Kesem-Kebena: A Newly Discovered Paleoanthropological Research Area in Ethiopia Author(s): Giday WoldeGabriel, Tim White, Gen Suwa, Sileshi Semaw, Yonas Beyene, Berhane Asfaw, Robert Walter Source: Journal of Field Archaeology, Vol. 19, No. 4 (Winter, 1992), pp. 471-493 Published by: Boston University Stable URL: http://www.jstor.org/stable/530428 Accessed: 19/08/2009 17:21
Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in the JSTOR archive only for your personal, non-commercial use.
Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at http://www.jstor.org/action/showPublisher?publisherCode=boston.
Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed page of such transmission.
JSTOR is a not-for-profit organization founded in 1995 to build trusted digital archives for scholarship. We work with the scholarly community to preserve their work and the materials they rely upon, and to build a common research platform that promotes the discovery and use of these resources. For more information about JSTOR, please contact [email protected].
Boston University is collaborating with JSTOR to digitize, preserve and extend access to Journal of Field Archaeology.
http://www.jstor.org 471
Kesem-Kebena: A Nervly Discovered PaleoanthropologxcalResearch Area in Etlliopia
GidayWoldeGabriel Los AlamosNational Laboratory, Los Alamos,New Mexico
TimWhite The Universityof Californiaat Berkeley,Berkeley, California
GenSuwa The Universityof Tokyo, Tokyo, Japan
SileshiSemaw RutgersUniversity, New Brunswick,New Jersey
YonasBeyene Ministryof Cultureand SportsAffairs, Addis Ababa,Ethiopia
BerhaneAsfaw Ministryof Cultureand SportsAffairs, Addis Ababa,Ethiopia
RobertWalter Instituteof HumanOrigins, Berkeley Geochronology Center, Berkeley, California
ThePaleoanthropoloyical Inventory of Ethiopia is a long-termproject designed to assessthe paleoanthropoloyicalresources of theEthspian rift system. Inventory work completed in 1989 hasestablished several new research areas. One of these, the Kesem-Kebena area near thenorthern terminus of theMain Ethiopian Rift, hasnow been investigated by the inven- toryteam, and several vertebrate paleontoloyical and archaeoloyicallocalities have been dis- covered.The deposits rangefi^om >3.7Ma (millionyearsof age; Pliocene) to Late Pleisto- cene.Important stratigraphic units were dated by conventional KlAr dating on basaltic lavasand 40Ar/39Aron feldspar separatesfi^oun keytephra horizons. Among the most signif - cantdiscoveriesfi^om theenctensive Kesem-Kebena sedimentary succession are Pleistocene Acheulianlithic assemblages and Acheulian fauna datingto ca. 1.0 Ma.
Introduction discussthe potentialof these new areasfor elucidating The geologicalhistory of easternAfrica has provided a humantechnological and biological evolution. uniquesetting for the paleontologicaland archaeological investigationof humanorigins and evolution. Research at Historical Background Laetoliand OlduvaiGorge (Tanzania),East and West Turkana(Kenya), Omo, Hadar,Melka Kontoure, the Initial EthtopianDiscoveries:1900-1960 MiddleAwash (Ethiopia), and other areas(FIG. 1) has Thefirst vertebrate fossils reported from the Ethiopian providedunique insights into hominid evolution and gen- rift systemwere found during an expeditionled by the eratedmany new questions. These questions, in turn,have Frenchexplorer de Bozaswho visitedthe areanear the stimulatedthe searchfor additionalresearch areas from mouthof the Omo Riverin southernEthiopia in 1902 whichmore data might be recovered. (FIG. 1). Thefirst geological reconnaissance and substantial Wereport here on a seriesof geological,archaeological, paleontologicalwork in thissoutheasternmost part of the andpaleontological discoveries made by the Paleoanthro- Ethiopianrift systemwas carriedout by CamilleAram- pologicalInventory of Ethiopiain 1988-1989. We first bourg and colleaguesin the early 1930s (Howell and outlinethe historyof paleoanthropologicalresearch in Coppens1983). No hominidremains were recovered by Ethiopia,introducing the inventory project, and then pro- that effort,leaving the initialhonors to anotherFrench videthe regionaland local geological framework for our projectin the samedecade at PorcEpic cave (1929 and discoveries.Finally, we describethe new discoveriesand 1933), nearDire Dawa(FIG. 1), farto the NE (Teilhard 472 Kesem-KebenaPaleoanthropolo,gical EthiopialWoldeGabriel et al.
and continueduntil 1982 (Chavaillonet al. 1979, and referencestherein). Meanwhile, in the late 1960s a joint French-American-Kenyanexpedition returned to the LowerOmo Basin.The workof this InternationalOmo Expeditionresulted in the recoveryof paleontologicaland archaeologicalresources from well-understood geological contexts(Howell 1978; Howell and Coppens 1974; Howellet al. 1987;de Heinzelin1983). The Omo project establishedstandards for multidisciplinary,large-scale pa- leoanthropologicalresearch in Africa. Chavaillon,who was then excavatingat MelkaKon- toure,encouraged a Frenchgraduate student in geology, MauriceTaieb, to proceedwith geologicalsurvey along the AwashRiver in the late 1960s.Taieb recorded a va- rietyof archaeologicaland paleontological localities in the Afardepression (Taieb 1974). He formedthe Interna- tionalAfar ResearchExpedition (I.A.R.E.) and began workat one of the areashe hadfound, Hadar ( Johanson et al. 1982 andreferences therein). One I.A.R.E.partici- pant,Jon Kalb,subsequencly dropped out of the Hadar work and formedthe Rift ValleyResearch Mission in Ethiopia(R.V.R.M.E.). Kalb's team proceeded to explore the MiddleAwash region immediately south of Hadar (Kalbet al. 1982a,1982b, 1982c, and references therein). In the early1970s, Fred Wendorf and associates inves- tigateda smallstudy area with Middle Stone Age localities in the Gademottaarea of the Lakesregion (FIG. 1), which is the centralsector of the MainEthiopian Rift (Wendorf and Schild1974). Anotherteam under the directionof Clarkundertook paleoanthropological work at Gadeb (FIG. 1), on the plateaualong the SE edgeof the Ethiopan Figure 1. A map of Ethiopiaand northernKenya showing some of rift (Clarkand Kurashina1979; Williamset al. 1979). the new paleoanthropologicalareas discovered by the Paleoanthropo- Clarkand White carried out reconnaissancestudies in the logicalInventory of Ethiopia.The Main Ethiopianand the AfarRifts MiddleAwash in 1981 (Clarket al. 1984). are stippledand darkareas represent rift valleylakes. Inset map shows By the autumnof 1982, the Omo projectwas inactive the distributionof the East Africanrift system. andwork at Gadeband Gademotta had ended. Fieldwork at MelkaKontoure persisted. Research at Hadarhad de Chardin1930; Vallois1951; Clarkand Williamson ceasedin 1977 butwas ready to resumein 1982, concur- 1984). Briefvisits to the Omowere made by LouisLeak- rentlywith work in theMiddle Awash. In thefall of 1982, ey'sassociates in 1942 (Howelland Coppens 1983). Ad- however,all foreignscientists in archaeology,paleontol- ditionalwork on a varietyof archaeologicallocalities ogy, and affiliateddisciplines were informedthat their throughoutthe Horn of Africawas accomplishedby Ethiopianpermits were no longervalid and that all field- J. DesmondClark during the SecondWorld War (Clark workwas suspended until new national policies governing 1954). researchhad been formulated and legislated. As a conse- quence,no Ethiopianantiquities-related field research was Progressox SeveralFroxts: 1960-1981 doneby foreignexpeditions between the autumnof 1982 Earlyin the 1960s,the Paleolithicarea of MelkaKon- andSeptember of 1990. tourewas discovered by hydrogeologistG. Dekkeron the Ethiopiainvests the authorityto issuepermits for field edgeofthe Ethiopianrift near Addis Ababa (FIG. 1). Here, researchwith the EthiopianMinistry of Cultureand in the headwatersof the AwashRiver, a groupled by SportsAiairs (EMC).The Centrefor Researchand Con- Frencharchaeologist Jean Chavaillon began work in 1965 servationof CulturalHeritage (CRCCH; Antiquities) is JoutnaZof Field Archaeoto,gylVot. 19,1992 473 the EMCunit responsible for managingEthiopian antiq- hominidorigins, diversification, and anatomicaland be- ultles. havioraladaptations. Fromthe Ethiopianand international perspectives, the TheDevelopmext of Ethiopiax Paleoaxthropology: traditionalapproach of exhaustinga singleresearch area 1982-1988 beforesearching for anotherarea is increasinglyinappro- priate.For manyperiods of interestto paleoanthropolo- Thehistory presented above shows a patternof mostly gists,localities are unknown or limitedin extent.Even for unsystematicreconnaissance for new paleoanthropological better-known periods, proper antiquities management and areas,followed by discoveryand intermittent or extended researchplanning are still impossiblebecause the extent exploitation.Paleoanthropological collecting areas, once of undiscoveredresearch areas remains unknown. The in- identifiedin Ethiopia,have then continued to attractpa- ventoryproject was thereforedesigned to integrateEthi- leontologists,geologists, and archaeologists. The results, opiannational interests with internationalscientific con- in the veryproductive fossil fields such as Hadarand the cerns, allowingpaleoanthropology to proceedin the Omo,have been spectacular increases in knowledgeof the knownresearch areas while additionalresearch areas are prehistoryof this partof the world. simultaneouslysought. Paleoanthropologicalresearch in Ethiopia,before the EthiopianRevolution in 1974,and for several years there- after,was conductedby foreign-basedexpeditions inves- Scape tigatinga few rich deposits,with little or no scientific The inventoryproject began its workin the Ethiopian collaborationof Ethiopianinstitutions or individuals. riftsystem v . . . and . . adjacent . highlandareas. The .. . most . . impor- ^thloplan partlclpatlon was most .y llmltec bto acmmlstra- tant databearing on humanorigins and evolutionhave tion; local scholarshipand facilitieswent undeveloped. been found embeddedand sealedin sedimentaryunits Artifactsand fossils were usually loaned for studyto for-
. . . . associatedwith the riftsystem. These primary and contex- elgn mstltutlons. tualdata (artifacts, fossils, sediments) become susceptible The 1970sand 1980switnessed the emergenceof sev- to discoveryas a resultof geomorphicprocesses aided by eralEthiopian scholars working in paleoanthropologyand faultingassociated with rift evolution. For geomorpholog- allieddisciplines, a developmentin whichthe CRCCH icalreasons, fossiliferous occurrences in easternAfrica, as playeda crucialrole. Furthermore, the Centre oversaw the a verygeneral rule, become rarer as theirage increases completionof laboratoryand storage facilities on the Na- the oldera deposit,the greaterthe chanceof subsequent tionalMuseum compound in AddisAbaba. With the rapid erosionor deepburial consistent with intensevolcanotec- developmentof nationalpersonnel and facilitiesfor pa- tonicevolution of the MainEthiopian Rift andthe Afar leoanthropologicalresearch came the recognitionthat the duringthe Neogeneand the Quaternaryperiods. For ex- historydescribed above had left Ethiopia'spaleoanthro- ample,middle and late Miocene deposits that might hold pologicalpotential largely unexplored. The Ethiopian rift cluesto the originof the Hominidaeare rarer and more systemhad proven capable of yieldingevidence needed to disturbedthan middle or late Pleistocenedeposits. This solvethe basicproblems of humanorigins, but the extent "pullof the recent"bias has serious consequences for the of paleoanthropologicalresources in this regionwas still mventory. project. . poorlyknown. A systematic,comprehensive program of Localitiesdiscovered by the inventoryproject span the discoveryand assessment was required. timeperiod beween the earliestEthiopian vertebrate pa- leontologicallocalities (perhaps Mesozoic; expected to be The PaleoanthropologicalInventory of Ethiopia rare)and the latest Stone Age andIron Age archaeological occurrences(far more common). Potential localities of all Goals agesare recorded as encounteredduring inventory work, The PaleoanthropologicalInventory of Ethiopiawas but projectfocus is on MiddleStone Age and earlier formedin 1988 to allowEthiopia to assessher overall archaeologicaland paleontological resources embedded in antiquitiesresource base. Located within the Ministryof theirgeological contexts. The inventoryproject is con- Culture,the projectis designedto documentthe distri- ductedoutside of alreadyestablished paleoanthropological butionand research potential of lateCenozoic sedimentary researchareas (the Lower Omo, Melka Kontoure, Gadeb, bodiesand to placethese deposits in a time-stratigraphicMiddle Awash, Gademotta, and Hadar;FIG. 1) because frameworkfor the studyof bioticand cultural change. It the EthiopianMinistry of Cultureis alreadycognizant of representsthe primary research step in studiesto elucidate the resourcesand potential of theseareas. 474 Kesem-IRebenaPaleoanthrrpological Ethspia/WoldeGabriel et al.
