momdjomo

Manbarla, Yoko(Vu), Moi'nan,Adinkol

Deng deng


5/80


becomes larger when the learning places ofthe artisans are included. For example, someof the artisans learned pottery techniques ina village other than that where they wereobserved. Because of the deeply rootedcharacter and resistance to change of certainstages of the technical process Gosselain1992a1, one may reasonably assume that theobserved techniques were also practised inthese learning places.

In the field, working with the artisansconsisted of a detailed observation of theentire technical process from theprocurement of raw materials until the potsare put into circulation); an interviewcovering the biography of each potter, themodalities of his/her learning process, andthe non-technical aspects of the activity

taboos, rituals, economic situation,determination of the reasons underlyinghis/her technical choices, and thevocabulary of the activity; a series ofmeasurements plasticity of the clays beforeand after treatment, firing temperatures-27firings were recorded with 3 to 11thermocouples; see some results inGosselain 1992b); and an extensive sampling

pots, tools, clays, nonplastic additives,ks used for post-firing keatments).

So far the results of the comparativestudy indicate that although none of thetechnical choices is truly original, thecombination of these choices in the chaineope ra to ir e permits differentiation of theartisans. The shaping stage proves to beparticularly relevant for differentiatingethnic groups. In total, nine shapingtechniques were identified and theirdistribution correlates quite well that oflinguistic groups Gosselain 1992~). Thestudy of the economic context shows thatmost of the groups have a domestic mode ofproduction. The only exception are theYamba, among whom the activity isrestricted to men in a aaft specialisation orsemi-specialisation context.

LaboratoryMost of the analyses are still in progress.

The clay and pottery samples were, andcontinue to be, analysed at the UniversitkLibre de Bruxelles Impedovo 1991, Yachou

19911, at the University of Oxford, and atTervuren Museum. The focus of suchanalysis is on the comparison of chemical,mineral, textural, and physical properties,and on confronting these with the empiricalreasoning of the artisan. The other aim is totest the feasibility and relevance ofprovenience studies in this part of theAfrican continent. Other analyses are beingconducted on the same pottery in order toidentify the traces of technical processesradiography, thin sections, and scanning

electron microscopy).Finally, organic coatings are being

analysed and experimentally tested in orderto evaluate the use of the pots.

Along with this laboratory work, aseries of typological experiments ondecorative systems and morphologicalparameters are being carried out. Theseexperiments will eventually lead to aselection of criteria for identifying differentproduction levels individual, community,ethnic, and regional levels), as well as forrecognizing the taxonomic systems used bythe artisans Gosselain and van Berg, inpress)

cknowledgements

This research was conducted under thedirection of P. de Maret and entirely fundedby the Universitb Libre de Bruxelles. InCameroon, research authorizations wereissued by the Ministry of Higher Education,Information and Scientific Research incollaboration with the Institute of HumanSciences. I would like to thank the numerouspeople who have contributed to my researchwith their support, criticism and advice, andalso all potters who kindly and patientlyresponded to my enquiries.

References

Arnold, D E

1985 Ceramic Theory and Cultural Process.Cambridge: Cambridge UniversityPress.

Barreteau, D., and Delneuf, M1990 La ceramique traditionnelle Giziga et

Mofu Nord-Cameroun): etude


6/80

NY ME KUIM No. 39 June 1993

cornpar des techniques, des formes td u vocabulaire. In Re1 a t o n sInterethniques et Culture Math icllc dunsle Passin du Lac Tchad. Actes du cColloquc Mega-Tchad, 12-12 septmrbtr1986. Park ORSTOM, pp. 121-49.

Childs, S. T.

Style in Technology: A View of Aficanarly m Age m Smd ting through its

Refmctory Ceramics. Ph.D. dissertation,Ann Arbor: University MicrofilmsIntemationaL

Style, technology, and iron smelting

furnaces in Bantu-speaking Africa.Journal o Anthropological rchaeology 10:332-59.

Transformations: iron an d copperproduction in Central Africa. InGlumac, P. (ed) , Recent Trends inA r c h a ~ t a l l u r g i c a l Research. MascaResearch Papers in Science andArchaeology, VoL 8, Part 1, pp. 3346.

David, N., and Hennig, H

1972 The ethnography of pottery: a Fulanicase seen in atrhaeological perspective.McCaleb Module n Anthropology 21: 1-29.

Delneuf, M.

1991 Un cha mps par t icu l ie r d el'expbrimentation en cdramique: lesateliers de poterie traditiomelle duNord Cameroun. In Xle RencontresInternationales d Archhlog ie t d Histoired Antibe: 25 ans de Technologie enPrthistoim. Juan-les-Pins: APDCA, pp.65-82

Dieu, M. and Renaud, P. (eds.)

1983 Situation Linguistique en AfriqueCentrale. lnvcntaire Prtliminaire: leC a m e r o u n . Atlas Linguistique del'Afrique Centrale (ALAC). Yaounde:ACCT-CERDOTOLA-DGRST.

Gosselain, 0 .

1992a Technology and style: potters andpottery among Bafia of Cameraon Man27 3): 559-86.

1992b Bonfire of the enquiries. Pottery firingtemperatures in archaeology: what for?Journal of ArchaeologiCP1 Science 19 3):243-59.

1992c Stylistic and archaeometricalapproaches to present potterytechnology in Cameroon. Paperpresented at the 1992 SAfA BiennalMeeting, Los Angela, March 26-29.

Gosselain, 0 P., and van Berg P.-L

in press Style, individualite et taxonomie chezles potieres Bafia du Cameroun.B u l l e t i n d u C e n t r e G e n e v o i s

d Anthropologie3

Ingold, T.

1990 Society, nature and the concept oftechnology. Archaeological eview fromCambridge9 1): -17.

Impedovo, V. R1991 Simulation de la Technologie des

Ceramiques Anciennes. Mdmoire deLicence. Universitb Libre de Bnucelles.

1977 Style in technology. Some earlythoughts. In Lechtman, H. and Memll,R S. (eds), Material Culture: Styles,Organization, and Dynamics ofTechnology. St. Paul: West PublishingCompany, pp. 3-20.

Lechtman, H., and Steinberg, A

1979 The history of technology: ananthropological point of view. InBugliarello, G. and Doner, D. B. (eds.),The Histmy and Philosophy o Technology.Urbana: University of Illinois Press, pp.135-60.

Lemonnier, P.

1986 The study of material culture today:toward an anthropology of technicalsystems. Journal of Anthropological

Archaeology 5: 147-86.1990 Topsy turvy techniques. Remarks on

the social =presentation of techniques.Archaeological eview from Cambridge91): 7-37.


7/80

N YA M E A K U M A N o 39 June 1993

1991 e la culture materielle la culture?Ethnologie des techniques etprkhistoire. In Xle RencontresInternationales d'ArchdoJogie et Histoire

Antibe: 25 ans de Technologie enPrehistoire. Juan-ledins: APDCA, pp.15-20.

Rice P

1987 Pottery Analysis: A Sourcebook.Chicago:The University of Chicago Press.

Sackett, J . R1982 App roache s to style in lithic

archaeology. JoumI of AnthropologicalAldmcology1: 59-1121990 Style and ethnicity in archaeology: the

c se for isochrestism. In Conkey, M. W.and Hastorf, C. A. (eds), he Uses ofStyle in Archaeology. Cambridge:Cambridge University Press, pp. 3243

Sterner, J

1989 Who is signalling whom? Ceramicstyle, ethnicity, and taphonomy among

the S i k Bulahay. Antiquity 63.45259.

Yachou, M1991 Differmtiation des Argiks et des Poteries

par Annlyse Chimique et Midralogique.Travail d e Fin d gtud e. I.E.S.T.E.Bruxelles.

Archaeological Research inNorthern Cameroon 1992-the Projet Maya-Wandala

Scott MacEachernDepartment of ArchaeologyUniversity of CalgaryCalgary Alberta CanadaT N N

The first field season of the Projet Maya-Wandala took place between 16 May and 31August, 1992, in the Extreme-Nord Provinceof Cameroon. The a e w for this season

included Scott MacEachern (project head),Claire Bourges, Ross Jamieson, and LaurieBeckwith from the University of Calgary,Canada. In addition, three Cameroonianarchaeology students from the University ofYaounde, Djime Rama, Rodger Ndombe andGerald Lamba worked with the project from27 July to 24 August, under an agreement bywhich the Projet Maya-Wandala providedthem with training. Finally, RaymondAsombang, from the faculty of t heUniversity of Yaounde, visited the researcharea and inspected sites in early August.

The Projet Maya-Wandala wasconceived of as a successor project, buildingupon the archaeological and ethnoarchae-ological research carried out by members ofthe Mandara Archaeological Project/RojetArcheologique Mandara (includingMacEachem) between 1984 an d 1991 (Davidand MacEachern 1988, David and Sterner1989). Research was conducted from Mora,in the DCpartement of Mayo-Sava, andfocused upon archaeological survey of andexcavation around the numerous inselbergsextending north from the MandaraMountains, which have their northernmostextension in this region. In addition,ethnohistorical inquiries were an importantcomponent of this season s research. Thereis some overlap between the MandaraArchaeological Project archaeological surveyin 1984 and the Projet Maya-Wandalasurvey area in 1992. This happened becauseof different research strategies in the twoseasons; our 1984 Mandara ArchaeologicalProject research (David and MacEachern

1988) was oriented toward establishing aregional culture-historical sequence, while1992 Projet Maya-Wandala research wasconsiderably more narrowly focused.

