L'industrie lithique préoldowayenne du site de Fejej FJ-1

(To cite this pre-press online article, please include the standard reference information, the digital object identiier and online publication date) The early OldOwan STOne-TOOl aSSemblage frOm fejej fj-1a, eThiOpia Deborah Barsky, Cécile Chapon-Sao, Jean-Jacques Bahain, Yonas Beyene, Dominique Cauche, Vincenzo Celiberti, Emmanuel Desclaux, Henry de Lumley, Marie-Antoinette de Lumley, François Marchal, Pierre-Elie Moullé & David Pleurdeau Abstract Résumé Located in the Omo-Turkana basin at the northern limit of the Koobi Fora sedimentary Formation, the Fejej region has recently proven to be a rich study area for understanding early hominin behaviour and paleoenvironmental conditions. Among the rich fossiliferous and stone artefact localities discovered so far at Fejej, the FJ-1a archeological site has yielded a faunal and lithic assemblage in primary context. The archeological level is situated within a 15 meter luvial sequence beneath a volcanic tuff. Geochronological data from the FJ-1 sequence indicate an age of nearly 1,9 Ma for the FJ-1a artefact level. The stone industry was knapped from locally available raw materials (mainly quartz and basalt) and rocks had been carefully selected according to speciic petrographical and formal criterion. Hominins mastered several distinct stone knapping methods and used more or less exhaustive reduction sequences in order to produce small lakes. The different techniques used for stone reduction are deined in this paper thanks to a series of reits of lakes onto cores. Along with the reits, an in-depth analysis of the lakes, cores and worked pebbles provides an overview of the technological capacities of hominins present at the site nearly 2 million years ago. After the Fejej FJ-1a site was abandoned the archeological materials were rapidly buried, leaving an almost undisturbed archeological level. This site appears to represent a short episode of hominin occupation. La région de Fejej est située dans le bassin Omo-Turkana, à la limite septentrionale de la Formation sédimentaire de Koobi Fora. Récemment, cette région s’est avérée être importante pour la compréhension du comportement des homininés et des conditions paléoenvironnementales. Parmi les localités découvertes à Fejej, riches en fossiles et en industries lithiques, le site archéologique de FJ-1a a livré un assemblage de faunes et d’industries dans un contexte primaire. Le niveau archéologique est renfermé dans une séquence sédimentaire luviatile sous un tuf volcanique. Les données géochronologiques de la séquence de FJ-1 indiquent un âge d’environ 1,9 Ma pour le niveau archéologique FJ-1a. Sur le site, des matières premières locales (essentiellement quartz et basalte) ont servi à tailler les industries lithiques. Les homininés ont sélectionné les roches selon des critères pétrographiques et géométriques spéciiques. Ils maîtrisaient plusieurs méthodes de débitage différentes et utilisaient des séquences de réduction des nucléus plus ou moins exhaustives, ain de produire des éclats de petites dimensions. Cet article décrit les différentes techniques de débitage des roches, par la description d’une série de remontages d’éclats sur les nucléus. Ensuite, une analyse approfondie des éclats, des nucléus et des galets taillés permet de comprendre les capacités technologiques des homininés présents sur le site de FJ-1a il y a près de 2 Ma. Après l’abandon du site par les homininés, les ossements et les industries lithiques furent rapidement enfouis, laissant un niveau archéologique en place, indicatif d’un épisode d’occupation de courte durée. Keywords: Plio-Pleistocene, reits, stone tool assemblage, Ethiopia, Mode 1, Oldowan, variability, technology, knapping Deborah Barsky (corresponding author) 8 [email protected] ; [email protected] * Area de Prehistoria, Universitat Rovira i Virgili (URV), Avinguda de Catalunya 35, 43002 Tarragona, Spain and IPHES, Institut Català de Paleoecologia Humana i Evolució Social, C/Escorxador s/n, 43003 Tarragona, Spain * Cécile Chapon-Sao / Jean-Jacques Bahain / David Pleurdeau Département de Préhistoire du Muséum national d’histoire naturelle, UMR 7194 du CNRS, 1 rue René Panhard, 75013 