Termixology geological,paleontological, and archaeological contents of The term"site" has a varietyof archaeologicaland pa- delimitedoutcrops of one or a few stratigraphichorizons. will contain leontologicalmeanings in easternAfrica, a circumstance Largefossil and artifact-bearing areas (fields) traceablefor more thatengenders substantial confilsion. For example, Hadar, manylocalities and may include beds et al. 1984). Olduvai,Laetoli, and the Omo are often called"sites," than30 km (de Heinzelin1983; Clark assessment whilemuch smaller archaeological and paleontological oc- The surfacecontent of the localitiesallows of currenceswithin these "sites" are also routinely identified of the paleoanthropologicalsignificance and potential To sum- as "sites."As Delson (1984) points out, areassuch as eacharea discovered by the inventoryproject. sig- thosenamed above are more properly called "fields" in the marize,we circumscribeareas of paleoanthropological vertebratepaleontology sense. nificanceby the outcropof depositsand we characterize within Differentsystems of nomenclatureand spatial recording the areasby the surfacecontent of the localities have "evolved"independently as differentpaleoanthro- them. pologicalresearch teams dealt with their respectivedis- coveriesin easternAfrica. For example, French and Amer- Methods ican teamsdefined collecting "localities" in the Omo, a The Ethiopianrift and Afarare largegeographic fea- systemalso appliedat Hadarwhere, as Johansonet al. tures(FIG. 1), but only relativelysmall outcrops within (1982: 379) describe,each localityrepresents ". . . a themconstitute targets of researchopportunity. To iden- discretegeographic area of variablesize andusually sam- tifythese targets, we employan approachcombining sat- pling a fairlyrestricted portion of the stratigraphicse- elliteand aerialremote sensing with ground-truthinfor- quence."In boththe Omo andat Hadar,locality bound- mationacquired by vehicleand foot survey.Rather than ariesare circumscribed on aerialphotographs. East of Lake startingsmall and working outward by chance,the project Turkana(FIG. 1), Leakeyand Leakey (1978: 8) describe beginscomprehensively and focuses incrementally (Asfaw threemajor "regions," with a seriesof numbered"areas," et al. 1990). much largerthan the Omo localities.These areasare This telescoping,or "stratified,"approach to areatar- "boundedeither by naturalvegetation cover or by easily getingmakes the inventorygeographically comprehensive recognizabletopographic features (sand rivers, etc.)." At withoutsacrificing spatial focus, meeting both scientific KoobiFora, specimen provenance is individuallyrecorded and EMC adrninistrativeconcerns. The 1988-1989 in- on aerialphotographs, and paleontological "localities," in ventorywork has alreadyestablished five new paleoan- the Omo sense,are not recognized.Archaeologists Isaac thropologicalresearch areas: Melka Werer; Burji; Fejej; and Harris(in Leakeyand Leakey1978) use the terms Bilate;and the subjectof the currentcontribution, the "sites"and "localities" synonymously at KoobiFora, using Kesem-Kebena(FIGS. 1, 2). the SASES(Standard African Site EnumerationSystem) At the top of our samplinghierarchy is LandsatThe- to place these spatially.In his surveywork in Kenya, matic Mapper(TM) and Space ShurdeLarge Format Pickford(1986) dividesthe countryinto six "areas,"rec- Camera(LFC) imagery. This imagery,when used with ognizingsmaller "areas" of paleoanthropologicalsignifi- ground-truthinformation already acquired by fieldwork, cancewithin these majorareas. The smallerareas (for allowsareas with exposedcontinental sediments to be example,the "Koru/Songhor/Muhoroniarea") contain targeted(Asfaw et al. 1990). Foot and vehicletransects "sites,"synonymous with "localities" (for example, Site 15 arethen usedto searchfor localitiesin the targetareas. at Koru).It canbe seenfrom this briefreview that there The localities,in turn,are used to characterizethe nature is no standardizedterminology. and potentialof each area.In largeand rich areas,the The PaleoanthropologicalInventory of Ethiopiaavoids discoveryand assessment of alllocalities is impossible,and the ambiguousterm "site." We recognizemappable de- the inventoryobjective is limitedto ensuringthat a rep- positsas "areas"(roughly equivalent to, but less formal resentativesample of localitiesis assessed. than the 'iformations''or "groups"of the geologist; Surveyroutes and discovernr data are recorded in a daily Bishop1967; Stein 1987). For example,the Omo and log, cross-referencedto imagery hard copy, aerial photo- Hadarwould constitute "areas" in our terminology. graphs,1 :250,000 and 1: 50,000 topographic sheets, GPS Likemost workers,we use the term"localiN' to de- data, localityrecording forms, and on-localityphoto- scribea relativelydiscrete paleontological or archaeologi- graphsand videall archivedin the Paleoanthropology caloccurrence within an areaof paleoanthropologicalsig- Laboratornrat the NationalMuseum of Ethiopia.It is not nificance.Localities are definedon the basis of the the intentof the projectto exploitthe researchpotential Jourwalof FieldArchaeoloflylVol. 19, 1992 475
0 5 10km 1 | .v-,';12 | + +1g3 4 En 5 @ 6 I SCALE ( 1: 114,000)
Figure2. Inset map (A) shows the generallocation of the studyarea. Solid lines with ticks represent normalfaults. Map (B) is a perspectiveview of a simplifiedgeological map of the Kesem-Kebenaarea, westernpart of the northemsector of the Main EthiopianRift. The symbolsrepresent: 1. recent sediments;2. basalticlavas and weldedtuff units coveredby a thin veneerof sediments;3. basaltic conglomeratedominated by sedirnentsintercalated with weldedtuff units exposedalong faultscarps and streamcuts; 4. rhyoliteand basalticlavas; 5. basalticand silicicrocks of the rift escarpment;and 6. vokaniccenters. The generallocation of the Kesem-Kebenaarea is indicatedin Figure 1. of localitiesdiscovered. Rather, the goal is documentation New Mexico.Analytical procedures for XRF andmicro- andgeological sampling; excavation and collection of an- probedata are presented by Valentine ( 1983) and Warren, tiquitiesare not undertakenunless the remainsare in Beyers,and Caporuscio (1984), respectively.Information immediatedanger. Geological samples of volcanicmaterial on K;/Aranalyses is givenin WoldeGabriel,Aronson, and suchas lavasor tephraare routinely taken for radiometric Walter(1990). The 40Ar/39Ar results were obtained at the dating,stratigraphy work, and chemicalfingerprinting. Instituteof HumanOrigins Geochronology Center in The last has becomean increasinglyimportant (Brown Berkeley,California. The center has a filllyautomated laser and Feibel 1986), but potentiallyunreliable (Aronson, fusion40Ar/39Ar datingsystem. Samples and monitors are Walter,and Taieb 1983), tool in correlatingdeposits irradiatedat the 8 MWOmega West reactor facility, Los withinand between areas of EasternAfrica. AlamosNational Laboratory, which has a fast neutron In the Kesem-Kebenawork, some of the stratigraphi- fluxof 5.7 x 1013ncm-2S-l.A sanidinefrom the Fish callyimportant samples were selected for major and minor CanyonTuff with a referenceage of 27.85 Mais usedas analysesby X-ray fluorescence and by electronmicroprobe a fluxmonitor. System blanks contain background levels analyticaltechniques at LosAlamos National Laboratory, of 2.7, 0.05, 0.03, 0.1, and 0.03 x 10-12 cc STP for 476 Kesem-KebenaPaleoanthropolo,gical EthiopialWoldeGabriel et al.