Research Goals

The objective of this work was toelucidate the historical relationshipsbetween populations found in this areatoday and in the recent past (i.e., over the

last 150-200 years), and earlier populations,now remembered as the semi-legendary Saoand Maya peoples (along with several local,even lesser-known groups). The ethnicmilieu in the area today is dominated by thedichotomy between plains-dwelling


8/80

No. 39 June 1993

populations (overwhelmingly Muslim,urban, literate, and organized into smallchiefdoms and states) and montagnardgroups (mostly animist and/or Christian,rural and, in this area, organized in small,acephalous ethnic units). The mountains aredensely settled, the plains much less so. Theresults of Mandara Archaeological Projectresearch in the 1980s indicated that thissituation had been reversed before perhaps200 years ago, with substantial Sao andMaya settlement on the plains and relativelysparsely occupied mountains. We wished tostudy the processes by which these earlierIron Age plains populations had beensupplanted by later groups, and especiallyby the Wandala, inhabitants of a small stateof the same name who linguistically andculturally seem to have resided in the regionfor a long time. The Wandala have beenMuslim for more than 250 years, and theirassociations with montagnards have oftenbeen ambiguous (MacEachern 1991,1992, inpress).

To that end, we concentrated ourresearch on areas which are associated inoral traditions with both Sao/Maya andWandala occupation. Wandala settlementsare usually associated with the inselbergswhich do t the plains north off the MandaraMountains, so it is not surprising that mostof our work took place around three of theseinselbergs, Grba, Doulo, and Aissa Hard6.Gr6a figures in a number of Wandalawritten histories and oral traditions, as animportant Sao centre and later as a largeWandala village, albeit one that retainedconsiderable independence from Wandalacentralized authority. Do do was the capitalof the Wandala state from the sixteenth untilthe nineteenth century, and it was the centreof a Maya chiefdom before that. AissaHard though much smaller than the othertwo inselbergs, was also desaibed in localhistories and oral traditions as an importantMaya and Wandala centre.

esearchesults

Field work over the period I June-25August, 1992, resulted in the discovery of 65archaeological sites (Fig. 1). In addition,surface collections were made in sevenlocalities where artefacts were found but

where their densities did not warrantdesignation as a site. Determination of theages of these sites must await radiocarbonand thennoluminescence analysis of therecovered material, but general examinationof the pottery and other artefacts indicatesthat the vast majority of sites and thecomponents that make them up are datableto the Iron Age, and probably to the last1,000-1,200 years. However, we have founda number of components (areas an d/orlevels within a site) that may date to theNeolithic period (before ca 1 A.D. , and twocomponents which may date to the LateStone Age. The latter are, unfortunately,probably undatable, since they are merelysurface scatters of eroded stone tools. In anycase, the objective of this project was thestudy of Iron Age cultural sequences, so wedid not extensively pursue analysis of theseearlier components.

Of the sites recorded, 27 were located eitherat the foot of or on the heights of Gr6ainselberg, and two were located about2.5 km away from the inselberg in an areaknown for its Sao occupation (Fig. 2).Similar sites were found in the same areaduring Mandara Archaeological Projectarchaeological survey in 1984. Gr6a wasobviously n important Iron Age habitationarea, as ethnohistorical and historicalsources also indicate. Manaouatchi-Gr6a(PMW 6021, with related sites, was anextremely large village (at least 2 x 1.5 k msurrounded by a 3 km long cobble wall thatprobably stood 1-1.5 m high. Gr6a inselbergitself was a long-standing focus foroccupation by Mafa montagnard groups;Gr6a Tlala-Mafa (PMW 603) is a large ridge-top site overlooking Manaouatchi-Gr6a.Eleven sites were found in the immediatevicinity of Doulo inselberg, and one in a drystream bed about 2 km to the east, threewere found around Aissa Hard6 (includingtwo very large, deep mound sites) and threemore in the immediate vicinity. Six werefound at the foot of or near Sera Ourdainselberg, and 12 were found close tosmaller inselbergs in the area (Mahola,Wawidive, Wakashimre, Sava, and anumber of others).

The concentration of sites around theedges of inselbergs is very noticeable. Projet


9/80

No 39 une 1993

Our d

ow i ve

: S Co u l o o r d e

Fig. 1 Sites in the plains north of the Mandara Mountains includes sites discoveredduring 1 98 4 and 1 99 2 surveys).


10/80


I V ~ s t b l e r e ar e a O u t e r

I W a l l

Fig. 2. Sites around Gr6a inselberg wi th the Gr6a wa ll feature indicated.

10


11/80

NY ME KUM No. 39 une 7993

Maya-Wandala 1992 research resulted in nodiscovery of large sites in the plains at morethan 1 k m distance from inselberg edges;even the two Sao mound sites found near

GrLa are small compared to many of theinselberg-edge sites. These findings contrastwith the results of the 1984 MandaraArchaeological Project research at andaround Mehe Djiddere, about 8 m east ofthe eastern edge of the Projet Maya-Wandala survey area, where three Iron Agemound complexes were discovered in theplains and one excavated. It should benoted, however, that extensive stratified-random survey by Mandara Archaeological

Project field crews in 1984 did not lead to thedetection of large Iron Age plains sites in thearea where Projet Maya-Wandala andMandara Archaeological Project surveyoverlapped, which reinforces our 1992findings.

It is not as yet clear whether our 1992results indicate (a) that inselberg edges wereactually vastly favoured over areas furtherout on the plains for Iron Age sites or (b)that depositional processes have concealedand/or destroyed sites which exist(ed1 onthe plains during the Iron Age. The latterscenario is possible, but we shouldremember that the depositional/erosionalprocesses that might be expected to affectplains sites are to a great extent caused byrunoff from inselbergs and the Mandaramassif itself. Under those circumstances,inselberg-edge sites should have been atleast as heavily impacted as sites further outin the plains. This also has implications for

our understanding of Neolithic/Iron Agepopulation shifts, since there seems to be arelative lack of Neolithic sites on the plains.

We excavated six of the sites found.These were ( in order of excavation)Mmaouatchi-GrLa (PMW 602-two 2 m x 2m pits, plus a number of shovel tests), GrLaTwin Peaks (PMW 618-0ne 2 m x 1 m pit),Doulo Igzawa I (PMW 63 6- 0n e 2 m x 2 m,one m x 1 m pit, plus shovel tests), DodoIgzawa 3 (PMW 652-two 1 m 1 m pits),Doulo Kwovre (PMW 63 1- on e 2 m x 1.5 mpit, plus shovel tests) and Doulo Chefferie(PMW 63 5- on e 2 m x 1 m pit started, butnot finished). All units were excavated tosterile levels except for Doulo Chefferie,

which was halted due to opposition fromone local Wandala noble. Given that thisoccurred in the last week of field work, wedecided not to restart excavation at this site,

leaving the pit uncompleted at 40 cm bd.Site depths varied widely. Our deepest unitwas PMW 602 Pit 2, excavated on a moundfeature on the very complex and diverse siteof Manaouatchi Grt5a; sterile levels werereached at 2.5 m bd, and we excavated to2.85 m bd Average site depth was 1.5 m.

Firm conclusions derived from theseexcavations must await artefact and faunalanalysis and site dating. Nevertheless, wecan make some preliminary judgments. In

the first place, it is fairly obvious that IronAge groups have lived around Gr6a andDoulo inselbergs at least for a considerableamount of time, perhaps 1,000 years atManaouatchi-GrLa and 500 years (andconceivably much longer) a t Doulo Igzawa 1and Doulo Kwovre. Site depths aresignificant in these cases, and evolution inceramic styles are easily detectable. It ispossible that the lower levels of these threesites date to the end of the Neolithic or early

Iron Age. The Iron Age settlement patternaround both inselbergs (extensivesettlements spread along slope edges) isdifferent from that of historical Wandalapeoples, who settled in defensible, nucleatedtowns like Manaouatchi-Grda and Doulo; itis also the pattern ethnohistorically linked tothe Sao and Maya, who do not seem to havesited their communities with defense as afirst priority.

n the second place, there are significantdifferences detectable between sites aroundboth Grda and Doulo inselbergs; thedifficulty is telling whether these are due todifferent site ages, site uses, or site ethnicoccupations---or to some combination ofthese (and other) factors. For example, theceramic and lithic assemblages from theGrda Twin Peaks site are very different fromthose from Manaouatchi-Grda, even thoughthe two sites are only a kilometer apart. Thisis probably due to ethnic (Mafa or proto-Mafa at GrLa Twin Peaks ver sus

Wandala a t Manaouatchi-Grda) andoccupational (stone bracelet and grindstoneproduction at the former site, no such workat the latter) differentiation between the


12/80

NY ME KUNI No 39, June 7993

sites. Similarly, although both Doulo andGrCa are now occupied by Wandalapopulat ions, the decorat ion andmorphology of prehistoric ceramics

probably related to proto-Wandalaoccupations from both areas are quitedifferent.

Substantial ethnohistorical data werealso gathered in the course of this season'sfieldwork. We gathered data frominformants on the location of sites, on sitefunctions and on the political and socialrelationships between different sites. Thissort of investigation is, of course, mostvaluable and reliable when concerned withrelatively recent sites; the recent Iron Ageand Wandala sites that we worked on aregood candidates. Thus, informants wereunanimous in saying that the Sao/Mayawere not warlike people, that this wasexpressed in their site settlement systemsand that this was the reason that theWandala were able to defeat them. Localtraditions also hold that conflicts betweendifferent Wandala centres were quitecommon, and that especially the settlementat Gr6a was frequently at odds with themore orthodox Wandala centres, probablybecause of its strong Sao traditions.