Paris, France Yonas Beyene * Academic and Research Vice President, University of Wolkite, SNNPR, Ethiopia Dominique Cauche / Emmanuel Desclaux * Laboratoire départemental de préhistoire du Lazaret, 33 bis bd. Franck Pilatte, 06300 Nice, France Vincenzo Celiberti * * Centre européen de recherches préhistoriques de Tautavel, Ave. Léon Grégory, 66720 Tautavel, France Henry de Lumley / Marie-Antoinette de Lumley Institut de Paléontologie Humaine, Fondation Albert 1er Prince de Monaco, 1 rue René Panhard, 75013 Paris, France François Marchal * Pierre-Elie Moullé * UMR 6578 - Unité d’Anthropologie bioculturelle CNRS/Université de la Méditerranée/EFS, Faculté de Médecine, Sect. Nord, Univ. de la Méditerranée, CS80011, blvd. Pierre Dramard, 13344 Marseille Cedex 15, France Musée de préhistoire régionale de Menton, rue Lorédan Larchey, 06500 Menton, France DOI 10.3213/2191-5784-10196 Published online November 2, 2011 © Africa Magna Verlag, Frankfurt M. Journal of African Archaeology Vol. 9 (2), 2011, pp. 207–224 207 D. Barsky et al. Discovery and general context of the Fejej FJ-1 site The Fejej region is located in the Ethiopian sector of the African Rift system in south-western Ethiopia, only 10 kilometres north of the border between Kenya and Ethiopia (aSfaw et al. 1991; lumley & beyene 2004) (Fig. 1). The Fejej fossiliferous sediments lie at the northernmost extremity of the Koobi Fora Formation on the east side of Lake Turkana. This Formation is part of the Plio-Pleistocene Omo Group, which also includes the Shungura, Mursi and Usno Formations in the Lower Omo Valley, and the Nachukui Formation west of Lake Turkana (heinzelin 1983; brOwn & feibel 1986; harriS et al. 1988). These sedimentary deposits include successive volcanic layers, basalt lows or tephras, which have allowed for the elaboration of a precise geochronological framework based on K-Ar and Ar-Ar dating and tephrochonological data. The Omo Group Formations have registered changing environmental conditions over a period of 4.5 million years. Paleogeographical reconstructions based on the nature of the deposits (deltaic, luvial, lacustrine) have revealed that two major hydrographical systems succeeded one another in the Omo-Turkana basin: 1) an episodic lake system and 2) a large axial luvial system named Turkana River or Paleo-Omo River, that lowed towards the Indian Ocean through the Anza Rift (harriS et al. 1988; brOwn & feibel 1991; Cerling 1994; rOgerS et al. 1994; feibel 1997). The Fejej region remained unknown in the extensively studied Omo-Turkana Basin until 1989, when the irst survey campaigns were organized (Ministry of Ethiopian Culture Paleoanthropological Inventory Team; Berkeley University, California; New York University, NY) (aSfaw et al. 1991; fleagle et al. 1992). Further surveying brought to light more than 50 paleontological and/or archeological localities, dating from the Oligocene to the Late Pleistocene (lumley & beyene 2004). The FJ-1 locality is a mesa Formation capped by a layer of volcanic ash and covering a surface area of about 450 x 250 m (aSfaw et al. 1991; lumley & beyene 2004). Numerous bone remains (including hominins) and a rich stone industry were discovered around the mesa. Given the large extension of the site, the area was subdivided into sectors (indexed FJ-1a through FJ-1k). Sector FJ-1a was selected for excavation since an especially large concentration of worked pebbles and lakes was discovered there. Systematic excavations were initially undertaken over a surface area of 9 m² (December 1992;– January 1993) and then further extended to 35 m² (May – June 1997) and inally to 80 m² (December 1998;– January 1999). The archeological material (~4 000 artefacts) is stored at the National Museum of Ethiopia, Addis Ababa, and a complete interdisciplinary study of the site was published in 2004 (lumley & beyene 2004). Fig. 1. Location of the Fejej region (a) and other Plio-Pleistocene sedimentary Formations of the Omo-Turkana basin (b, after feibel 1993). 