40Ar,39Ar, 38Ar, 37Ar, and 36Ar, respectively. More than five the largestof thesebeing the Kesemand the Kebena.The individualgrains were analyzed for each unknown sample. plantationsof Sabure(Awara Melka) and YaloKebena Theestablishment of a permanentpaleoanthropological have been established on the Awashfloodplain (elevation inventoryproject within the EthiopianMinistry of Cul- 750-850 m), irrigatedby thesetwo majorAwash tribu- tureenables monitoring of Ethiopia'santiquities and helps tarieswhich drain the adjacenthighlands to the west (el- establishpriorities for the scientificinvestigation of these evation2800 m). These and adjacent,more ephemeral resources.What is neededmost in paleoanthropologyis drainagessuch as the Sisaled'Ar and Tunfeta slice through the discoveryof new evidence.Although evidence will the elevatedPlio-Pleistocene sediments from west to east, continueto comefrom areas that are already known, it is formingdeep canyons that expose dramatic stratigraphic the evidencefrom undiscovered areas that may hold the sectionsin someplaces (FIG. 2). most potentialfor advancingscientific knowledge about The standardgeological map of Ethiopiadesignates humanorigins and evolution. In summary,the inventory rocksin the Kesem-Kebenaarea as Holocenesediments projectprovides a basicframework for moretraditional, (1:2,000,000scale; Kazmin 1975), whereasTaieb and intensive,area-oriented research to follow.We report here othersmapped this areaas Trapbasalts and Upper Pleis- on the recentlyidentified Kesem-Kebena area. tocenesediments (1:500,000 scale; Taieb 1974). Accord- ing to a recentlycompiled map (1:100,000), the Kesem- Kebenaarea consists of severalvolcanic and sedimentary The Kesem-KebenaArea: Geography rocks that become progressivelyyounger toward the In Decemberof 1988, the EthiopianMinistry of Cul- Awashflood plain (EIGS 1989). The majorunits are turelaunched its inventoryproject in a paleoanthropolog- representedby olderpre-Miocene Trap basalts that are icallyunexplored region near the northern terminus of the probablyunderlain by Mesozoicsedimentary formations MainEthiopian Rift. Furtherwork in 1989 showedthat and crystallinebasement rocks (seen on the westernes- this region,referred to here as the Kesem-Kebenaarea, carpment);Miocene and Plio-Pleistocene volcanics of the bore importantarchaeological and paleontologicalevi- rift margin;and Plio-Pleistocenealluvial fans, slope de- denceembedded in a substantialstratigraphic succession posits,and gravelterraces intercalated with pyroclastics of sedimentary,basaltic, and silicic tephra units. andbasalt flows of the riftfloor. Geologic cross-sections TheKesem-Kebena area (K-K) straddles two tributaries alongthe Kesemcanyon upstream from the Kesem-Ke- of theAwash River upstream from the MiddleAwash and benaarea expose basaltic and silicicrocks that spanthe Hadarpaleontology fields (FIG. 1). The arealies at the Mioceneperiod (23-7.7 Ma) (Rex, Gibson,and Dakin foot of the westernEthiopian Rift Escarpment,opposite 1971; Justin-Visentinet al. 1974; Jones1976). Our at- Chorora(Sickenberg and Schonfeld1975; Tiercelin, tentionwas drawn to the Kesem-Kebenaarea by TM and Michaux,and Bandet 1979; Asfaw et al. 1990), about40 LFCimagery which suggested, on the basisof reflectance, kmNW of AwashStation and 220 kmSSE of Hadar(FIGS. structural,and geomorphological considerations, that late 1, 2). Beforeour inventorywork, the Kesem-Kebenaarea Cenozoicsediments might be found(Asfaw et al. 1990). waspaleoanthropologically unknown; no vertebratefossils Initialwork in 1988 introducedthe paleoanthropological hadbeen reported from this side of theAwash River along potentialof Kesem-Kebena,and furtherwork in 1989 the 300+ km betweenMelka Kontoure in the southand establishedthe extentand content of theserocks. the MiddleAwash to the north(FIG. 1). In 1989 we establisheda campat Yalo-Kebenaon the The areais semiarid,characterized by severalgently edgeof the KebenaRiver, adjacent to the cottonplanta- droppingstep-faulted blocks dominated topographically tion headquarters(FIG. 2). This site is centralto the sed- by the adjacentrift escarpment to the westand the Awash imentaryoutcrops discovered by our pilot workthe pre- Riverand its floodplainto theeast. Access from the paved viousyear. We performedfoot andvehicle transects from AddisAbaba-Awash Station highway is viaan all-weather, the AwashRiver floodplain toward the westernrift es- unsurfacedroad that passeswest of the volcanoFentale. carpment,penetrating the roughterrain via tributariesof Fentalelies ssw of the Kesem-Kebenaarea, and offset theAwash River such as the Kebena, Tunfeta, and Kurteti, drainagesnorth of the volcanoDofan delimit the areaon whichslice from west to eastthrough the variablytilted thenorth (FIG. 2: insetmap) . A seriesof faultshas dropped sedimentarysuccession (FIG. 2). Multipletransects were theAwash River base level relative to the Plio-Pleistocene madeacross the areabetween Sabure and the NWedge of sedimentsof the Kesem-Kebena,leading to dissectionof the volcanoDofan, a distanceof approximately45 krn thesebeds by streamsdraining toward the AwashRiver, (FIG. 2). - ., L-- KK-1A; '.-,- (2.23 2025Mayz S K-K4C b *o -* -
Journalof FieldArchaeolo,gylVol.19, 1992 477
S METERS METERS METERS METER' METERS 45 * *; - . . 6 / -w.
. - . * *.---:-:: s_ ... .. - K-K6 .,-.. . ;0.
...... * * * : t - *D ,t, T Basalt Tunfetz ::. Tephra Box C.anyon - Sandstone . . .- . * *. . . * ' : ,,,s Sandy conglomerate *; -. ---. Li S; _ . ; * -. . . - 4 SiltyClay O-: ..:- L K-K1
Figure3. Simplifiedstratigraphic columns of the varioussections studied in the Kesem-Kebenaarea. Indentedlines representsofter rocks. The locationsof these sectionsare shown in Figure2. The ages of importantstratigraphic units, fossil and artifactlocations are indicatedon the sidesof the columns.
Kesem-Kebena:Pliocene exposesa sectionmore than 150 m thick (FIG. 3) that comprisesa basalfossiliferous sedimentary deposit (ca. 75 Geolo,gyand Geochronolo,gy m thick)capped by a 20 m thick,partially fractured and Reconnaissancegeological survey along the western weatheredaphanitic basalt (KK-1A) that grades to a ves- marginof the northernsector of the MainEthiopian Rift icular,amygduloidal, and porphyritic flow (KK-1B).The benveenthe Kesemand the Kebenarivers north and west thicksedimentary sequence is fluvialand lacustrine in or- of theQuaternary rift axis volcanoes of Fentaleand Dofan, igin. Calcite-cementedconglomeratic lenses underlie the respectively,revealed volcanic flows intercalated with flu- fossil-bearingsilty-clay sedimentary horizon. The baseof vialand lacustrine sedimentary deposits (FIG. 2). The rift the overlyingbasalt at the contactwith this sedimentary marginin this regionis broadand dropsgently toward unitis weathered,probably due to interactionof theinitial the rift floor.It is cut by denselyspaced and partially lavaflow with water or wet sediments.The basaltis cov- buriedNE-sw-trending normal faults that form syntheti- eredby poorlyconsolidated conglomeratic fluvial deposit callyand antithetically rotated step-faulted blocks. Several (ca. 50 m thick).The top of the sectionis cappedby a sectionswere studied and sampledto constrainthe age stronglywelded (volcanic rocks fused by heatsoon after andstratigraphic relationships of the variousunits across eruption)crystal- and lithic-richgreenish tuff (KK-1C) severalfault blocks throughout the Kesem-Kebenaarea thatis 15-20 m thick.Across from this section,similar (FIG. 3). unitsare exposed in the easternwall of a deepcanyon cut In theNW partof thestudy area, several lithologic units by a seasonalstream that runs along a transversefault areexposed along a majorNE-sw-trending fault scarp up zone.There, a thick(ca. 20 m) weldedtuff (KK-1D) caps to 200 m high.The fossiliferous K-K1 and K-K2 localities unconsolidatedgravel-rich sediments (ca. 60 m) thatare occuralong normal fault scarps and areless than2 km underlainby a 5 m thick,lithic-rich, platy, and perlitic apart(FIG. 2). TheNw-sE-trending normal faults that run weldedtuff (KK-1E).At the baseof the sectiona calcite- betweenthese two localitiesprevent the units at each from filledvesicular and aphanitic basalt (KK-1F) is exposed. beingtraced laterally, even across that short distance. Be- Geochemically,the KK-1A/Bbasalt is transitionaltho- causeof thesecross-cutting faults, the lateraldistribution leiiticwith low K20 andhigh totalFeO (TABLE 1). It is of the variouslithologic units is localin extent. enrichedin Sr,V, andBa anddepleted in Rb, likemany The rift axis-orientedfault scarp at the K-K1locality contemporaneousrift floor basalts. The weldedtuff units 478 ICesem-ICebenaPaleoanthropoloyical EthiopialWoldeGabriel et al.
Table1. X-rayfluorescence major and trace element analyses of samplesfrom the Kesem-Kebenaarea of che northernsector of the MainEthiopian Rift. Element 89KK-1A 89KK-2A 89KK-2E 89K-1 89K-2* 89K-10 SiO2 48.06 45.28 46.81 47.87 45.20 47.87 TiO2 2.71 3.01 2.85 2.25 3.02 2.40 Al203 16.39 15.81 14.54 18.29 15.76 16.38 Fe203 14.11 14.58 15.22 12.06 14.58 13.08 MnO 0.18 0.19 0.20 0.17 0.19 0.15 MgO 4.81 7.06 6.36 4.68 6.77 5.97 CaO 8.61 9.07 8.99 9.65 9.04 9.99 K20 0.98 0.92 1.06 0.75 0.93 0.80 Na2O 3.70 2.87 3.13 3.34 2.91 3.14 P20s 0.41 0.56 0.67 0.29 055 0.36 Total 99.97 99.34 99.83 99.35 98.94 100.14 Rb 4 6 8 8 6 6 Sr 548 745 396 498 504 484 Ni 24 48 43 15 9 38 Ba 483 569 703 282 302 302 Zr 170 165 211 138 123 138 Nb 15 12 17 12 12 12 Y 31 30 58 25 22 33 Zn 119 117 116 86 91 101 *Replicateanalysis of 89KK-2A
Table2. 40Ar/39Ar singlecrystal laser fusion age data on rocksand minerals of the Kesem-Kebenaarea and adjacent regionof the northernsector of the MainEthiopian Rffic. All locationsare indicated in Figure2. SampleID Location Matenal 37ARI39Ar 36Arl39Ar 40Arl39Ar % RAD J Age (Ma)
89KK-4A K-K4 Sanidine 0.0089 0.0093 3.0915 89 1.88 x 10-4 1.049 + 0.082 89KK-6A K-K6 Sanidine 0.0039 0.0023 3.0558 88 1.885 x 10-4 1.038 + 0.011 89KK-6G K-K6 Sanidine 1.7145 0.0761 2.9863 63 1.883 x 10-4 1.002 + 0.041 89K-7 Tunfeta River Sanidine 0.0296 0.0037 5.8955 96 1.885 x 10-4 2.007 + 0.022 89K-ll Tunfeta River Sanidine 0.0158 0.0085 5.4354 90 1.845 x 10-4 1.847 + 0.026
Table3. ConventionalKlAr age dataon rocksand minerals of the Kesem-Kebenaarea and adjacent region of the northernsector of the MainEthiopian Rific. All locationsare indicated in Figure2. K20 40Ar*-10 Age SampleID Location Rock type (Wt. o/0J (x 10 molelzg) 40Ar* % (MaJ 83W-43B Dofan Obsidian 4.15 0.074 20 0.12 + 0.01 89BT-128 AwashGorge Basalt 1.03 0.301 28 2.02 + 0.13 89K-10 Tunfeta Basalt 0.77 0.024 7 2.16 + 0.07 89KK-1A K-K1 Basalt 0.99 0.316 13 2.23 + 0.04 89KK-2E K-K2 Basalt 1.05 0.057 40 3.76 + 0.06 83BT-123 AwashGorge Basalt 0.96 0.779 30 5.6 + 0.32
areperalkaline to mildlyperalkaline in composition(TA- The basalflow (KK-2A)of the lava sequenceis fairly BLES 2, 4) andprobably erupted from the same center (e.g., porphyriticwith plagioclase phenocrysts. In the middleof Fentaleor Dofan;FIG. 2). A Klk datingof this basalt the section,several beds of tephraunderlie the lavasand (KK-1)above the fossiliferoussediments yielded a dateof consistof poorlyconsolidated and weathered ash fall (KK- 2.23 Ma (TABLE 3), demonstratingmajor tectonic and 2B) that is underlainby a weldedand fissilevitric tuff geomorphicmodification of the regionsince the eruption (KK-2C).The baseof the tephraunits consists of bedded of thesevolcanic units in the LatePliocene. andlaminated vitric ash fall layers(KK-2D). The basalt Thestratigraphic sequence at the K-K2locality consists flow (KK-2E)below the ash beds is stronglyweathered of basalticflows, tephrabeds, an alteredbasalt, and a andcrops out atopthe fossil-bearingsediments. fossiliferoussediment at the baseof the section(FIG. 3). Geochemicaldata (TABLE 1) indicatethat the younger Journalof FieldArchzzeolo,!y/Vol.19, 1992 479