Traditions about Mafa montagnardoccupation on the heights of Gr6a area arealso of interest. Oral traditions state thatMafa arrived at Gr6a just after the arrival ofthe Wandala people, which, i confirmed,would indicate an occupation at least 250-300 years old. Mafa communities were

traditionally located in the MandaraMountains to the south, and not on isolatedinselbergs; this is the northernmostextension of Mafa temtory. Such traditionsalso indicate fairly close, and constant,cooperation between Mafa and thedominant Wandala. This situation is quiteunusual in this area; I am now examiningWandala-montagnard interactions aroundthe isolated inselbergs (Keroua, Grka, Doulo,Urza) where important Wandala settlements

are located.

Conclusions

As have said, only preliminary analysisof materials recovered from this season's

research has been completed, and anyconclusions drawn at this point must ofnecessity also be preliminary. However,available data and ethnohistorical

information gathered during this fieldseason have implications for our ideas aboutthe makeup of the Wandala state. There is atendency to think of traditional states asorganized along modem models-that is,unitary, monolithic, and with well-definedborders and internal organizations. Wemight tentatively conclude that the Wandalastate was set up according to a differentorganizational model, one in which differentimportant settlements-to some extent

isolated geographically and conceptually bytheir proximity to inselbergs on the plains,and possessing considerable defensivecapabilities for the same reason-were onlyloosely governed by the ruling lineages ofthe Wandala. In addition, our excavationsindicate that there was considerable culturalcontinuity from the Sao and Maya to theWandala occupations at both Gr6a andDoulo.

cknowledgements

The first season of field research by theProjet Maya-Wandala was very rewarding.Our three-year programme of research isfunded by the Social Sciences andHumanities Research Council of Canada(Grant number 410-92-1860). I would like tothank the Government of Cameroon, andespecially the Ministry for Scientific andTechnical Research, for their assistance in

facilitating our work. would also like tothank a very fine field crew, bothCameroonian and Canadian, for their effortsin making this season a success. I lookforward to the next Projet Maya-Wandalafield season, just across the Cameroon-Nigeria border in Nigeria, between May andAugust, 1993.

References

David, N., and MacEachern, S.1988 The Mandara Archaeological Project:

preliminary results of the 984 season.In Barreteau, D and Toumeux, Heds.), e mili u t 2es hommes: recherchs

comparitives t historiques duns l bassin


13/80

N YA M E A K U M A N o 39 June 1993

du Lac Tchad. Collection colloques etdminaires. gditions de I ORSTOM,Paris, pp. 51-80.

David, N., and Sterner, J

1989 The Mandara Archaeological Project1988-89. N y a m Akuma 3 : 5 9.

MacEachern, S.

Du Kunde: Processes of MontagnardEtnogenesis in the Northern MandaraMountains of Cameroon. UnpublishedPh.D. dissertation, Department ofArchaeology, University of Calgary.

Les gens de Ngolele: an examination ofprehistoric ethnic relations in thenorthern Mandara Mountains. InBoutrais, J (ed.), D u po li ti que d1'Cconomiqw: u s historiques dans lebassin du Lac Tchad. C o l l e c t i o ncolloques et seminaires. gditions deI ORSTOM, Paris, pp. 165-92

Selling the iron for their shackles:Wandala-montagnard ethnic relationsin northern Cameroon ]ournnl of AfncanHistory.

Preliminary Results of the1991-1992 Field Season a tShum Laka NorthwesternProvince Cameroon

P. de MaretUniversitd Libre de BruxellesAnthropologie CP 12450 Av. F RooseveltB- 1050 BruxellesBelgique

R. Asombang

Unversitd de Yaoundd8. P 6544Cameroun

E Cornelissen, Ph. Lavachery,J Moeyersons, and W Van NeerMusde Royal de / A rique Centrale8-3080 TervurenBelgique

Shum Laka, a huge rockshelter of some1,200 m2 is situated in the NorthwesternProvince of Cameroon. This area is thoughtto be the homeland of Bantu-speakingpeople, whose archaeological past is stillpoorly known. Test pits from 1981 and 1984

de Maret et al. 1987, Asombang 1988)revealed the potential of the rockshelter forinterdisciplinary research to identify moreprecisely the nature of cultural andenvironmental changes during the last10,000 years.

The preliminary resu lts of geomor-phological research, conducted in 1991-92,indicate that the Shum Laka rock shelterprobably originated as a weathering pocketin a cliff of tertiary basaltic tuffs. Followingthe evacuation of fines, the original cavitybecame larger and deeper due to subsequentphases of weathering and removal of loosematerial. The bedrock floor has an irregularshape and is covered by deposits derivingfrom chemical and physical weathering ofthe overhang. The resulting 7.5 YR claymatrix contains at least two stone-lines ofcoarse fragments. The deposit also containsboulders of m and more. In the center, the


14/80


lowest area of the cave, the deposit containsa member of fine dust in the upper part,mixed with thick ashlayers. Here, an area ofsome 48 m2 was excavated this year to adepth of f 50 an. The excavations wereconducted with small brushes only. Thesediment was first dry sieved through 5 mmmesh and after flotation, wet-sieved through2 mm mesh.

The general impresssion is that thematerial remains and structures are inprimary context. Human skeletons werefound in anatomical association; stonestructures, hearths, and small concentrationsof lithic, ceramic, and faunal remains werealso unearthed. These concentrations arevery well limited both horizontally andve&cally. So far we s a difference &tweenan a rea that yielded mainly small lithicconcentrations, containing the famoustools of the axe/hoe type and also ceramicsand fauna, and an area with human remainsand few artefacts or animal bones. Datingthe various occupations is still a problem,due to the very fragmented nature of thecharcoal. One date of 3,140 f 80 BP (Beta-51834) (1,230 BC s from a pit with a humanskeleton, stone artefacts, and finelydecorated pottery; another date of 2,120 f110 BP (Beta-51836) (200 BC s associatedwith coarse pottery and stone artefacts.Previously, dates between 8,000 and 2,000BP were obtained (de Maret et al. 1987: 574).

The lithic industry can be tentativelydescribed as a blade industry. The rawmaterial consists mostly of basalt and

quartz, while a smaller fraction consists ofother rocks such as devitrified lava orobsidian. The first stages of the lithicreduction sequence apparently did not takeplace in the cave itself, since corticalelements are rare. In this respect it isinteresting to mention a test excavation in asmaller cave nearby (Shum Laka 111, thatyielded precisely the reverse: many corticalelements and few finished tools. Therelationship between the two caves needs

further investigation.The faunal remains are numerous but

heavily fragmented. All identifiablefragments represent species typical of aforest environment throughout the wholesequence. Not a single savanna element was

found. Inquiries among older people of thevillage indicate that the deforestation mustbe recent. Many of the larger speciesrepresented, which include gorilla (Gorillagorilla) and commom chimpanzee (Pantrogbdytes), have now disappeared from theregion. Despite this local extinction of forestspecies, however, there has never been areplacement of forest animals by wildsavanna species.

In the 5 m2 which were close to theopening of the cave, the excavation wasbrought down to 90 an below the surface.The ash-layer here attains a thickness ofsome 70 rn and consists of many discreteoccupation layers. Here also a stone circlewas uncovered surrounding a bundle ofhuman long bones.

A trial trench outside of the cave showsthat the deposits continue here for morethan 2 m deep. According to thegeomorphological studies, these depositswhich lso contain archaeological material,are apparently in situ as well. From thesedimentation rate inside the cave it can be

inferred that these deposits must be at least20,000 years old. The presence of even oldersediments in pockets in local bedrockdepressions is possible.

n the southwestern part of the site,there are dense concentrations of humanbones. In one instance, a vertical shaftcontained, a t the top, a complete skeleton ofan infant with a stone arrowhead stuckthrough its ilium, probably hit while on itsmother's back. Only two meters away from

the foregoing is another example consistingof two skulls, of which the most completebelongs to a young individual whose headwas resting on one of the large rocks.

In conclusion, the results of the newfield season in Shum aka have been moresuccessful than expected and if fundingpermits, we will carry on for several moreyears. This will give us, for the first time, anidea of the physical an d cultur alcharacteristics of the early inhabitants of thisarea and the nature of the changes occurringat the end of the Stone Age.


15/80

NYAME AKUMA No. 39 June 1993

Acknowledgements

The research was made possible bygrants from the Belgian National Fund forScientific Research, the L. S. B. LeakeyFoundation and the Royal Museum ofCentral Africa. We would like to thank theauthorit ies of Cameroon for researchclearances and the Fon of Bafotchu-Mbu forletting us conduct research at Shum Laka.We would also like to thank the numerouspeople who assisted in the field andprovided logistical support: the workers ofBafotchu-Mbu, Mr. J. J. Ayiboto, Mr.Emtcheu Keumaleu, Mr. Atlee Dzeawoni,the Ambassador of Belgium, F. Michils andhis wife, Mr. and Mrs. B. Geurinckx, Mr M.Rifflet, Mr. H. Doutrelepont and his family,Ms. S. Collette, Mr. E. De Crits, Ms. S.Fagnard, Ms. S. Fenaux, Ms. N. Nyst, andMr. F. Renoirte.

References

Asombang, R1988 Bamenda in prehistory: the evidence from

Fiye Nkwi Mbi Crater and Shum akurockshelters. Ph D dissertation,University of tondon, London

de Maret, P., Clist, B. and van Neer, W.1987 Rhultats des premiares fouilles dans

les abris de Shum Laka et d'Abeke auNord-Ouest du Cameroun,L'Anthropologie, 91: 559 84.