208 Journal of African Archaeology Vol. 9 (2), 2011 The Early Oldowan Stone-Tool Assemblage from Fejej FJ-1a Stratigraphy and sedimentology of the FJ-1a site at Fejej The FJ-1 depositional sequence is described from three isolated outcrops, stratigraphically correlated to compose a single sequence measuring 15 m. The entire sequence is divided into ive major units (Fig. 2). - At the base of the sequence, Unit 1 shows lateral variation with two facies (ChapOn 2007): the first is composed of silt rich in sands and the second of sands with some silt. Sediments are poorly sorted and correspond to high-energy fluvial deposits. - Unit 2 is a polymictic conglomerate with granitic sands. Channelling igures were observed at the base of the Unit as well as oblique and crossbedded stratiication indicating a luvial origin. Metamorphic heavy minerals from Units 1 and 2 show a low degree of polish, indicating shortdistance transport. The Fejej region is dominated to the east by the metamorphic Hamar mountain range, culminating at 2000 m and a basaltic plateau gently sloping towards Lake Turkana. The immaturity of the sediment and the type of pebbles in the conglomerate (quartz, basalt, gneiss, pegmatite and amphibolite), indicate that this Unit was deposited by temporary rivers lowing from the northeastern Hamar mountain range towards the southwest into the Omo-Turkana basin. - Unit 3 is composed of pale yellow silty sands with two layers of calcrete. The Unit shows a bimodal grain-size distribution with poorly classed, mature sediment of luvial origin. - Unit 4 is a light grey volcanic tuff with horizontal lamina and ripple-mark structures. The tuff is mainly composed of glass shards < 2 mm long and includes tubular carbonated concretions and rhizomes. The presence of angular quartz and feldspar grains indicates that the volcanic glass is contaminated by detritus. This Unit corresponds to a pyroclastic deposit in a calm aquatic environment. - Unit 5 is made up of light brown, silty sands with carbonated concretions. Partially dismantled by erosion, it has not yet been the object of a detailed sedimentary analysis. A trench cut into the sedimentary complex of the FJ-1 mesa revealed that archeological level “C1” is located in the basal part of Unit 3, about 3 m below the tephra layer (Unit 4). Fig. 2. Schematic representation of the lithostratigraphy of the FJ-1 locality deposits (lumley et al. 2004a). Geochronology of the FJ-1a site Geochronological data from the FJ-1 sequence is provided by magnetostratigraphy, tephrochronology and ESR dating. Magnetostratigraphic analyses show reverse polarity in Unit 1 and normal polarity in Units 3 and 4 (ChapOn 2007). The Fejej FJ-1 Tuff (Unit 4) was chemically analyzed by different methods and there has been debate concerning its correlation with other tephra from the Shungura or Koobi Fora Formations (aSfaw et al. 1991; haileab & feibel 1993; feibel 1999). New results from more recent surveys in the Fejej region suggest that the FJ-1 Tuff may be correlated to the Borana Tuff (Koobi Fora Formation) which, in turn, is correlated with an unnamed tuff from the Upper G Member of the Shungura Formation (ChapOn 2007; ChapOn et al. 2008; ChapOn et al. 2011). The age of the archeological layer is therefore between 1.95 ± 0.03 Ma, onset of the Olduvai Sub-Chron, and 1.869 ±0.021 Ma, 39Ar–40Ar age of KBS Tuff (Cande & KenT 1995; mCdOugall & brOwn 2006). This proposed age is supported by two Electron Spin Resonance (ESR) dates of 2.40 ±0.54 Ma and 1.96 ±0.32 Ma, obtained from quartz grains sampled from Units 1 and 3 respectively (lumley et al. 2004a). Journal of African Archaeology Vol. 9 (2), 2011 209 D. Barsky et al. Bioindicators the Fejej FJ-1 assemblage is therefore biostratigraphically dated to between 2.33 and 1.78 Ma (eChaSSOux et al. 2004). Biochronology Faunal remains were systematically collected from stratigraphical Units 1, 3 (archeological layer C1) and 5 (Tab. 1). Biostratigraphical data suggests relatively homogeneous faunal associations in all three Units corresponding to the Upper G and H Members of the Omo Shungura Formation and to the Upper Burgi and KBS Members of the Koobi Fora Formation. The age of Fejej FJ-1 Primata Theropithecus sp. Paracolobus sp. Xenocyon africanus Canis sp. Hyaenidae Herpestes (Galerella) sp. Felidae Deinotherium bozasi Elephas recki ssp. Elephas recki cf. atavus Ceratotherium simum Diceros bicornis Equidae Equus sp. Hexaprotodon aethiopicus