Table4. Electronmicroprobe glass analyses of varioustephras from the Kesem-Kebenaarea normalized to 100 wt % (n = numberof glassshard analyses). ElenKnt 89KK-ld 89Kl-]e 89Kl:4al 89KK4£ 894C SiO2 72.79 72.17 75.82 67.71 69.14 TiO2 0.24 0.28 0.12 0.39 0.52 A1203 8.86 11.48 11.05 13.67 16.02 Fe203 7.17 4.32 2.89 5.30 3.12 MnO 0.24 0.26 0.10 0.31 0.09 MgO 0.02 0.14 0 0.07 0.50 CaO 0.34 0.15 0.18 1.19 1.59 BaO 0.05 0.09 0.03 0.07 0.11 Na2O 5.57 5.45 4.61 5.37 4.66 K20 4.71 5.64 5.21 5.90 4.26 Total 94.5 94.82 94.7 94.7 93.2 n 7 7 4 3 3
89K1r4F 89KK4E 89KK4G 89KK4J 89KK4K SiO2 71.97 74.89 68.07 75.95 73.12 TiO2 0.26 0.24 0.49 0.20 0.33 A1203 13.62 8.83 15.95 10.55 11.52 Fe203 3.31 7.27 3.63 3.23 5.16 MnO 0.11 0.30 0.13 0.14 0.19 MgO 0.01 0.03 0.29 0.01 0.01 CaO 0.53 0.33 1.44 0.14 0.36 BaO 0.07 0.02 0.12 0.03 0.06 Na2O 4.75 3.38 5.74 4.21 3.62 K20 5.37 4.70 4.13 5.54 5.62 Total 91.62 92.61 92.17 93.54 91.39 n 7 6 8 7 7
891RK-6A 89KK-6G 89K-3 89K-7 89K-8 89K-11 89SD-1 SiO2 75.07 69.60 73.71 71.08 71.56 73.34 72.71 TiO2 0.20 0.30 0.37 0.26 0.30 0.36 0.24 A1203 11.36 14.91 9.59 12.14 14.22 11.99 11.82 Fe203 2.97 4.06 5.98 4.03 3.14 3.88 4.62 MnO 0.12 0.16 0.16 0.24 0.20 0.23 0.17 MgO 0 0.20 0 0.1 0.21 0.11 0.01 CaO 0.17 0.26 0.30 1.51 0.70 0.27 0.39 BaO 0.03 0.10 0.01 0.05 0.15 0.03 0.08 Na2O 3.33 4.77 5.12 4.13 2.74 5.89 5.15 K20 6.74 4.65 4.78 6.46 6.75 3.90 4.81 Total 93.48 92.29 95.68 96.99 92.28 97.64 91.52 n 7 7 7 8 6 4 7 basalt(KK-2A) is slightlymore enriched in MgO,TiO2, fall (K-8 andK-9), a porphyriticbasalt (K-10) with a 1 A1203 thanthe earlyPliocene (3.76 Ma) basalt(KK-2E) m-thickbaked zone, and a crystal-richand strongly welded abovethe fossiliferous sediments. The Late Pliocene basalt tuff (K-ll). Becauseof the attitudeof the unitsand the (2.23 Ma) at the K-K1locality that capsthe fossil-rich widespreadfaulting, most of theseunits cannot be traced sedimentsis slightlydifferent from the K-K2flows. laterallyexcept in the adjacentbox canyonof the Kebena Northeastof the K-K1locality, the KebenaRiver cuts River.Based on the degreeof weldingand proxiniity, the acrosssoutheasterly dipping (S18RE) welded tuff, basaltic weldedash flow and ash fall tuff units were probably lava,and sedimentaryImits forming the wallsof a deep eruptedfrom either of the nearestQuaternary center of (ca.300 m) andnarrow box canyon.Here, a basaltflow Fentaleor Dofanto the sw andSE of the section,respec- (K-1) and a weldedtuff (K-3) were collected.About 2 tively,whereas the basalticflow is fissuralin origin.These km NE of thiscanyon, the Tunfetaephemeral stream runs unitsare buried under thick (>200 m) basalt-dominated througha similarbox canyon exposing the same lithologic sand,gravel, and conglomerate deposits that were trans- successionwith an averagedip of 20° to the SE. As shown portedfrom the adjacentwestern rific escarpment (FIG. 2). m. Figure3, the Tunfetasection consists of beddedsedi- The threetuff unitsfrom the Tunfetaseciion are gen- ments)Sthic-rich welded tuff (K-7), beddedpumiceous erallyrnildly peraLkaline, whereas the samplefrom the 480 Kesem-KebenaPaleoanthropoloyzal EthspialWoldeGabriel et al.
KebenaRiver is peralkaline,with higher iron (ca. 6 wt%) thecusrepresented by an immature,edentulous mandible contentthan the others (TABLE 4). K/Arand 40Ar/39Ar data anda rightlower third molar. The latterspecimen mea- on threesamples from the TunfetaCanyon yielded Late suresas follows:length: 22.1 mm; breadth:12.7 mm; Plioceneages (TABLES 3, 4). The 40Ar/39Ardating of sani- heightat entoconid:9.9 mm. In size andrelative crown dine separatesfrom the basallithic-rich (K-7) and the height,the specimenlies intermediate between the Hadar uppercrystal-rich (K-ll) weldedtuffunits yielded results T. dartiand the Omo ShunguraMember C specimensof of 2.01 and1.85 Ma,respectively, whereas the basalt flow T. brumpti.We attributethis specimento Theropithecus betweenthe two tuffsprovides a slightlyolder K/Ar date brumptilbaringensis. of 2.16 Ma (TABLE 3). The minorvariation in agemay be Biochronologically,theK-K1 locality is clearlyPliocene, dueto contaminationby olderrock or phenocrystsin the bestfitting between Hadar and Omo ShunguraMember basalt.The basalt is compositionallysimilar to otherflows C. Thus,a biochronologicalplacement of ca. 3.0-2.5 Ma fromthe adjacentlocalities (e.g., K-K1,K-K2, and K-1 wouldbest fit the knownfauna. The discrepancy between fromthe KebenaRiver). thisestimate and the superimposedbasalt date of 2.23 Ma will only be resolvedby furtherfield research, but both Paleontolo,gy linesof evidenceplace the K-K1occurrence firtnly in the Pliocene.This is a periodof considerableimportance in LOCALIItYK-K2 mammalian,and particularly hominid, evolution in Africa. As noted in the precedingsection, two vertebratelo- TheKesem-Kebena Pliocene sediments described in the calitieswere described for the Kesem-Kebenaarea, K-K1 previoussection on geologycomprise many beds depos- and K-K2. The older locality,K-K2, is locatedin an ited by rapidlyflowing water. This, and the factthat the embaymentof an escarpmentcalled "Adon-Adas" (FIG. bedsare not well exposedhorizontally, contribute to the 2). Here,the faunacomprises fragmentary chelonian and limitednumber (two) of Pliocenevertebrate localities es- proboscideanremains found in situ and on an outwash tablishedby inventoryproject personnel. Isolated verte- lag surfacebelow some silts that underlie the basaltdated bratefossils, Hippanon teeth and a phalanx,were found at 3.76 Ma. on Pliocenestrata exposed in thewalls of the Kebenariver canyonNE of theK-K1 and K-K2 localities. More detailed LOCALITYK-K1 paleontologicalsurvey will undoubtedlyproduce addi- Themost important Pliocene vertebrate locality thus far tionalvertebrate paleontological localities in the Kesem- discoveredin the Kesem-Kebenaarea is K-K1,found by KebenaPliocene strata. The inventoryteam found no YonasBeyene in 1988 on a steepslope ca. 13.5 krnnorth evidenceof lithictechnology or hominid-basedbone mod- of Sabure(FIG. 2). Dozensof vertebratefossils were ex- ificationin the Pliocenedeposits. posedon the surfacehere, and a searchof the slopeiden- tifiedthe fossiliferousstratum as a graysilty sand poorly exposedon the steepslope about25 m belowthe 2.23 Kesem-Kebena:Pleistocene Ma basalt.The excellentpreservation of the vertebrate Geolo,gyand Geoclyronolo,gy remainsand theirabundance indicate that a productive quarrycoid be establishedon thishillside. The volcanicflows and the interbeddedsediments of The K-K1 surfacefossil assemblageis diversein taxa the Kesem-Kebenaarea get progressivelyyounger in age and body parts,with fish, turtle,crocodile, and hippo- towardthe riftaxis and away from the western rift escarp- potamusremams . mc. lcatmg. . a nearry permanentwater ment.The LatePliocene rocks are generallycovered by sourceat the timeof deposition.The bovidassemblage is tephrasin elevatedterranes and tephras and sediments in diagnostic,comprising, cf. Reduncini,Tragelaphini( ?), cf. the low-lyingareas. Several sections consisiing of Earlyto Damaleps,Aepyceros cf. shun,gurae,and SynceruslU,gandancMiddle Pleistocene volcanic and sedimentary deposits oc- sp. The Aepycerosis largerthan the Hadarform, and curalong fault scarps and stream cuts. In theBara locality, matchesthe Omo Shungura B/C or morerecent represent- about3 km SE of the K-K1localibr (FIG. 2), vertebrate atives.The bovinemandible shows a fourthpremolar bone-bearingconglomeratic and silbr sediments occur be- whosemetaconid slants posteriorly, a trait seen in Syncerus low reworkedtuff and ash fall that are cappedby an fromOmo ShunguraB and C, fromKoobi ForaArea agglomeraticunit. Northeast of theBars section and alsout 203, andfrom Hadar. Other details, however, differ from a kilometereast of the K-K1 localibr,clastic beds and the Hadar Uanduc. tephraunits crop out alonggullies. Most of thesesedi- One primateis knownfrom K-K1, a formof Theropi- mentsare devoid of fossilsexcept for localized,low den- lournal of FieldArchaeoloflylVol. 19, 1992 481
Figure4. The KK-4 localityviewed from the west. The arrowpoints to two paleontologistsstanding on the hillsidewhere a rich vertebratefauna was documented. sity,faunal and floral assemblages, with no fossilfragments callyvariable (TABLE 2). Theyrange from alkaline to per- foundin situ. alkalinein composition,suggesting variable source areas TheLate Pliocene rocks exposed along the Kebenaand or a centerthat eruptedcompositionally variable flows. TunfetaRiver Canyons are overlain by lithic-richwelded SampleKK-4A contains tsvo types of volcanicglasses tuffs,ash falls, and gravel and silty clay deposits of fluvial (KK-4A;KK-4A2) suggestive of magmamixture prior to origin.The weldedtuff units formresistant cliffs along eruption.Some of theglasses are silica-poor and iron-rich, faultscarps and stream cuts. whereasthe oppositeis truefor the othertype (TABLE 2). In the K-K4locality, north of the Tunfetastream, sev- Someof thetephra units were recognized in othersections eraltephra layers are exposed betsveen fossil-bearing sed- aroundthe K-K4locality. Major element contents of KK- iments.There, a weldedtuff (KK-4A)occurs below and 4G and KK-4Care similar, whereas some of the glasses abovea pumice(KK-4B) and an ash fall (KK-4C)unit, in KK-4Aare compositionallycorrelative to glassesin respectively(FIG. 4). Thesetephras blanket light-brown KK-4J. fossiliferoussediments. About a kilometerto the NW of Sanidineseparates from the lithic-rich welded tuff (KK- thissection, a beddedvitric ash (KK-4E) that grades down 4A) abovethe fossiliferoussediments yielded an 40Ar/39Ar to a stronglywelded fiamme-rich greenish tuff (KK-4D) ageof 1.05 Ma.Based on the localstratigraphic relations, cropsout. Theseunits are gently tilted to the SE. About thisunit is olderthan most of the ashfalls except for the 3 km NE of the fossil-richsection (KK-4A), tsvo ash fall southeasterly-tilted,green welded tuff (KK-4Dand E). units(KK-4F and G) anda denselywelded tuff (KK-4H) Northeastof the K-K4locality, several fossil- and arti- areintercalated with light reddish-brownsilty claysedi- fact-richsediments crop out alongstream gullies (FIG. 2). mentsand occur along deeply dissected ephemeral stream At the K-K3locality, a vitrictuff (KK-3)occurs betsveen banks.Southeast of the KK-4Asection, thin (<10 cm) tsvoconglomeratic and silty clay deposits along an older ashlayers (KK-4I, J, andK) occurinterbedded with thick streamterrace on the southside of the Kurtetiseasonal (15-20 m) andconsolidated light orange-brown sedimen- streambed. The fossiliferoussediment overlies the vitric tarydeposits along the SissaleD'ar stream banks. tuff. The silicictephras of the K-K4locality are geochemi- The K-K6locality to the NWof the KK-3outcrop is a 482 Kesem-KebenaPaleoanthropoloyical EthiapialWoldeGabriel et al.