Prospection des Sitesd Habitat dans lesArrondissements de Djoumet Mintom (Sud-Cameroun)

Jean Paul Ossah MvondoE N S Universitb de Yaound6BP 47 YaounddCameroun

Techniques et Mdthodes

Dans le cadre du programme derecherche sur le peuplement ancien du Sud-Cameroun, des recherches furent menbesdans le departement du Ntem etl'arrondis semen t de Zo6telk. Cesprospections ont permis de mettre enevidence des sites d'occupationprehistoriques et mbtallurgiques. Afin detester l'existence des points de peuplementanaen dans les parties les plus reculees etfrontalieres de la fork3 du Sud-Cameroun,

une campagne de prospection fut men duler au 10 aoQt 1992 dans les arrondissementsde Djoum et de Mintom. Cette prospection apermis aussi de completer les recherchescornmenc s depuis quelques annks dans leSud-Cameroun, afin d'obtenir une cartecomplete des zones d'occupation anciennesde cette partie forestiere du Cameroun.

Pour realiser ces prospections troiszones de prospection furent choisies:l'arrondissement de Djoum, l'arrondisse-

ment de Djoum et les rives du Dja. Djoumest une ville cr e vers 1922 parl'administration franqaise Elle est le siege detrois groupes de populations es Bulu, lesFang, et les Pygmees. Cette cite colonialerelie le Cameroun et les populations Bantuau Gabon et au Congo. Les enquetes etrecherches Djoum ont permis de tester cessites anaens de peuplement et d'apprbcier lesens de l'occupation de cet arrondissementpar les populations Fang-Bulu.

La seconde zone de recherche Mintomest 80 m au Sud Est de Djoum. C'est unarrondissement frontalier qui relie laprovince du Sud l'Est Cameroun parLomie-Ngoila d'une part, mais aussi le Nord


16/80

NY ME AKUlUA No. 39 June 1993

du Congo. Cet arrondissement est peupUpar les populations Fang et Pygmbes. Lesrecherches menQs ont permis d e suivre lesens de l'occupation du sud forestier et de

v&ifier les informations recueillies Djoum.Enfin la rive droite du fleuve Dja fut

importante prospecter dans la mesure oiiles points d'occupation anaeme ont depuisla pdhistoire souvent 6tti les rives des coursd'eau, car les hommes dans l'histoire s'ysont install& pour le ravitaillement en eau.

Compte tenu des difficult& deprospection archthlogiques en for& (OssahMvondo 19911, une d6marche approprik fut

adopt&. Une grande place fut accordhe auxenqu&tes rales. insi les entretiens avec lesous pr6fet nous ont permis de connaitrel'occupation spatiale des d if fhn ts groupesde population de l'arrondissement, et der6erer les grandes chefferies de troisiemedegr6. Dans les chefferies, le choix desinterlocuteurs fut port4 sur les chefs eux-mCmes et les personnes les plus I g h duvillage. a technique interrogatoire fut lalibre conversation sur des questions prdcises

portant sur la vie mattirielle pr6coloniale etleurs anaens sites d'habitat, leurs origines etle sens de l'occupation de l e u espaceforestier.

Ces dances de conversation oraledonnerent lieu la descente sur le terrainavec les anciens sur les antiques sitesd'habitat, fin de les localiser, les refirer eteventuellement chercher les temoinsmat4riels de cette occupation anaenne. Desinterrogations sur le terrain permirent de

mieux orienter les textes et surtout unebonne compr6hension du site d'habitat et del'espace forestier. Toutes les informationsrecueillies Mintom-Djoum-Route d'Ovenget la rive du Dja affirment que lespop ula tio ns Fang-Bulu de cesarrondissements ont pour origine communele site d'Ako'oafem.

Les sites re@rCs, il fallait passer larecherche des tkmoins d'une occupationancienne par le test au sondage. Lespremiers indices furent les palmiersanthropiques, les termitieres, et lesmonticules considerees comme desdtipotoirs. Tous les tests rkali&s ont donnedes rbsultats positifs grace l'apparition desvestiges quelques centimetres de

profondeur. Les prospections ont itti aussifaites partir des cartes. Elles permettaientde voir le sens de l'occupation actuelle del'espace sud forestier, l'emplacement des

villages par rapport aux grands axes routierset aux 414ments naturels.

partir de ces elements, on pouvait,avec les autres donnees orales, choisirobjectivement le village d'enquete etl'emplacement possible des sitesd'occupation. Le materiel ldger fut recueillialors que les objets lourds furent la isds insitu sur ce site a p r b des prises de vuesphotographiques.

R sultats des Prospections

Les enquCtes orales menees aupres de lachefferie de Djoum village-Ekom 1 km surla route Sangmelima-Djoum, chefferieoccupee par les Boulou, ont permis derefirer l'anaen site d'habitat d'Ekom. setrouve l'Ouest du village aduel, gauchede l'axe routier. Nous fumes conduits sur leslieux p r Edou'ou Mvele Samuel. Selon lessources orales l'origine, Djoum etaitcompod des villages Engogom et Djom. Cesvillages furent dunis par d'administrationfrancpise pour former la atti de Djoum, vers1922. Akoafem est le premier pointd'occupation ancienne, ensuite Engogom etDjom sont les autres points d'occupationancienne anterieurs la formation de laville.

Le site actuel est envahi p r les arbres,les palmiers, les plants de cacao et quelquesplan tes alimentaires. Le ma tCriel recueilliapres le sondage de surface se composed'une enclume en place, de quelquesmorceaux de ceramique et des scories.L'enclurne atteste la pratique des activitbsm~tallurgiques. Elle a une largeur de 36 cmet une 6paisseur de 25 an La partie droite aune largeur de 15 an et une 4paisseur de 10an a longueur totale est de 68 an Cetteenclume presente des traces d'usureplusieurs points de sa surface, dedimensions differentes. Ces diffkrenteslocalisations des traces d'usure indiquent lesdiffkrents plans de travail et le moded'utilisation de l'enclume. Tous les autreselements qui participaient au travail dumetal n'ont pas eti retrouvbs.


17/80

NYA ME AKUIMA No 39 June 1993

La ckamique s limite un morceau delevre. Elle indique un vase avec un profil enS. La surface est noire, la partie interne duvase Ctait engobbe. La technique de

fabrication est grossike avec des intrusionsdes morceaux de quartz dans la pate. Lesscones de la derniere catkgorie de vestige sepresentent sous forme de petits nodules. Cesd o n n k permettent d'affirmer que le travaildu fer fut une pratique technologique dansla vie des populations de Djoum. Cettetechnologie selon nos informateurs auraconnu sa fin definitive en 1958. Les deuxblbments spatiaux qui apparaissent sur lesite sont l'atelier et le depotoir. Ces vestiges

sont importants pour la suite des travaux. Ilsauront pennis d'attester la presence d'unsite d'habitat Ekom, de programmer lescampagnes de fouille, et de dater la dureed'occupation du peuplement ancien de cettepartie du sud forestier camerounais.

Apres l'arrondissement de Djoum, uneseconde prospection fut effectueel'arrondissement de Mintom, suivant le sensdu deplacement des populations vers leSud-Est, selon les sources orales. Lesenquetes orales nous conduisirent dam levillage de Ze, 1 krn de Mintom ville, sur laroute Mintom-rive du Dja. L'ancien sited'habitat fut indique par Mr Simon AssaeNsan La prospection du site aujourd'huioccupe par les arbres et les plantations, apermis, a p r h quelques tests, de dkouvri rles temoins materiels d'une occupationancienne de Ze.

Quatre types de vestiges composent lesite une fosse, des morceaux de &ramique,une perlek, et un eclat. La fosse est de forrnerectangulaire, d'une profondeur de 1.20 m,1.50 m de longuer sur 1.10 m de largeur. Ladecouverte de cette structure pose leprobleme de l'hygiene dans la vie antiquedes populations du S u d b e r o u n . On peutse demander si la fosse mait servi dedepotoir pendant ou apres sa premiereutilisation. Le sondage effectu6 dansquelques endroits de la fosse n'a permis ded6couvrir aucun vestige casd, ou b rig.

La cQarnique compose la majorit6 desvestiges du site. Dix-huit tessons furentrecoltbs. Quatorze tessons sont desmorceaux du corps du vase et quatreappartiennent l'ouverture. l est difficile de

preciser les parties du corps auquelappartiennent les ceramiques. Au niveaudes caractbristiques externes des tessons, onpeut noter deux colorations: des tessons

noirs et des tessons marron. Les tessonsmarron rep rbe nte nt pres de 80 del'kchantillonnage. Les deux types de couleurindiquerait deux modes de cuissonchm ique Ze: une cuisson oxydante et unecuisson reductrice. De plus, la presence surle m@me ite de deux types de dramiquepose le p ro b l h e de l'origine des diffkrentesceramiques. a faible quantitd des tessonsnoirs implique-t-elle une importation Zede ce type de ceramique? Une etude

approfondie du site permettra d'aborderd'autres problemes tels que les modesd'acquisition, l es provenances, lescirculations et la commercialisation.L'observation des tessons permet deconstater un traitement fin de la surface etde l'interieur des vases avec une engobe. Led6graissant est constitub de petits grains dequartz repartis dans la pgte. Cette techniquefine caract&ise surtout la ckamique marron,alors que les tessons noirs sont

grossierement faconnbs. Les dimensionsmoyennes des tessons noirs sont de 7 m delong et de 5 mm d'kpaisseur. Alors que lestessons marron ont une Cpaisseur moyennede 6 mm et une longueur moyenne de 5 anAucun tesson des deux groupes n'estd6co1-6.

Le troisieme type de vestige est uneperle bleue ciel perforce. En effet plusieurssites du Cameroun ont d6jA l i d ce type deperle: Mdaga dans l'Extr4me-Nord et EligKono dans le Centre. Cette perle 6taitassembl6e avec plusieurs autres l'aided'un fil. C'est ce qui explique la presenced un trou de 2 mm. Cette perle indique laprise en compte du corps comme objet dedecoration et support esthktique chez lespopulations Fang de Ze. Mais plusieursquestions se posent: Quelle etait la partie ducorps dk o r k , s'agissait-il d'une esthetiquefeminine ou masculine. Les perles 6taient-elles import6es ou fabriqu6es localement?

Les prochains travaux vont eclairer cesaspects de la vie esthetique dans l'histoiredes populations de l'arrondissement deMintom. Le quatrierne vestige est un objetlithique. s agit un eclat 3 an ongueur.


18/80


La prospection le long du Dja sur leprincipe de l'occupation ancienne par leshommes des rives des cours d'eau a permisde localiser le site metallurgique d'AlatMakay. l se localise dans le village dumeme nom sur la piste reliant Bi Lomie.Les vestiges se trouvent en surface devantl'eglise EPC du village. En effet les fortespluies de cette region ont provoquk uneQosion du sol de 50 cm de profondeur, cequi, en lessivant la couche d'argile, a mis ensurface de nombreux scories temoins de lametallurgie du fer. e site par la dispositiondes vestiges a une structure spatialecirculaire. La dispersion des scories estcirculaire sur un diametre de 4 m. l secompose de centaines de petits scoriesmesurant entre 6 cm, 3 cm, e t 1 cm delongueur. Deux morphologies de scoriessont attes tees Alat Makay: la formecylindrique, de petites dimensions, et laforme globulaire, de grandes dimensions.Ces types de scories de leur taille serapprochent des scories des sites de Z t6lC.

a zone de concentration maximum desscories est au centre. Un objet en silex futaussi trouv6 a c6t6 d'une lentille de charbon.Cette lentille se presente comme unedispersion de charbon plus de la cendre deforme dong . Elle semble re postitrieureau site dans la mesure oii elle n'appartientpas au niveau de la couche d'argde dkap 6epar YQosion dans laquelle se trouvent lesscories. Ces vestiges attestent la pratique dela m6tallurgie de fer dans l'arrondissement.

Wapres les sources orales, Alat Makay

est une ancienne zone habitation Fang, lespopulations Esawam avaient migre vers lesite d'Akoafem, avec la fondation du postepar l'administration allemande. Unedeuxieme migration de ces populationspartit dfAkoafem vers Alat Makay entraversant cette fois le Dja avec la chutte duposte allemand d'Akoafem pen dan t lapremiere guerre mondiale. A fut dans lepas un centre de transit. 11reliait les deuxrives du Dja. l Ctait aussi relie aux pistes quiallaient vers Bengbis, Messamena, Lomi-lfEst, au Sud-Est, il reliait Mintom-Lele-Souanke. Pendant la periode allemande ilservait de transit entre Kribi-Ebolowa etLomiC. Sur cet axe Ctaient acheminbs, lemais, le sel, le caoutchouc et les autres

produits d'echange. Ainsi, ce site depuis lapQiode prCcoloniale, etait un centre dedistribution des produits locaux eteuropeens sur un rayon de plus de 100 km

Ce qui pose le probleme de la logistique, desCchanges et du commerce au Sud-Camerounpendant la pe o d e prkoloniale.

La p r o s p e c t i o n d e s d e u xarrondissements a permis d'apres lesinformations orales de dbtecter les autressites: ainsi sur la route Djoum-Oveng,Minko'o est un site de m6tallurgie de fer, dememe que Akontang et enfin Akoafem.Tous ces sites n'ont pas kt6 visitits pendantcette campagne. Ces prospections ontpennis d'attester que la forst r e d e des sitesarcheologiques thmoins des occupationsanciennes, malgre les difficult deprospection propres la for . L'hypothesed o n aquelle cette bande frontaliere auraitles zones d'occupation les plus anciemes dela province du Sud merite d'etre nuancee.En effet toutes les enqustes orales menkesDjoum-Minko'o, Ekom, Mintom, Ze,Zoulabot et Alat Makay affirment que cettezone est une occupation rkcente. Lespopulations Fang-Bulu qui occupentactuellement ces departements jusqu'auxfrontieres du Congo-Gabon, ont pourberceau Akoafem d'oh ils sont venus apresla dislocation de la ville. Cette dispersion apermis aux diffkents groupes d'occuper lesnouveaw espaces.

Mais deux problemes se posent: Peut-onconsidher cette zone comme inhabitke avantla dispersion? La datation des sites et ladecouverte de nouveaux objets permettrontde repondre cette question. a civilisationFang-Bulu n'a-t-elle pas occupee toute laregion depuis Akoafem jusqu'auxfrontikes?

Dans tous les cas, l'anteriorite d'unecivilisation Fang-Bulu par rapportl'occupation allemande est 6nonc . Comptetenu de l'antkriorit6 d'Akoafem par rapportaux autres sites prospect6s, il est irnpQatifd'organiser une campagne de prospection

dans ce site. Elle permettra avec les fouilleset sondages de verifier les donnees dessources orales sur l'anciennetb du site, dedeterminer avec precision les deux phasesd'occupation prkoloniale et coloniale dusite. Enfin, la fouille permettra d'apprkier


19/80

NYAME AKUMA No. 39 June 7993

les premiers blbments de la culturematbrielle Fang-Bulu qu'on pourracomparer avec les sites des arrondissements.Les premiers sites prospect& mdritent d'etredat& car ils pennettront d'btablir le sens del'occupation de cette partie sud de la for& etde suivre les mouvements migratoires despopulations Fang-Bulu, de l'antiquitk auxpbriodes contemporaines, avec leurpermanence et rupture.

La deuxieme hypothese selon lessources orales est que les populations Fangont migrb d'Alat Makay vers Akoafem etd'Akoafem vers Alat Makay. Ce quiimplique deux centres de dispersion despopulations Fang-Bulu: Alat Makay etAkoafem. La con&quence est une originebipolaire des populations. a datation dusite d'Alat Makay et les r&ultats d'Akoafempennettront dr6tablir scientifiquement lescircuits migratoires et le sens de l'occupationpar les populations Fang-Bulu de cette zoneforestike.

a troisieme hypothese est que les sitesd'Alat-Djoum-Akoafem, Ze, Minko'o ont

connu antbrieurement l'occupation Fang-Bulu des vagues anciennes de populationKaka-Djem qui auraient progressivementmigr6 la fin du XIXe siecle vers Kribi,Lolodorf, et le Congo. Les recherchesarch6ologiques dans le Sud du Camerounouvrent de toute Cvidence de nouvellesperspectives sur l'histoire du peuplement duCameroun dans l'antiquit6.

Remerciements

Cette campagne de prospection difficilea kt6 rbalis& gr3ce la partiapation activede Alain Boula Meva'a des enquetes oralesjusqu'i la prospection. e ravail de terrainfut facilit6 par le Sous-Prkfet Mr Pipa deDjoum, et les anciens Assae Nsan, Ze, etEdou'ou Mvele que nous remeraons tous ici.

Ossah Mvondo J P.99 Problemes et methodes de la recherche

archeologique en pays forestiers: laprovince du Sud et ses environs. NyameAkurnn 36: 47 52.

EGYPT

ArchaeologicalInvestigations in the BadariRegion Egypt: A Report onthe 1992 Season

Diane L HolmesInstitute of ArchaeologyUniversity College London3 7-34 Gordon SquareLondon WClH OPY

U K

This contribution outlines theresults of the second season of theTrehistory of Badari Project which ispr incipal ly concerned withPredynastic period b4000-3100 calBC developments in the Badari.region near Assiut in Middle Egypt. Ashort first season was conducted in1989. A preliminary report on thisappeared in Nyame kuma (Holmesand Friedman 19891, and the finalreport on this season has now beenaccepted for publication (Holrnes andFriedman forthcoming). The objectivesof the 1992 season were to extend thesurvey begun in 1989 and to surfacesample and test excavate furtherPredynastic sites.

The Badari region lies on the eastbank of the Nile, and may be definedas the area between (and including)the two large wadis, Wadi el Asyutiand Qau Bay (see Fig. 1). This is astretch of low desert approximately 60

m ong, and includes Brunton's threesectors, Badari, Mostagedda, andMatmar (Brunton 1937,1948; Bruntonand Caton-Thompson 1928).A major concern of the present project

is the modern encroachment of thelow desert and consequent damage tothe archaeological record. Modernland-use activities were noted ashaving a significant impact on the


20/80

NYAME AKUMA No 39 une 1993

c u l l i v t i o n

e s c r p m e n t w d i sa r o i n i n g o f i p l t e o u

m o d e r n t o w no r v i l l g e

Fig 1 Map o f t h e B a d a ri r e g io n d e r iv e d f r o m a SPOT s a t e l l it e s c en e t a k e n6 D e c e m b e r 1987


21/80

NYAME AKUMA No 39 June 1993

survival of archaeological sites in theareas surveyed in 1989 (Holmes 1992).Since then modem use of the lowdesert has intensified, the principal

activities being the creation of newfields, house-building and theexpansion of existing cemeteries. Inorder to get a quantitative measure ofthese developments, the writer iscurrently analyzing two satellitescenes of the region.