heavilydissected badland topography containing the rich- area(TABLE 4). For example,a tuff unit (KK-1D)from est fossiland artifact concentration in the Kesem-Kebena the K-K1locality is similarto KK-4Ethat crops out in a area.The fossil-rich sediments are underlain by thick (>30 streambelow the KK-4Awelded tuff (1.05 Ma). Both m), pumiceoustuff units that are exposed along deep and KK-4Aand KK-6Ayielded similar ages (TABLE 2) and narrowbox canyonsof the Harafteliseasonal stream and havevery similar chemistry (TABLE 4). Anotherunit (KK- its adjacenttributaries (FIGS. 2, 3, and6). Thesetephras are 4J)from the vicinity of the K-K4locality is alsocorrelated tiltedto the SE (5-10°). The basalunit (KK-6A)consists to both tuffs. Some of the youngertephras from the of poorlysorted coarse pumice fragments (<15 cm di- adjacentK-K4 and K-K6 localitiesare also sirnilar.For ameter)embedded in a matrixof vitricash, and grades to example,KK-4G and KK-6G (1.0 Ma)have sirnilar major an ash fall (KK-6B).This unit is followedby a fines- elementchemistry. The SD-1 tuff exposedSE of the K- depletedpumice fall (KK-6C). A fairlyweathered yellow- K1 locality(FIG. 2) iS alsocorrelated to K-ll (1.85 Ma) ishpumice bed blanketed by a pumicefall (KK-6D) forms fromthe top partof the TunfetaCanyon (FIG. 3). the top of the pyroclasiicsequence. The upperhalf of the sectionis dominatedby the fossil-and artifact-rich lacus- Paleontoloflyand Archoleolofly trineand fluvialvolcaniclastic and silty claydeposits. A The Pleistocenedeposits in the Kesemdrainage area partiallyreworked pumice fall (KK-6G)occurs in the westof AwaraMelka (Sabure) are sterile, except for abun- lowerpart of the sedimentarydeposit and most of the dant rootcasts,but their counterpartsto the north are fossil-and ariifact-rich horizon crops out justbelow it. A increasinglyfossiliferous and irnplementiferous,particu- thin(ca. 10 cm),dark-gray ash layer (KK-6E) intercalated larlywest of the volcanoDofan (FIG. 2). Al1of the tran- with the silty clayoccurs at the top partof the section. sectsnorth of the Tunfetadrainage encountered archaeo- To the NNW of the K-K6locality, a 20 m-thick,southeas- logical evidencein the form of flakes,debitage, and terlydipping (ca. 10°), denselywelded tuff (KK-6F)un- occasionalisolated handaxes. The establishedlocalities all derliesbedded pumice fall. Isolated low densityfossils and representoccurrences with concentrationsof vertebrate artifactswere discovered in the sedimentsabove the pum- fossils.Our survey of middlePleistocene rocks established ice fall. fourpaleoanthropological localities in the Kesem-Kebena Northeastof the K-K6locality, a 1.5 m thickuncon- area. solidatedvitric ash (KK-7)was recognized between light brownsilty clay deposits. Along the seasonal stream banks LOCALITYK-K4 of the K-K7locality, the sandyclay deposit above the tuff The most southerlylocality, K-K4, is a richvertebrate is about10 m thick,whereas the underlyingmassive de- localitywith a diversesurface and exposedin situ fauna positis about4 m thick.Fossils and artifacts eroded from (FIG. 4). No artifactswere seen directlyassociated with sedimentsabove the vitrictuff horizonwere recognized the K-K4vertebrate assemblage, but about20 m to the m. t le sectlon.. west,in thesame geological unit, a denticulatedsidescrap- Geochemicaldata from the basalash flow tuff (KK-6A) er madeof weldedtuff was foundin situ.Excavations of andthe partiallyreworked pumice fall (KK-6G)indicate the poorly-consolidated,fossiliferous silty sandswhich a peralkalineand mildly alkaline composition, respectively. yieldthe diverse,well-preserved fauna will be necessaryto The 40Ar/39Ardata, however, suggest that the tephrasare establishthe archaeologicalcontent of the locality-a few similarin age.Ages of 1.04 and1.0 Mawere obtained on flakesand chunks, mostly of weldedtuff, wereidentified the basalpumiceous tuff and the pumicefall abovethe on adjacentslopes, and moresurface artifacts, including fossiliferouszone, respectively (TABLE 2). Thesimilarity in handaxes,sidescrapers, and unretouched flakes on welded age andthe variationin chemistrysuggest that the units tuffand chalcedony, were found ca. 500 m to the west. wereerupted from separate source areas more or less at The K-K4 faunalassemblage shows a wide rangeof thesame time as the tephras of theK-K4 locality, probably bodyparts, some articulated, and includes the taxaMet"- fromDofan and Fentale(FIG. 2), the closestQuaternary diochosorusmodestus (partial cranium), Tragelaphini, Alcel- volcaniccenters in the area.At Dofan,just east of the K- aphini,Antidorcas cf. recki,a largeEgyuas, Theropithecus K3, K-K6,and the K-K7localities, some of the youngest oswaldi(distal tibia), Colobus sp. (2 maxillae),and Viver- obsidianflows (83W43B)are much youngerthan the ridae.The biochronologicalplacement of this faunais tephraunits encountered at the K-K6locality to thenorth rendereddifficult by the lirnitedsample size of the surface (TABLE 3). specimens,and by a rarityof well-datedmiddle Pleistocene Basedon majorelement glass chemistry, some of the assemblageselsewhere in Africa.It is notablethat aquatic tephraunits were correlated to eachother within the study elementsare absent in this assemblage.The depositsare Journalof FieldArchaeologylVol. 19, 1992 483
vertebratefauna including the suidKolpochoerus majus, a taxon also found in the Bodo depositsof the Middle AwashValley (Harris and VVhite 1979; Clark et al. 1984), as well as the remainsof Rhinocerotidae,Hippopotami- dae,Alcelaphini, and Antilopini. The bonewas well fos- silized,and specimensrange from unabradedto highly rolledbone pebbles. The unityielding these remains also containeda few in situwelded-tuff flakes and chunks, but no handaxeswere present.The artifacts,although una- braded,were judged to be in secondarycontext. Outcrops of the same geologicalunit were tracedacross to the westernbank of theKurteti where they were also observed to yieldvertebrate fossils and occasional artifacts.
LOCALITYK-K6 TheK-K6 locality is anextensive outcrop of Pleistocene sedimentsfarther west of the volcanoDofan in the Har- aftileand adjacentdrainages (FIGS. 2, 6). Concentrations of artifacts(particularly Acheulian bifaces) and vertebrate fossilswere found in an areameasuring approximately 50,000 sq m, but we choseto identifythe entirecontin- uousoutcrop as one localitybecause at leasta thinscatter of artifactsand fossils was presentwherever the relevant sedimentsoutcropped. The locality is dissectedinto thirds by the two forksof the Haraftiledrainage (below whose junctionthe drainageis calledthe "Edegaho").Fossils and artifactswere found from the top to the bottomof this section(FIG. 5), but a widespread,light-grey tuffaceous deposityielded the greatestconcentration of surfacearti- factsand vertebrate fossils. Abundantvertebrate fossils found on surfacesof the K- K6 outcrops,particularly on the lightgrey tuffaceous de- posits, includedremains of Elephassp., Bovini,Alcela- Figure5. The K-K6 locality.View is towardthe NNW, acrossa tribu- phini, Reduncini/Hippotragini,Kolpochoergs olduvaiensis, taryof the Harafteli.The stratigraphicsection visible in this photo- Phacochoergssp., HippopotamusSp. (FIG. 6), Rhinoceroti- graphis depictedin Figure3. Acheulianartifacts and fossil vertebrates arefound betweenthe volcanicstratum seen croppingout on the dae,Equus sp., Crocodilia,Pisces, and Aves. Fish, croco- lowerleft (KK-6A)and the one croppingout in the upperright (KK- dile,and hippopotamus fossils were particularly abundant, 6G). Theseunits are datedat 1.04 and 1.00 millionyears, respectively. indicatinga permanentwater source at the time of de- position.Found on surfacesof the sameoutcrops were a not indurated,and the lens of fossilsis extremelyrich. low-densitybackground scatter of flakes,and a fewbifaces, Excavationof thislocality would probably yield a sizeable mostlyon weldedtuff and rhyolite. fauna.Survey in the nearbySissale D'Ar drainage located In onelocation, KK-6b, benveen the Haraftiledrainage additionalfossils, including a smallHipparion and an al- arms,an approximately20 x 20 m remnantpatch of celaphine. tuffaceoussilts had exposedon its surface21 unabraded handaxes(FIG. 7). Onewas found- in situin the lightgray LOCALITYK-K3 tuffaceousdeposit. Very few whole flakes and sidescrapers TheK-K3 locality lies on theeastern bank of theKurteti werefound with the handaxes(they had obviouslybeen drainage.The stratum-yieldingthe K-K3paleontological madeelsewhere and imported), in contrastto otherK-K6 and archaeologicalremains is an indurated,bioclastic exposuresof the samestratum where flakes were common gravel-sandmeasuring up to 1 m thickand containing but bifacesrare. Melanoidestuberculata shells. This unit yieldeda small In additionto the miscellaneoussmall tools andflakes, 484 Kesem-KebenaPaleoanthropolo,gical EthiopialWoldeGabriel et al.
shallowtrimming is presenton the ventralfaces. Flaking wasprobably done with hard hammer (handaxes found at otherlocations within the K-K6locality showed evidence of soft hammerretouch). Many of the handaxesretain corticalbutts, showing that the raw material took the form of streamworn bouldersand cobbles.All of them are heavy-duty;sections are plano-convex for the specimens conservingthe cortex,and bi-convexfor those without cortex.The lone exception is a trihedralpick with a quad- rangularsection. These tools and flakesare extremely fresh,with unalteredphenocrysts in the flakescars and with sharp,unabraded edges, all circumstances indicating minimaldisturbance or transport. The handaxesfrom K-K6 fall into fourtypes, pointed handaxes,trihedral picks, converging cleavers (hachereau bifaces),and cleaver and core axes. Dimensions of 19 K- K6bbifaces range in maximumlength from 152-222 mm Figure6. A fossilizedhippopotamus cranium from the K-K6 locality, (mean=183mm), maximumbreadth from 72-120 mm
. . lRl SltU. (mean=99mm). Thicknessvalues show a homogenous "thick"aspect to thesetools. Bifaces from other outcrops the K-K6bartifact assemblage comprises 7 rhyoliteand of the samegeological unit were not so spatiallyconcen- 14 weldedtuff handaxes(FIGS. 8A, B). All of these are trated,and included thinner, more finely-flaked forms than madeon largeside- and end-struckflakes. Subsequent thoseobserved at KK-6b. deep,oblique flaking scars mark the dorsalsurfaces, and
Figure7. A view of localityK-K6 west of KK-6B, an archaeologicaloccurrence with abundantAcheu- lian handaxes.View is to the east. The hills on the most distanthorizon represent the top of Dofan volcano(see FIG. 2). Journalof FieldArchaeolo,gylVol. 19, 1992 485
A
Figure8. Bifacesfrom the KK-6Blocality, dorsal and ventralfaces illustrated. B 486 Kesem-KebenaPaleoanthropoloyical EthiopialWoldeGabriel et al.
The depositionalcontext of the Kesem-Kebenalocality Polleontoloflyomd Archaeolofly K-K6 fossils and their co-occurrencewith unabraded The thin veneerof late Pleistocenesediments between stonetools on the freshly-erodedoutcrops strongly implies localitiesK-K3 and K-K6 has deflateddue to erosionat contemporaneityof theseremains in undisturbed,primary localityK-K5, revealinga rich concentrationof Later contexts ideal settingsfor archaeologicalinvestigation. StoneAge artifacts(FIG. 10). The localityis placedon a Thereare manyoutcrops of the fossiliferous,soft grey terraceabout 20 m abovethe localdrainage floors. The tuffaceoussilt in locations wherevery little overburden artifactsinclude fire-altered rocks, endscrapers, flakes from wouldhave to be removedin a horizontallylarge exca- preparedcores, sidescrapers,microliths, blade cores, vation.Furthermore, study of the lateralvariation in as- blades,and debitage. Most of the artifactsare on welded semblagetypes and stone tool concentrations(as in Jones tuff, but some areon obsidianand chalcedony. There is 1979; Potts 1989) shouldprove highly effective in this friablevertebrate bone and some charcoal associated with region.Thus, the Kesem-Kebena MiddlePleistocene de- the assemblage.There are many large blades on obsidian posits are full of potentialfor futurearchaeological re- andwelded tuff at this extensivelocality, but noneof search. the faunawas identifiable.Smaller occurrences with similar contentwere seen in much of the areawest of Dofan LOCALITYK-K7 volcano,always unconformably superimposed on the ear- The K-K7locality lies NE of K-K6.It yieldedremains lierPleistocene sediments. of Kolpochoeruscf. majus, Phacochoerus sp. and medium and MiddleStone Age assemblagesare notably absent from smallbovids, as well as lateAcheulian artifacts. The sedi- the Kesem-Kebenaarea, with the exceptionof surface ments themselvesare judged,from field aspect,to be occurrencesnear the "AsaKela" drainage where Levallois youngerthan those of K-K6.These sediments represent flakes,and cores,small bifaces, denticulated sidescrapers a good lithologicaland archaeologicalmatch for occur- andnotched flakes on weldedtuff wereencountered on renceson the easternside of the AwashRiver at Melka the surfaceduring survey. No localitywas assignedto Werer(discovered by the inventory project in 1988-1989; theseassemblages. (FIG. 2). Diminutiveovate and elongate ovate bifaces on obsidianand welded tuff (with biconvex sections and shal- low retouch)are common in thisarea, but all were surface Plio-PleistoceneEvolution of the Kesem-Kebena finds.The K-K7locality marked the northernmostextent Area of the 1989 inventoryproject work in this area.Further In its northernsector, the MainEthiopian Rift (MER) researchto extendknowledge northward is planned,as is broadenstoward the Afardepression. The riftmargin in workto correlatebetween the volcanic horizons of Kesem- the Kesem-Kebenaarea is characterizedby broad,step- Kebena(particularly at K-K7)and those of MelkaWerer. faulted,synthetically- and antithetically-tilted,and exten- sivelydissected blocks that gentlydrop towardthe rift Kesem-Kebena:Late Pleistocene floor.According to Mortonand Black(1975), this is a manifestationof riftingand crustalthinning that devel- Geolofly oped as a resultof normalfaulting and blocktilting and Siltyand conglomeraticsediments are disconformable ductiledeformation within the upperand lowercrustal abovethe youngestvitric ash depositsalong stream ter- layers,respectively. Mohr (1983), however,disputed this racesin the vicinityof most of the fossil-and artifact- hypothesisand suggested an alternativeprocess of dilation bearinglocalities described above. These later sedimentary andductile deformation that was enhancedby heatflow depositsoccur primarily as isolatedpatches. One of the fromdike injections to createthis kindof riftingmecha- well documentedoutcrops occurs at the K-K5 locality nism.Although no dikeswere observedin the Kesem- betweenthe K-K3and K-K6Pleistocene sections (FIGS. Kebenaarea, magma injections that led to basaltfissural 2, 9). The sedimentaryrocks at localityK-K5 consist of a eruptionsalong border faults of themarginal grabens may moderatelyconsolidated light-brown, silty clay with con- havefacilitated dilation and block tilting, resulting in the glomeraticlenses exposed along a streambank to thenorth presenttectonic and geomorphic features of the broadly- of the Kurtetiseasonal stream. In a terraceadjacent to the faultedwestern rift margin. The riftshoulder of the west- northernbank of the streama dense concentrationof ern escarpmentof the northernsector of the MER be- artifactswith small fragmentsof bone was discovered tweenAddis Ababa and Debre Berhan (FIG. 2) consistsof (FIG. 10). Oligoceneand Miocene basaltic and silicic flows, whereas Joxrnal ofFieldArchaeolo,gylVol. 19, 1992 487
Figure9. LocalityKK-5, a Late Stone Age occurrence.View is to the east. Artifactsare seen on the surface,recently deflated from a thin veneerof late Pleistocenesediments that mantlesthe Earlyto MiddlePleistocene beds below.
Figure10. Artifactsfrom LocalityKK-5. 488 Kesem-KebenaPateoanthropotogicat EthiepialWotdeGabriet et at.
the youngervolcanic sequence is dividedinto Miocene the Kesem-Kebenaarea directly across from the Chorora 'sAlajiSeries" (15-16 Ma) and Pliocene"Balchi Series" sedimentsare solely confined to the foothillsof the west- ( Justin-Visentinet al. 1974; Zanettinet al. 1974).In the ernescarpment, and such sediments are totally absent from AddisAbaba sector of the westernriE-margin Early Mio- the AwashGorge to the south.The AwashGorge near cene (22 Ma) silicicsare the dominantunits (Mortonet the KarayouLodge of the AwashNational Park exposes al. 1979). Riftingin the northernsector began in the morethan 150 m of LateMiocene to LatePliocene (5 6- MidclleMiocene (15-14 Ma) as a broaddownwarped 2.1 Ma;TABLE 3) basaltflows that are capped by densely basinwith marginalgrabens along the foothillsof both weldedPleistocene tuff layers. The tuff beds are separated escarpments(Kazmin and Berhe 1978). Althoughthis by a ca. 50 cm-thickpaleosol. kindof structuralfeature is locallypresent in the region As presentedin the geologicdescriptions of eachlocal- today,the ancestralstructural and geomorphicsetting ity,the Kesem-Kebenaarea is characterizedby intercalat- probablylooked like the marginalgrabens of the central ing depositsof sedimentaryand volcanicunits that have sectorof the MER. The rift axisin the centralsector is beencut by swarmsof riE-axis(NE-SW) orientednormal bifurcatedand runs alongmarginal grabens represented faultsthat areoccasionally displaced by transversefaults by the Wonjiand Silti-Debre Zeit FaultBelts of the east- (FIG. 2). Today,there is more than 300 m of vertical ern and westernescarpments, respectively (Mohr 1967; displacementbetween the K-K1and K-K2 localities and Di Paola1972; WoldeGabrielet al. 1990). The median the floodplain of the AwashRiver and its majortributar- partof the riftfloor acts as a watershedfor bothmarginal ies, the Kesemand the KebenaRivers, that drainacross grabens,and most of the lakesin thisregion are confined thestudy area from the adjacentrift shoulder (FIG. 2). The to the easternmarginal graben (i.e., LakesZiway, Lan- Kesem-Kebenaarea has dramaticallyevolved geomor- gano,Abyata, etc.). The fossiliferous fluvial and lacustrine phologicallyfrom a gentlydipping and downwarped ter- sedimentsof the Kesem-Kebenaarea occur at the baseof ranedraped by floodbasalt and tephra units of theMiddle severalstep-faulted blocks along the foothillsof the west- to LateMiocene time (FIG. llA) to an areadominated by ernrift escarpment. These sediments were not recognized antithetically-tiltedfault blocks that bounded several par- in the AwashRiver Gorge close to the presentday axis of allelmarginal grabens characterized by flood plainand the riftfloor that exposes Plio-Pleistocene basalts and sil- freshwater depositional environments during the Pliocene icictephras (WoldeGabriel 1987). The confinement ofthe (FIG. llB). In the Pleistocene,the Kesem-Kebenaarea sedimentsto the Kesem-Kebenaarea is attributedto mar- continuedto be affectedby majorbasaltic and silicic erup- ginalgrabens that acted as sedimenttraps during the ev- tionsand intense faulting and tilting and evolved to a set olutionof the escarpmentand the adjacentrift floor.A of gentlydropping step-faulted and rotatedblocks (FIG. generalizeddiscussion on the structuralevolution of the llC) that currentlyfeed the Awashfloodplain along the Kesem-Kebenaarea that highlights the existenceof ances- axisof the rift with abundantsediments of volcanicand tralmarginal grabens is presentedbelow. sedlmentary. orlgm.. . Theexistence of ancestralmarginal grabens in thenorth- Thefossiliferous sediments of the K-K2locality under- ernsector of the MERis indicatedby the occurrenceand lyingthe Pliocenebasalt (3.76 Ma) arewell sortedand the confinementof Neogenesediments along the present aredominated by silty clay probably deposited in a shallow foothillsof both escarpments(FIG. 2). For example,it is waterenvironment. The presenceof shallowwater or wet suggestedthat the LateMiocene Chorora lacustrine sed- sedimentsis also indicatedby the strongweathering in iments(FIGS. 1) 2) weredeposited in a narrow,marginal the overlyingbasalt, most likelyas a resultof waterand grabenthat was bounded on its westernmargin by a horst hot lavainteraction during the eruption.This basaltis andseveral Neogene silicic centers (Gara Ghumbi, Asebot, generallyweathered, unlike the overlyingtephra and bas- Atcem, Wo.c oyi, etc.) that eruptedalong the western alticunits. Thus the eruptionof thesevolcanic units filled borderfault of this marginalgraben (Kazmin and Berhe the basinat the K-K2locality, disrupting the sedimenta- 1978).The thickening of theNazret Group Silicics within tlon. processes. this old marginalgraben and theirtotal absencein the Thefossiliferous sedimentary beds at the K-K1locality AwashGorge to the westand close to the presentday rift gentlydip to the westand are terminated by a transverse axissupport the argument.The basaltflows at the Awash faultthat runs between this section and that of the K-K2 Gorgesection are younger (Pliocene), however, and the locality.As shownin Figure3, the lithologicunits at the lateMiocene sediment or tephrasof the Chororadeposits adjacentK-K1 and K-K2 localities are different, consistent maybe buriedbelow the Pliocenebasalts of the gorge. withthe agesof the overlyingbasalts, the lithologyof the The Plio-Pleistocenefluvial and lacustrinesediments of sedimentaryunits, and the typeof fossilassemblages. The Journalof FieldArchaeolo,gylVol.19, 1992 489
Figure11. Schematicblock diagram on the geologicaland structuralevolution of the Kesem-Kebena area.Figure 11A representsa downwarpedregion prior to EarlyPliocene intense faulting and marginal grabenformation (FIG. llB). Antitheticblock faulting and tiltingwere dominant.The marginalgrabens actedas sedimenttraps that depositedin freshwater environments during rift evolution.In the Pleisto- cene the marginalgrabens were extendedand filledwith sedimentsand volcanic rocks, and subsequently faulted.Older rocks and structureswere buriedunder younger units, and todaythe areais characterized by synthetically-tilted,step-faulted blocks (FIG. 11C).