In July-August 1992 some 40 kmof low desert were field-walkedbetween Gebel el Haridi (the southernlimit of Qau Bay) and the modernvillage of El Matmar. In addition, twolate Predynastic (Gerzean) localities(BD-36 and BD-51) weresystematically surface sampled, andan early Predynastic (Badarian) site

3400) was test excavated.

rchaeological Survey

As in the first season, the aims ofthe field survey were to record sites of

all periods, historic as well asprehistoric, to document theoccurrence of any modem activity onthe sites, and to identify Predynasticlocalities worthy of more detailedinvestigation. Sites visited in 1989were checked again. This allowed anupdate on their condition with regardto modem encroachment. Further-more, the locations of most of the sitesvisited this season were recorded

using a GPS (Global PositioningSystem) satellite receiver. This canprovide a position accuracy of around30 m.

During the second season, more than140 archaeological sites were visited,at least 24 of which have not beenrecorded before (a further 4 newsites were found in 1989). Most of thelocalities in the region were originallyrecorded by Brunton. While the

majority of his site s could berelocated, around a dozen could not,and this is because despite beingactively sought, they either havedefinitely been destroyed, or are

probably destroyed (except in the caseof one group of rock tombs whichcould not be found). There are also afew small areas which were not

checked where sites had been notedby Brunton as well as by Gabra (1930).However, from the level of modemactivity seen from a distance thosesites are also presumed to have beendestroyed, but i t was not consideredtime-effective (time being somewhatlimited) to walk over those areas toconfirm this.

The destruc tion of sites bymodern agriculture and otheractivities is certainly proceeding at arapid rate. To give an idea of this rate,the condition of the archaeologicallocalities visited in 1989 can becompared with observations madeduring the second season. Of the 42localities visited during the firstseason, 5 (12%) were found to havebeen completely obliterated with afurther 3 having been partly destroyedby modern land-use. In 1992 the total

of destroyed sites had increased to 14(33%) with another 4 sho wingsubstantial modem damage, includingsite 34 U discussed below.

The newly recorded localities areprimarily Roman, a period whichBrunton tended to dismiss asunimportant. There are, in fact,numerous Roman sites in the Badariregion including cemeteries in the lowdesert, and rock tombs and quarries in

the cliffs of the limestone plateau. The1992 survey also resulted in thediscovery of a t least 7 previouslyunreported Predynastic sites (labelledBD-4, BD-6, BD-17, BD-51, BD-52, BD-57, and BD-59). BD-4 and BD-6 aresituated in the southern part of QauBay. BD-4 shows indications ofBadarian settlement remains, butunfortunately most of the site hasdisappeared through modern sand

quarrying; BD-6 is a large Gerzeancemetery. BD-17 is located betweenCaton-Thompson's site of North SpurHemamieh and the l imestonepromontory just to the north. It occurs


22/80

NY ME KUM No 39, June 1993

on a small spur, or talus slope, andshows evidence of Badarian andGerzean occupations. BD-51 and BD-52 occur at the foot of the limestone

promontory to the south of Nag 'ArabMiteir. BD-51 appears to have been alarge Gerzean occupation site whichwas reused as a cemetery during theGraeco-Roman period. BD-52 consistsof both Protodynastic and Dynasticcemetery components. BD-57 is asmall Predynastic occupation localityopposite the village of El Nazla elMustagidda with Badarian and laterPredynastic pottery. Finally, BD-59 is

a large site to the north of ElKhawaled with evidence of Badarianoccupation and possibly a Badariancemetery. Part of the site was alsoused as a burial ground in Dynastictimes.

urface ampling

Two localities with Gerzeanoccupation remains were chosen for

systematic surface sampling: BD-36and BD-51. These were tested using asurface scrape technique rather than anormal hand collection straight fromthe surface. The surface deposits andartifacts were scraped off using atrowel and passed through a fine-mesh sieve (approximately 3 rnm).This ensured the recovery of allartifacts including very small itemsmany of which would have beenmissed if the collection had been bydirect pick up, as well as any largerartifacts that might otherwise have

een obscured by a thin layer of dust.

BD-36 is one of the series of sitesBrunton recorded between the villageof Sheikh 'Esa and the tomb of SheikhAhmed, approximately 4 k m to thenorth. Unfortunately, it has not beenpossible to determine the Brunton siteto which BD-36 corresponds. The lowdesert is dissected by numerous s-mallwadis or gullies and Brunton referredto the intergully patches of desert as

spurs. Most of the spurs in theSheikh 'Esa-Sheikh Ahmed stretch ofdesert have sites, but they could notbe exactly matched to Brunton's set of

site and spur numbers. Hence eachlocality has been assigned a BDnumber. BD-36 is likely to be eitherBrunton's site 3000/10 or his 3OOO/ll(Brunton and Caton-Thompson 1928:47 .

While BD-36 does have i n s i t udeposits, there are also someindications of older disturbance. Thereare few Dynastic burials which hadbeen plundered and presumablyBrunton also dug in one or two spots.Nevertheless, BD-36 was in reasonablecondition and it was obvious that itsmain period of use had been duringthe Predynastic. Two 5 x 5 m areas (Aand B were marked out for surfacesampling. Square A was positionednear the tip of the spur, while square Bwas placed where the spur slopeddown into the gully on the south side.These units yielded a combined total

of 712 Redynastic sherds, 86.0% ofwhich are of straw tempered Nile silt(the Predynastic ware types of theBadari region are discussed in Holmesand Friedman forthcoming). Sherds ofuntempered Nile sil t ware(representing polished red and black-topped red vessels) form 9.6% of thecollection with the remaindercomprising sherds of a coarse organicfabric and a marl ware. This collection

undoubtedly points toward a Gerzeandate, a conclusion corroborated by adecorated marl ware sherd (with apainted spiral) found a short distanceaway from one of the collection units.

There were 447 lithic (flint) itemsalso recovered from the two units. Themajority of the pieces (80.1 ) consistof small-sized debitage (i.e., flakes,flake fragments, and shatter thatwould pass through a 20 mm sieve).

The remainder is composed of 35flakes, 13 blades, 16 bladelets, 12chunks, a core, a core preparationflake, and I1 tools. The latter includetwo truncations, a burin, and a


23/80


denticulate which are made on fairlyregular blades and bladelets.

BD-51 is one of the newlydiscovered sites mentioned above. Itshows extensive evidence ofPredynastic occupation and was usedin the Graeco-Roman period as a largecemetery. As a consequence there areonly a few undisturbed patches ofPredynastic sediment.

A 5 x 10 m area was marked outfor surface scrape sampling. Although

this was disturbed with evidence ofplundered later graves, it also showeda concentration of Predynastic sherds.Any disturbance is likely to havemerely redistributed the Predynasticartifacts rather than to have alteredtheir composition. Unfortunately,completely undisturbed Redynasticlocalities are now rare in the Badariregion.

The 5 x 10 m sampling unitproduced a total of 1 P27 Predynasticsherds. As at BD-36 the most commonware is the straw-tempered fabric(71.9%). Next in abundance are sherdsof untempered Nile silt (17.4%). Alsopresent are sherds of coarse organicware (9.O%), and marl ware (1.3%). Inaddition, there are five sherds (0.5% ofthe collection) of a fine Nile silt fabricwhich is characteristic of Badarianceramics, and two of the pieces showtypical Badarian-type rippling. Whilethe bulk of the collection clearlyindicates a Gerzean age, these fivesherds suggest that the site has aBadarian component as well.

The lithic collection numbers 574items mainly composed of small-sizeddebitage (69.5%). The rest of thecollection includes 93 flakes, 10blades, 17 bladelets, 5 cores, and 25tools. Among the tools are twoendscrapers, five notches, a burin, atruncation, two broken blade toolswith alternate backing retouch, andtwo unusually small bifacial crescentdrills.

Test xcavations a t Site 34

Site 3400 lies about 0.5 k m northof the modem village of Deir Tasa,and is one of Brunton's Tas ian andBadarian villages (Brunton 1937: 3,13-14). When visited in 1989, itappeared to be largely undisturbedand indeed was noted as the bestpreserved Badarian locality visitedthat season. However, in 1992, morethan half of the site was found to havebeen cut down (by more than a metre)and converted into irrigated fieldplots. What was left of the site stillappeared undisturbed and a 2 4 mtest pit (TP1) was excavated. Thisrevealed an in s tu trash pit containinglayers of ashy brown sediment andabundant charcoal.

These sediments yielded a pureBadarian artifact assemblage. A totalof 354 potsherds and 2112 lithicartifacts were obtained. Thepredominant ceramic category is anorganic tempered ware (90.4%) with

black cores and brown surfacesvariously wet-smoothed, irregularlyburnished or evenly burnished allover. The other main Badarian waretype is the untempered fine Nile siltfabric (8.2%). he surfaces of this wareare generally slipped and burnished(usually a lustrous even polish) andmany show the characteristic Badarianrippled decoration as well. Theexcavation also produced two sherdsof what Brunton would have classedas Tasian beaker fragments (seeAg. 2 .

Besides a large element (75.0%) ofsmall-sized debitage, the lithicassemblage contains 330 flakes, 64blanks of blade o r bladeletproportions, 10 cores, and 29 tools aswell as a few other items. The basicblank technology is geared for theproduction of flakes together with a

few elongate specimens havinglength/width ratios of 2 or greatergiving them a blade or bladelet-likeappearance. However, there is noindication of a separate blade orbladelet reduction process. The tools


24/80


from locality 3400 are generally ratherad hoc specimens, including end-scrapers, denticulates, notches,perforators, and irregular truncations.There is also a small bifaaal triangle.

Fig. 2.An incised beaker shardfr om Locality 3400.