calcite-cementedconglomeratic sediments and the over- 50 m of poorlysorted, non-fossiliferous sands and gravels lying freshwater sedimentsat the K-K1 localityare over- weredeposited above the basaltflow. The eruptionof an lain by a 2.3-Ma basalt.The poorly-sortedbasal conglom- earlyPleistocene, densely welded tuffthat caps the section eraticlenses represent a masstransport of basalticcobbles heraldedthe endof this basin.Most of the K-K2locality fromthe adjacentOligocene and Mioceneflows of the rift rockswere probably falllted before the accumulationof shoulder.The volcaniclasticbeds were depositedin a fresh the K-K1units. Both localities were subsequently affected water environmentas indicatedby the fossil remainsof by majorfaulting episodes in the EarlyPleistocene. Today bovids,fish, turtle,crocodile, and hippopotamus.Fissural thispart of the studyarea is characterizedby step-faulted basaltflows probablyerupted along borderfaults of the blocksthat drop gently toward the riftaxis. The low-lying marginalgrabens and filled the sedimentarybasin about areais dominatedby Pleistocenegravel deposits that are 2.23 Ma. This basinpossibly continued subsiding, and ca. intercalatedwith minortephra deposits. These units are 490 Kesem-KebenaPaleoanthropolo,gical EthiopialWoldeGabriel et al. cut by swarmsof denselyspaced normal faults of recent in the deepsea drillcores from the nearbyGulf of Aden age (FIG. 2). (Sarna-Wojcickiet al. 1985) holdsgreat promise for the Northeastof the K-K1and K-K2 localities, the Kebena futureof paleoanthropology.The wealth of Achelllianlo- Riverruns along a majortransverse fault that has displaced calitiesin and adjacentto the Ethiopianrift and Afar mostof the olderrocks to the east.The olderunits rec- depression,spread across a wide rangeof environments, ognizedat the K-K1 and the K-K2 localitiesare not holdsgreat promise for archaeologists and paleontologists. exposedin the deep canyoncut by the lower Kebena The PaleoanthropologicalInventory of Ethiopiahas River.The sectionhere consists of youngerrocks and is confirmed,in the two yearsof its existence,the presence cappedby a denselywelded early Pleistocene tuff. The of severalnew areasof greatpotential for geological,ar- faultingprobably preceded the depositionof theserocks. chaeologicaland paleontologicalresearch. For example, The poorlysorted and consolidatedgravel deposits that theBurji area is Ethiopia'sfirst Middle Miocene vertebrate makeup most of the section(>250 m) were probably paleontologicalarea (WoldeGabriel et al. 1991). The Bi- transportedfrom the adjacentrift shoulder.In Figure3, late areahas extensiveoutcrops of late Pleistocenesedi- the lithologicunits are representedby intercalatingvol- mentswith a richrecord of lithictechnology. Inventory canicflows and sedimentary deposits as in the K-K1and work has establishedthe Chororaand Fejejregions as K-K2localities, suggesting the existenceof grabensthat large,stratigraphically complex, multiple component re- actedas depositionalenvironments. Subsequent faulting searchareas with greatpotential for futureinvestigation andtilting of olderfault blocks and basaltic eruption mod- (Asfawet al. 1990;Asfaw et al. 1991).The Plioceneand ifiedthe basin,halting sedimentation processes. The age Pleistocenedeposits of the Kesem-Kebenaarea, with their of the rocksdecreases eastward and toward the rift axis. interstratifiedvolcanics, well-preserved vertebrate fossils, The denselywelded tuff thatcaps most of the unitswest andextensive archaeological evidence, will now join other of the K-K4locality thickens to the sw andwas possibly bewerknown paleoanthropological research areas as the eruptedfrom Mt. Fentale.In the NE halfof the Kesem- questto revealhuman origins and evolution continues. Kebenaarea, the ashflow tuffsand vitric ashes are blan- ketedby fluvialand lacustrinesediments consistent with the freshwater fossil remains recognized in the K-K3,K- Acknowledgments K6, andK-K7 localities. The thickgravel deposits of the ThePaleoanthropological Inventory project of theEthi- KebenaRiver section are not apparentin K-K4,K-K6, opianMinistry of Cultureand Sports Affairs was begun andK-K7 localities (FIG. 2). underMinister Girma Yilma and is supportedby the Com- miKeefor Researchand Exploration of the NationalGeo- graphicSociety (4134-89), the Anthropology Program of Conclusion the NationalScience Foundation (BNS88-19735), the TheKesem-Kebena area, like some of the other,bewer- Wenner-GrenFoundation for AnthropologicalResearch, knownpaleoanthropological areas in Africa,samples a theJapan Society for Promotion of Science,and the Japan substantialrange of time. The abundanceand variety of ShipbuildingIndustry Foundation. Part of this workby volcanicrocks interstratified with fossil-and artifact-bear- GWGwas done underthe auspicesof the U.S. Depart- ing stratain the Kesem-Kebenahold the promiseof ex- mentof Energy,Office of BasicEnergy Sciences Research, cellentgeochronological calibration and eventual tephro- andby the Earthand Environmental Sciences Division of stratigraphiccorrelation with other areasin the region. theLos Alamos National Laboratory. Typing and editorial Forexample, the Middle Awash (now "Messalu-Gewane"; work by BarbaraHahn and drafting by AnthonyGarcia, Clark1987; White1986; Williamset al. 1986), located both of Los AlamosNational Laboratory, is greatlyap- farthernorth in the Afardepression of Ethiopia,shares preciated. with Kesem-Kebenaa long sequenceof Pleistocenesedi- Thanksto J. Heirtzler,Head, Geology and Geomagne- mentswith Achellliantools and vertebrateremains. Ar- tism Branch,NASA/Goddard Space Flight Center,for chaeologistsare currently seeking pawerns in the climatic, providingcollaborative access to LandsatThematic Map- biological,and technologicalrecords of the middleand perand Large Format Camera images; and to C. Ebinger latePleistocene hominids to bewerunderstand the origins and D. Hardingfor their advice,encouragement, and of our species(Binford 1981, 1984; Clark1988; Right- support.Thanks to ElisabethVrba for assistancein bovid mire 1985; Wolpoff1984). The possibilityof linking identification,and to AndyCohen for assistancewith the these areasby isochronoustephra, and of linkingtheir invertebratefossil identification.The 1988-1989 field terrestrialAfrican records of globalclimatic change found team at Kesem-Kebenaincluded Tadesse Terfa, Head, Joutnalof FieldArchaeolo,gylVol. 19, 1992 491
Centrefor Researchand Conservationof CulturalHeri- Binford,Lewis R. 1981 Bones:Ancient Men and ModernMyths. Orlando, FL: tage, and Alemu Ademasu,Tekele Hagos, Yohannes AcademicPress. Haile-Selassie,Pelaji Kyauka, Agazi Negash, Chernet Ti- lahun,and Tamirat Wodajo. Thanks to the officialsof the 1984 FaunalRemains from Klo/sies River Mouth. Orlando, FL: AcademicPress. NationalMuseums of Ethiopia(Mamo Tessema and Wol- desenbetAbomssa) and Kenya for accessto comparative Bishop,William W. 1967 "AnnotatedLexicon of QuaternaryStratigraphical No- materials,and to J. DesmondClark for encouragement menclaturein EastAfrica," in WilliamW. Bishopand J. and advice.Thanks to AkelomMohamed, Ishmail Hus- DesmondClark, eds., Bacround to Evolutionin Afiica. sein,Ali Abduke, Ebrahim, and Ali Babile theAfar peo- Chicago:University of ChicagoPress, 375-395. ple whosefield support made the projectpossible. Brown,Frank H., andCraig S. Feibel 1986 "Revisionof LithostratigraphicNomenclature in the Koobi Fora Region, Kenya,"Journal of the Geolo,gical Societyof London 143: 297-310. Giday WoldeGabrielis ageoloyist at the LosAlamosNational Laboratotyin New Mencico.MailinB address:Earth and Envi- Chavaillon,Jean, Nicole Chavaillon, Francis Hours, and Marcello Piperno ronmentalScience Division, EES-1ID462, LosAlamosNa- 1979 "Fromthe Oldowanto the MiddleStone Age at Melka- tional Laboratory,LosAlamos, NM 87545. Kunture(Ethiopia), Understanding Cultural Changes," Tim White is a professorin the Departmentof Anthropolofly Quaternaria21: 87-114. MailinB address: at the Universityof California in Berkeley. Clark,J. Desmond Departmentof Anthropolofly, U.C. Berkeley,Berkeley, CA 1954 ThePrehistoric Cultures of theHorn of AfTica. Cambridge: 94720. CambridgeUniversity Press. is an assistantprofessor in the Departmentof Gen Suwa 1987 "Transitions:Homo erectus and the Acheulian:The Ethi- Anthropolofly,University of Tokyo.MailinB address:Depart- opiansites of Gadeband the MiddleAwash," Journal of ment ofAnthropolofly, Faculty of Science,University of Tokyo, HumanEvolution 16: 809-826. HonBo,Bunkyo-ku, Tokyo 113, Japan. 1988 "TheMiddle Stone Age of East Africaand the Begin- SileshiSemaw is agraduate student in the Departmentof ningsof RegionalIdentity," Journal of WorldPrehistory AnthropoloflyatRutBers University.MailinB address:Depart- 2: 235-305. ment ofAnthropolofly, RutBers University, New Brunswick, Clark,J. Desmond,and Hiro Kurashina w 08903. 1979 "HominidOccupation of East-centralHighlands of ConasBeyene is a paleolithicarchaeoloyist at the Minishy of Ethiopiain the Plio-Pleistocene,"Nature 282: 33-39. Culture and SportsAJfairs,PO Booc5717, AddisAbaba, Ethiopia. Clark,J. Desmond,and Kenneth Williamson 1984 "AMiddle Stone Age OccupationSite at PorcEpic cave, BerhaneAsfawis a paleoanthropoloyistin the Paleoanthro- DireDawa (East-central Ethiopia)," TheAfticanArchae- poloflyLaboratoty, CRCCH, Ministty of Culture and Sports olo,gicalReview 2: 37-71. AJfairs,P.O. Booc5717, AddisAbaba, Ethiopia. BerhaneAsfaw, Getaneh Assefa, John W. K. Geochron- Clark,J. Desmond, R. C. Walter is ageochronoloyistat the Berkeley Harris,Hiro Kurashina,Robert C. Walter,Tim D. White,and oloCgyCenter. MailinB address:Berkeley Geochronolofly Center, MartinA. Williams Institute of Human OrzBins,2453 RidBeRoad, Berkeley,C:A 1984 "PaleoanthropologicalDiscoveries in the MiddleAwash 94709. Valley,Ethiopia," Nature 307: 423-428. de Heinzelin,Jean, ed. 1983 TheOmo Group. Strat,igraphic and Related Earth Sciences Studiesin theLower Otno Bo/sin, Southern Ethiopia (An- Aronson,James, Robert Walter, and Maurice Taieb nales,Ser. 8, no. 85). Tervuren:Musee Royal de l'Afrique 1983 "Correlationof Tulu Bor Tuff at Koobi Forawith the Centrale. SidiHakoma Tuff at Hadar,"Nature 267: 323-327. Delson,Eric Asfaw,Berhane, Yonas Beyene, Sileshi Semaw, Gen Suwa,Tim 1984 "CercopithecidBiochronology of the AfricanPlio-Pleis- White,and Giday WoldeGabriel tocene:Correlation Among Eastern and Southern Hom- 1991 "Fejej:A New PaleoanthropologicalResearch Area in inid-bearingLocalities," in PeterAndrews and Jens Fran- Ethiopia,"JournalofHumanEvolution21: 137-143. zen,eds., TheEarlyEvolution ofMan, with Special Empho/sis on SoutheastAsia and Afiica. CourzerFonchun,gsinstitxt Asfaw,Berhane, Cynthia Ebinger, David Harding, Tim D. White, Senckenber,g69: 199-218. andGiday WoldeGabriel 1990 "SpaceBased Imagery in PaleoanthropologicalResearch: Di Paola,G. Mario An EthiopianExample," National Geo,graphic Research 6: 1972 "TheEthiopian Rift Valley(Between 7° and 8°40' Lat. 418-434. North),"Bulletin Vokanolo,gique 36: 517-560. 492 Kesem-KebenaPaleoanthropoloyical EthiopialWoldeGabriel et al.
EIGS(Ethiopian Institute of GeologicalSurveys) Kalb,John E., ElizabethB. Oswald,Assefa Mebrate, Sileshi Te- 1989 "GeneralizedGeologic Map of the NorthernEthiopian bedge,and Clifford Jolly Rift Valley,"scale 1:500,000, Addis Ababa. 1982c "Stratigraphyof the AwashGroup, Middle Awash Val- ley, Afar,Ethiopia," Netvsletter of Strati,graphy.11: 95- Harris,John M., andTim D. White 127. 1979 "Evolutionof the Plio-PleistoceneAfrican Suidae," TransactionsoftheAmerican Philosophical Society 69(2): 1- Kazmin,Vladmir 128. 1975 "TheGeological Map of Ethiopia,"Geological Survey of Ethiopia. Howell,F. Clark 1978 "Hominidae,"inVincentJ.MaglioandH.B.S.Cooke, Kazmin,Vladmir, and Seife Michael Berhe eds.,Evolution of Afican Mammals.Cambridge: Harvard 1978 "Geologyand Developmentof the NazretArea," Ethio- UniversityPress, 154-248. pian Geolo,gicalStfrvey Note 100: 1-26. Howell,F. Clark,and Yves Coppens Leakey,Meave G., andRichard E. Leakey,eds. 1974 "Inventoryof Remainsof Hominidaefrom Pliocene/ 1978 KoobiFora Research Project, Volume 1. Oxford:Claren- PleistoceneFormations of the LowerOmo Basin,Ethio- don. pia (1967-1972),"Amerzcan Journal of PhysicalAnthro- Mohr,Paul polo,gy40: 1-16. 1967 "TheEthiopian Rift System,"Btflletin of the Geophysical Observatory,Addis Ababa 11: 1-65. 1983 "TheOmo Group.Stratigraphic and Related Earth Sci- encesStudies in the LowerOmo Basin,Southern Ethio- 1983 "TheMorton-Black Hypothesis for the Thinning of Con- pia,"in Jeande Heinzelin,ed., TheOmo Group. Strati- tinentalCrust Revisitedin WesternAfar," Tectonophysics ,graphicand RelatedEarth Sciences Studies in the Lower 94: 509-528. OmoBasin, Southerx Ethiopia. (Annales, Ser. 8, no. 85). Tervuren:Musee Royal de l'AfriqueCentrale, 1-5. Morton,William H. andRussell Black 1975 "CrustalAttenuation in Afar,"in A. Pilgerand A. Rosler, Howell,F. Clark,Paul Haesaerts, and Jean de Heinzelin eds., Afar Depressionof Ethtopia.Stuttgart: E. Schwei- 1987 "DepositionalEnvironments, Archaeological Occur- zerbartscheVerlagsbuchhandlung, 55-61. rencesand Hominidsfrom Members E andF of Shun- guraFormation (Omo Basin,Ethiopia)," Journal of Hu- Morton,William H., JohnG. Mitchell,David C. Rex, andPaul manEvolution 16: 643-664. Mohr 1979 "RiftwardYounging of VolcanicUnits in the Addis Johanson,Donald C., MauriceTaieb, and Yves Coppens AbabaRegion, Ethiopian Rift Valley," Nattfre 280: 284- 1982 "PlioceneHominids from the HadarFormation, Ethio- 288. pia (1973-1977): Stratigraphic,Chronologic, and Pa- leoenvironmentalContexts, with Notes on Hominid Pickford,Martin Morphologyand Systematics," Amertcan Journal of Phys- 1986 "CainozoicPaleontological Sites of WesternKenya," cal Anthropolo,gy57: 373-402. MtinchnerGeownssenschaftliche Abhandltfn,gen, Reihe A. Geolo,gietfnd Palaeontolo,gie 8: 1-151. Jones,Peter R. 1979 "Effectsof Raw Materials Potts,Rick on BifaceManufacture," Sci- 1989 ence204: 835-836. "Olorgesailie:New Excavationsand Findingsin Early and MiddlePleistocene Contexts, Southern Kenya Rift Jones,Peter W. Valley,"Joxrnal of Hxman Evolxtion 18: 477-484. 1976 "Ageof the LowerFlood Basaltsof the EthiopianPla- Rex,D.C., I.L. Gibson,and F. Dakin teau,"Nature 261: 567-569. 1971 "Ageof the EthiopianFlood Basalt Succession," Natxre PhysicalScience 230: 131-132. Justin-Visentin,Ester, M. Nicolette,L. Tolomeo,and Bruno Za- nettin Rightlllire,G. Phillip 1974 "Mioceneand PlioceneVolcanics of the AddisAbaba- 1985 "TheTempo of Changein the Evolutionof Mid-Pleis- DebraBerhan Area: Geopetrographic and Radiometric toceneHomo," in EricDelson, ed., Ancestors,The Hard Study,"Bulletin Volcanolo,gique 38: 237-253. Eridence.New York:Alan R. Liss,255-264. Kalb,John E., ElizabethB. Oswald,Sileshi Tebedge, Assefa Me- Sarna-Wojcicki,Andre M., CharlesE. Meyer,Peter H. Roth, and brate,Emmanuel Tola, and Dennis Peak FrancisH. Brown 1982a "Geologyand Stratigraphy of Neogenedeposits, Middle 1985 "Agesof TuffBeds at EastAfrican Early Hominid Sites AwashValley, Ethiopia," Nature 298: 17-25. andSediments in the Gulfof Aden,"Natxre 313: 306- 309. Kalb,John E., CliffordJolly, Assefa Mebrate, Sileshi Tebedge, CharlesSmart, Elizabeth B. Oswald,Douglas Cramer, Paul Sickenberg,O., andM. Schonfeld Whitehead,Craig B. Wood,Glenn C. Conroy,Tsrha Adefris, 1975 "The ChororaFormation Lower Pliocene Limnical LouiseSperling, and B. Kana Sedimentsin the SouthernAfar (Ethiopia)," in A. Pilger 1982b "FossilMammals and Artefacts from the MiddleAwash andA. Rosler,eds.,AfarDepresszon ofEthiopia. Stuttgart: Valley,Ethiopia," Nature 298: 25-29. E. SchweizerbartscheVerlagsbuchhandlung, 277-284. J1 rTw-1 1 A 1 1 1 r t lN l^nN j_ o2Ma oJrtelaArcUaeoloM7lZOl. 1y) 1yMY *^;w5
Stein,Julie K. WoldeGabriel,Giday, Tesfaye Yemane, Gen Suwa,Tim D. White, 1987 "Depositsfor Archaeologists,"Advances in Archaeoloyical andBerhane Asfaw Methodand Theory11: 337-395. 1991 "Ageof Volcanismand Fossils in the Buri-SoyomaAreaX AmaroHorst, SouthernMain EthiopianRift," Journal Taieb,Maurice tfAfricanEarth Science 13: 437-447. 1974 "EvolutionQuaternaire du Bassinde l'Awash,"unpub- lishedPh.D. Thesis,University of ParisVI. Wolpoff,Milford H. 1984 "Evolutionin Homoerectus: The Questionof Stasis," Teilhardde Chardin,P. Paleobiology10: 389-406. 1930 "LePaleolithique en Somaliefranc,aise et en Abyssinie," L'Anthropoloyie40: 331-334. Zanettin,Brano, A. Gregnanin,Ester Justin-Visentin, M. Nico- letti, C. Petrucciani,E.M. Piccirillo,and L. Tolomeo Tiercelin,Jean-J., Jacques Michavrx, and Yves Bandet 1974 "Migrationof the Oligocene-MioceneIgnimbritic Vol- 1979 "LeMiocene superieur du Sudde la Depressionde lAfar, canismin the CentralEthiopian Plateau," Neues Jahruch Ethiopie:Sediments, Faunes, Ages Isotopiques," Bulletin fiErGeoloyie und Palawntoloyie 9: 567-574. SocieteGeoloyique FranJcais 7: 255-258. Valentine,Greg 1983 "Proceduresfor Analysisof SilicateRocks and Minerals at Los AlamosNational Laboratory by X-rayFluores- cence,"Los Alamos National Laboratory Report LA-9663- MS. Vallois,Henri V. 1951 "LaMandibule Humaine Fossile de la Grottedu Porc Epic,pres Dire Daoua(Abyssinie)," L'Anthropoloyie 55: 231-238. Warren,Richard G., FrankM. ByersJr., and F. A. Caporuscio 1984 "Petrographyand MineralChemistry of Units of the TopopahSpring, Calico Hills, CraterFlat Tuffs and OlderVolcanic Units, with Emphasison Samplesfrom Drill Hole USW G-1, YuccaMountain, Nevada Test Site,"Los Alamos National Laboratory Report LA- 10003- MS. Wendorf,Fred, and R. Schild 1974 A MidXleStone Age Sequencefiwnthe Central RifE KalleyJ, Ethiopia.Warsaw: Polka Akademia Nauk InstytytHis- toriiKultury Materialnej. White,Tim D. 1986 "Cutmarkson the Bodo Cranium:A Caseof Prehistoric Defleshing,"AmencanJoumalofPhysicalAnthriolofly 69: 503-509. Williams,Martin A., FrancisWilliams, Franc,ois Gasse, Garniss H. Curiis,and Don Adamson 1979 "Plio-PleistoceneEnvironments at Gadeb Prehistoric Site,Ethiopia," Nature 282- 29-33. Williams,Mariin A. GetanehAssefa, and Don A. Adamson 1986 "DepositionalContext of Plio-PleistoceneHortiinid- bearingFormations in the MiddleAwash Valley, South- ernAfar Rift, Ethiopia," in L.E. Frostick,R.W. Renaut, J. Reid, andJ.J. Tiercelin, eds., Sedimentatonin theAf- ncanRifats. Palo Alto: BlackwellScientific, 241-251. WoldeGabriel,Giday 1987 "VolcanotectonicHistory of the CentralSector of the Main EthiopianRift," unpublished Ph.D. thesis,Case WesternReserve University, Cleveland, Ohio. WoldeGabriel,Giday, James Aronson, and Robert C. Walter 1990 "Geology,Geochronology, and Rift BasinDevelopment in the CentralSector of the MainEthiopian Rift,' Geo- loyicalSociety of Amer Bulletin102: 439458.