The botanical and faunal samplesrecovered have still to be analyzed.However, the animal remains do

include several fish vertebrae. Twocharcoal samples and a potsherd havebeen submitted for accelerator dating.At the moment the Badarian is poorlydated:Existing dates suggest that itextends b ck to more than 4000 cal BC,but its chronological limits areunknown.

Conclusions

Although only a small tcstexcavation has been carried out at sitc3400, the results obtained make asubstantial contribution to ourknowledge of Badarian settlcmentartifactual remains. Thcre have bccnclaims for a Badarian presence beyondthe Badari region, but thcsc havegenerally bccn made on thc basis ofinadequate information (discussed inHolmcs 1988; 1989: 180-83). Indccd,

until the Badarian tradition ispropcrly understood in its homcrcgion, it will not bc possiblc todcfinitively state whcther it is prcscntin other arcas. Thc early Prcdynasticof the Armant rcgion (Mycrs and

Fairman 1931; Mond and Myers 1937:1-3,6-8,61,166-76, pls. LIV VI , forexample, is intriguing and its apparentBadarian character needs explaining

(cf. Holmes and Friedman forth-coming). However, given that there isregional variation during theAmratian period (ca. 3900-3500 calBC), it seems reasonable to supposethat there were regional culturalvariants before then. The problem 1s adearth of Predynastic sites in UpperEgypt, besides those of the Badariregion, that are older than around4000 cal BC.

The ceramic assemblage from site3400 with its two beaker sherds alsosupports the view that what itcmsBrunton considered Tasian do, infact, form part of the normal Badarianrcpcrtoire. The lithic data a f f i hecharacterization of the Badarianindustry described on the basis ofBrunton and Caton-Thompson'scollections in the Petrie Museum atUniversity College London (Holmes1988,1989), i.e., that it is a generalizedflake-blade industry with simple flaketools, tools made on elongate flakes ofblade proportions, and occasionalbifacial tools. However, the rare toolsmade on regular blades seen in theolder excavators' collections that wcreapparently from Badarian contcxts arcin fact probably not Badarian, butrathcr bclong to the later Mostagcddaindustry.

Brunton thought that the post-Badarian settlements werc primarilyAmratian in date (Brunton and Caton-.Thompson 1928: 48, Brunton 1937:7582). This conclusion was considcrcdsuspcct whcn the lithic artifacts fromC a t o n - T ho m p s o n ' s a n d h i sexcavations wcre analyzcd (Holmcs1989: 174, 390). Thc recent ficldobscrvations suggcst that the sitcs,

which yicldcd thc lithic artifactsassigncd to thc Mostagcdda industry,arc at least prcdominantly Ccrzcan asis thc casc with thc surfacc samplcdlocalitics, BD-36 and BD-51. Thcrc isan cxtrcmc paucity of matcrial in thc


25/80

NYAME AKUMA No 39 June 1993

Badari region that can be asaibed tothe Amratian. Given the apparentcontinuity of occupation at thestratified site of Hemamieh (Holrnesin press, Holmes and Friedmanforthcoming), it does not seem likelythat the region was abandoned at thattime, but rather that the Badariantradition continued, though perhapsin some evolved or transitionalform.

However, the most pressingproblem is the rapid destruction ofsites as the low desert is reclaimed foragriculture and building space. It ishoped that further fieldwork,concentrating on excavation, can becanied out next year.

cknowledgements

I should like to thank Ted Brock,Director of the Canadian Institute inEgypt, and his assistant, Saad BadiyaMohammed, for their help with theproject's preparations in Cairo. I amalso very grateful to Roxie Walker ofthe Bioanthropology Foundation forthe loan of the GPS receiver. Mythanks too go to my field assistant

hris Ellis of Southampton Universityand our field inspector 'EsamMeghazy Abdel Rahman. The 1992field season was made possible bygrants from the British Academy andthe Central Research Fund of LondonUniversity.

References

Brunton, G

1937 Mostagedda and the Tasian CultureLondon: Quaritch

1948 Mntmar London: Quaritch.

Brunton, G., and Gton-Thompson, G.

1928

Gabra, S.

1930

The Badarian C ivi l i sa t i on London:

British School of Archaeology in Egypt.

Fouilles du Service des AntiquitesDeir Tassa. Anna les d u Se rv ice des

Holmes, D. L

The Predynastic lithic industries ofBadari, Middle Egypt new perspectivesand inter-regional relations. WorldArchaeology 20: 70 86.

The Predynastic Lithic Industries ofUpper Egypt: A Comparative Study ofthe Lithic Traditions of Badari, Nagadaand Hierakonpolis. Oxford: BARInternational Series 469 (2 vols.).

Archaeological cultural resources andmodern land-use activities: someobservations made during a recentsurvey in the Badari Region, Egypt.

J o u d f the American Research Center inE g y p f29: 67-80.

Recent investigations in the BadariRegion (Middle Egypt). In Krzyzaniak,L, Kmeper, K nd Kobusiewicz, M.(eds.), Znt mgi ona l Contacts in the LaterPdtistory of Northeastern Afnca Poznan:Poznan Archaeological Museum.

Holmes, D. L, and Friedman, R F.1989 The Badari Region revisited. N y a m e

kum 31: 15-19.

npm s Survey and Test Excavations in theBadari region, Middle Egypt.Proceedings of the Pnhis tmic Society

Mond, R, and Myers, 0. H.

1937 emeteries of A m n t I London: EgyptExploration Society.

Myers, 0 H., and Fairman, H. W

1931 Excavations at Armant, 1929-31. Journalof EgyptiPn Archaeology 17: 223-32


26/80


27/80

No 39 June 7993

Fig 1 A general area of the Remboud River project; Okala site

27


28/80


Fig 2 Location of archaeological sites along th e Remboud River

28


29/80

NYAME KUM No 39 une 1993

archaeological surveys and excavations inan as yet unkown area of the province andto provide interesting data along a riversystem connecting the Estuaire trade systemto the Og o od trade system known to havebeen active during the seventeenth tonineteenth centuries. This represents animportant contribution to the results of workcarried out along the Estuaire since 1985.

Palaeoenvironmental results

It was decided from the start to usewood charcoal and carbon 13 content of soilsto trace environmental changes in the

RembouC River area. Soils samples weretaken on the Rembou4 at locations 2,8,9,15,19, and at OMa. Charcoal samples were setaside for wood identification by R.Dechamps, Tervuren Musem, Belgium, butthis study i s not yet completed. All the soilcolumns show the same pattern: forest coverthroughout the clay cover without any traceof savanna vegetation. All 1 X eadings f l lb e t ~ e e n - - 2 4 . 9 ~ / ~ ~nd -30.0/00. One site isimportant for dating the palaeoenviron-

mental evidence: location 9, where aradiocarbon date of 18,020 f 80 BP (Beta-53553) has been obtained from a depth of-110 n n the clay, sampled from a sub-horizontal and continuous charcoal layer.This date shows that the lower half of thedeep clay cover is of Pleistocene age, at leastat some locations. The date also agrees wellwith the c. 40,000 BP date from Okala (seeabove) at the base of the clay.

rchaeological resultsThe RembouC Project resulted in the

discovery of 25 sites of various periods with29 archaeological components, i.e., 4 aremulti-component sites. Two sites can beidentified with the Middle Stone Age. One isa quarry at the Kango 1 platform, locatednorth-west of Kango. Previous surveys hadnot identified the site. It was only during theplatform's survey that the quartzite andquartz tools were found: one proto-handaxe,

one very nice discoid nucleus, one choppingtool, one pick, one simple nucleus and twoflakes. The second site is location 1 on aforesters' dirt road going south from theRembouC. It is contained within a layer of

quartz gravel at the base of the clay mantle:of the 24 artefacts found, 95.5 are onquartz and 4.5 on flint. Flakes and simplenuclei were recovered. There is no clusteringof sites in the countryside. The Middle StoneAge locations are randomly distributed andrare.

Nine Late Stone Age sites have beenlocated but none were dated. The sites occurat varying depths in the clay cover on hilltops. Tools are rare except for location 25where quite a few flakes are retouched(c. 8 ). Raw material is mainly quartz (from60 to I, with flint (from 6 to 35 ) andquartzite for the rest. LSA sites can be foundnear rivers, like the RembouC r quite fari n l a n d 4 kilometres from the main riversystem.

No Neolithic sites were found, but 12Iron Age sites were located. Pottery is foundon all the sites and is associated withgrinding stones, iron slag, charcoal, andlaterite blocks.

Refuse pits were dug through the clay atdifferent sites. At location 9, the pits were

dug on the south slope of the hill and arec. 0.50 to 0.80 m in diameter with a depth ofc. 0.60 m. They are thus of small volume.There s one exception: a pit dug on a northslope leading down to the RembouC river a tlocation 25 which cut through the clay andwas also dug into the laterite and the upperpart of the weathered rock. Though itsopening could not be exactly pinpointed, itis thought to have been at 0.40 m belowsurface while the pit extended to a depth of2.50 m with a mean diameter of 0.70 m. It isthe first time such a large pit has been notedin the province.

Several Iron Age sites were radiocarbondated. They suggest the area was settled byIron Age inhabitants around 2,180 BP(location 12; Beta-53555: 2180 f l BPI andwas more or less constantly occupied fromthen on: location 3 at 1,750 BP (Beta-53552:1,750 i 70 BPI, location 9 at 1,650 BP (Beta-53,554,1,650 0 BPI, location I1 at 1,360 BP

(Beta-54220: 1360f70 BP), and location 15 at670 BP (Beta-54221: 670 50 BPI. Thereseems to be a concentration of sites along theRemboue within a 4 m comdor as a surveyof foresters' roads to the south did not yieldany Iron Age sites. Test pits at location 15


30/80

No. 39 June 1993

dated to c. 670 BP showed the villagecovered a minimum of 900 square metres.The density of artefacts is not high on IronAge sites. This is clearly different fromcontemporary villages on the Gabonestuary. It can thus be proposed the villagesof the Remboue were smaller.

Three such historical sites were found.They are conspicuous for their importedEuropean artefacts, such as glazed pottery,white clay tobacco pipes, gun flints, guns,iron implements or objects, glass beads, andglass bottles. Two sites were dated: location22 yielded a date of 10 60 BP (Beta-53556)and location 15 s date was modem (Beta-54222). Locations 15 and 22 also hadGabonese pottery, showing that pottery-making may still have been practised in thefirst half of the nineteenth century, as thesites were dated by the tobacco pipes andbeads to A.D. 1760-1840.

It must be stressed that no site withimported European objects was foundinland. This is in keeping with the first goodEuropean descript ions of trade in theprovince showing European goods movingup and down the main rivers; the Rembou6was the main economical route between theEstuary and the Ogooue river at the time(Bowdich 1819). It is possible our location 15is the old village of Akonjdo known fromEnglish and French texts, since the earlytwentieth cen tury village of Akondjo,abandoned sometime around 1930, lies just800 m upstream.

By comparison with Iron Age sites

historical villages differ markedly in densityof artefacts (European and Gabonese), whichare much more abundant at the latter. Thismay point to larger villages developing atstaging posts of European trade. This isfurther evidenced by test pits revealing ac. 8,100 m2 extension of the archaeologicallayer at location 15.

Part of a trade gun was found at location15. This is the first such weapon ever to befound during excavation in Central Africa

see Dubrunfaut 1984).

iscussion

The main vegetation type in the studyarea since c. 40,000 BP appears to have beena forest cover. Absolutely no trace ofextensive savannas was found. Furthermore,the lower half of the clay cover-onlocations where the sediments have not beenheavily eroded away-apparently datesback to 40,000/18,000 BP, while the upperhalf is substantially younger; this agreeswell with what is known of Late Stone Agesites further north (Estuary) dated to after8,000 BP and lying in the upper part of theclay. This apparently contradicts thehypothesis that the forest was limited torefuge areas during the dry climaticepisodes of the late Pleistocene of CentralAfrica, with savannas taking over the rest ofthe land. It has been shown recently thatCentral Africa may have witnessed a switchfrom rain forest to montane forest duringthe dry climatic episodes (Bengo and Maley1991, Elenga 1992). This would be in accordwith our data, as 13C readings can notdistinguish between two types of primaryforests.

J Maley (1987)-followed by B Peyrot(1989bhas proposed three forest refugeareas: the Monts de Cristal, the Mayombemountains, and the Massif Du Chaillu.Previously my work near the Feman Vazhad shown the Mayombe refuge limits toextend further to the west by c. 80 kmm n mum (Clist 1991,1992: 6). The 1992 datafrom north and south of the Gabon Estuaryagain has shown since c. 40,000 P a forestvegetation where large expanses of savannasshould have been. Three models can besuggested to explain this:

Montane forest took over from rainforest in the Estuaire Province.The forest refuges have beenmisplaced; this has been suggestedby J Reitsma, A. Louis, and J JFloret (Reitsma, Louis, and Floret,1992). They think the Monts deCristal refuge area was on the

western slopes of the mountainextending towards the Atlantic (i.e.,our area).

The Gabon Estuary and RemboueRiver areas were included in a


31/80

N Y M E K U M No. 39 une 1993

refuge area of a river gallery type,separate from the Monts de Cristalrefuge.

Our paleoenvironmental data, brieflyoutlined here, will be developed in aforthcoming paper (in preparation).

By finding connections betweenarchaeological sites and their environmentaland palaeoenvironmental features, we havebeen able to map the probability ofdisturbing sites during future oilexploration/ production work in unexploredareas of the Estuaire Province. This mapshould lead to better ArchaeologicalHeritage Management by British Gas Gabon.Moreover, the approach to reducing impacton cultural resources that we havedeveloped in Gabon is also being used byBritish Gas in operations in other countries,notably Tunisia.

The main archaeological results of our1992 field work are the following:

It can now be suggested on more securegrounds that Middle Stone Age sites foundon and in stone-lines at the bottom of clays

in the Estuaire Province date to before40,000 BP At Okala, flint and quartzartefacts were found under the c. 40,000 BPcharcoal layer. They are markedly differentrom stone-line industries found in Gabon

(Lock0 1991 .

It is possible to find 40,000 to 12,000 BPsites in situ in the clay; they are rare butpresent.

While Neolithic sites (4,000-2,100 BPIwere frequent on the estuary to the north nosuch sites were found on the Remboub.Though not conclusive, this is stronglysuggestive of absence of human settlementsin this area at that time.

Location 12 dated to c. 2,180 BP withiron slag pushes back some 200 years theiron smelting technology on the Atlanticcoast of Gabon. Up to now the earliest datewas of 1,900 BP from Kango (Clist 1990,JCzbgou et Clist 1991).

It has also been shown that the densityof Iron Age sites is high, as in other areas ofGabon that have been well surveyed. TheIron Age sites around the RembouC do notconcentrate on the river and they are notvery large.

Perhaps during the Late Iron Age, butsurely during historical times, villages fallback on the RembouC river and are larger inextent at least partly because of European

trade. The archaeology of eighteenth andnineteenth European trading posts in Gabonappears to hold great promise.

The most interesting labwork has beenthe develo pment of a GeographicalInformation System linked to our 14C datesand archaeological site data bases. A map ofarchaeological risk of the Libreville area was

finished in November using SPOT satelliteimages and the Multiscope/Windowssoftware. Next year (1993) work willcontinue with the Institut National deCartographie using their MapInfo andDbmbter softwares. Our objective is todevelop a CIS cum archaeological atlaswhich may be put to use in countries ofCentral Africa for Archaeological HeritageManagement purposes.

knowledgementsI am grateful to Dr. Roger Dechamps of

the Section d Anatomie des Bois Tropicauxof the Tervuren Museum in Belgium for thewood identification from Okala and to theBritish Gas Gabon company for their activeinvolvement in Archaeological HeritageManagement and good working conditions.

References

Bengo, M. D. and Maley J.1991 Analyses des flux pollinitiques sur la

rnarge sud du Golfe de Guin6e depuis135.000 ans. Comptes-rendus de2 Academie des ciences e Pans 313,843-49.

Bowdich, T. E.

1819

Clist, B.

1990

Mission from Cape Coast Castle toAshanfee. London: W. Ward.

Des derniers chasseurs aux premiersm6tallurgistes: skdentarisation etdebuts de la metallurgie du fer(Cameroun, Gabon, Guinke-


32/80

N YA M E A K U M A No 39, une 1993

Equatoriale). In Lanfranchi, R. andSchwartz, D. (eds.), P a y s a g e sQua terna ires de I'Afrique CentraleAtlant ique. Paris: Didactiques,

ORSTOM, pp. 458-78.

1991 Archaeology and palaeoenvironment.In Maguelou: An Enmronmental Study ofthe Ofoubou A m or Dupont and P n08BV. Libreville: Africa Forest, pp. 59-64.

Clist, B.

1992 1991 excavations and laboratory work

in Gabon Nyame Akuma37: 4-8.

Dubrunfaut, P.

1984 Armes rt F e u de Traite en A @ w Noire:Aspects Technique, Artist ique e tEthno log ique-Propos i t ion d 'uneTypologie.M A. thesis, Universite Librede Bmxelles.

Elenga H.

1992 VCg ation et Climat du Congo depuis24 000 Ans BP: And ys e Palynologique deSkquenccsSedim entaim du Pays Batekeetdu Littoral . These de Doctorat enSciences-SpCcialitB GBologie,UniversitC d Aix-Marseille.

JkBgou, M. P and Qit, B

1991 L,'Age du Fer Ancien: Gabon. InLanfranchi, R, and Clist, B. (eds.), Auxorigines de 1'Afnque Centmle. Paris:

Centres Culturels franqais d'AfriqueCentrale/CICIBA, pp. 203-7.

Locko, M.1991 Ages de la Pierre Ancien et Moyen:

Gabon In Lanfranchi, R, and Clist, B(eds), Aux origines e I'Afrklue Centrale.Paris: Centres Culturels franqaisd'Afrique Centrale/CICIBA, pp. 67-70.

Maley, J.

1987 Fragmentation de la foret densehumide et extension des biotopesmontagnards au Quaternaire r6cent:nouvelles donnees polliniques etchronologiques . Impl ica t ions

palbclimatiques et biogkgraphiques.Palueoecologyof Africa 18: 207-334.

Peyrot, B.

1989 Pal6oCcologie du Gabon 15.000 ansBP: situation th6orique et implicationssur la pal606cologie de la fin duPlBisto ne. Nsi 6: 115-25.

Powers, A.

1990 Phytolith Analysis of Five SedimentSamples from the Estuaire Province ofGabon, West Africa. Unpublishedreport.

Reitsma, J. M. Louis, A., and Floret, J.-J.1992 Flore et vegetation des inselbergs et

dalles rocheuses: premiere etude auGabon, Bulletin du Museum Nationald'Histoire Naturelle 14 rie B: 73-97.

MADAGASCAR

The Central Androy ProjectSouthern Madagasc

nyame akuma issue 039

Documents