Notochoerus scotti Metridiochoerus andrewsi Metridiochoerus modestus Kolpochoerus limnetes Giraffa pygmaea Pelorovis sp. Alcelaphini Aepyceros shungurae Tragelaphus nakuae Hippotragini Reduncini Antilopini Orycteropus sp. cf. Coleura afra Lepus capensis Hystrix sp. Arvicanthis gr. niloticus/primaevus Heterocephalus cf. atikoi Chelonia indet. Crocodilus sp. Anura indet. Sauria indet. Colubridae indet. cf. Claria sp. Siluriforma indet. Fish indet. Unit 1 X X Unit 3 X X X X X Unit 5 X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X Tab. 1. Synthetic table showing presence/absence of different vertebrate taxons in each Stratigraphical Unit of the Fejej FJ-1 site (eChaSSOux et al. 2004). 210 Although Unit 1 is stratigraphically earlier than the other fossil bearing Units, the relatively poor preservation of the fossils and the low frequency of remains for each species make it dificult to precisely evaluate the biostratigraphical position of this level. In addition, species’ presence/absence in Unit 1 is not necessarily signiicant (for example, the absence of Metridiochoerus modestus and Equus, Tab. 1). However, the presence of Elephas recki cf. atavus indicates that this level cannot be older than the Omo Shungura Formation’s Upper Member F (2.33 Ma). The presence of the genus Equus in Units 3 and 5 suggests that the lower chronological limit for these levels may be situated at 2.32 Ma since this species irst appears in Member G of the Omo Shungura Formation. According to the appearance/disappearance grid for suids (whiTe 1995), the presence in Unit 3 of Metridiochoerus modestus and Notochoerus scotti situate this level between 1.89 and 1.80 Ma. The micromammals from archeological level C1, including Arvicanthis morphotype niloticus/primaevus, Heterocephalus cf. atikoi, Lepus capensis and cf. Coleura afra indicate an open environment and a semiarid climate. This faunal association is comparable to the upper levels of the Shungura Formation in the lower Omo Valley (Members E, F and lower G; weSSelman 1984) and also to the Koobi Fora Formations in East Turkana (blaCK & KriShTalKa 1986). The fauna Fejej FJ-1 can be placed within the paleoclimatic and paleoenvironmental evolution described for the Omo-Turkana Basin for the last 4 million years and may be assigned to the arid event recognised in this area between 2.3 and 2 Ma. Paleoecological and taphonomical features of the fauna Paleoecological conditions in the Fejej FJ-1 region at the time of the hominin occupation of the FJ-1a site have been assessed from palynological data and faunal associations, both of which translate an open, grassy landscape with wooded areas (eChaSSOux et al. 2004; umer et al. 2004). The pollen analysis indicates an evolution towards increased humidity that probably favoured the development of a mosaic, riverside landscape with forested mountain ranges. The presence of Crocodylus and ish remains conirm that there was a nearby water source, as does that of the grass-eating antelope Reduncini (Kob) which is generally found in wet areas such as loodplains. In coherence with the pollen analysis, species’ diversity indicates a variety of ecological settings around the site. Journal of African Archaeology Vol. 9 (2), 2011 The Early Oldowan Stone-Tool Assemblage from Fejej FJ-1a Although surface finds tend to be altered and strongly mineralized, bones from archeological level C1 are well preserved. Surface remains from Fejej FJ-1 had undergone signiicant taphonomical and secondary biological modiications. In contrast, fossil bones recorded from archeological level C1 do not show traces of secondary modiication. Fossils from the archeological level do not appear to have been rolled nor dispersed and their burial in situ by luvial sediments was probably rapid. Unlike the surface inds, no carnivore remains were found in level C1 and bones show no traces of carnivore or rodent activity. While surface inds relect a mixed ensemble, those from level C1 were often found in anatomical connexion, with reitting bone fragments discovered adjacent to one another, suggesting a highly localized activity area. Fossils are found associated with stone artefacts. Long bones from level C1 were systematically broken and fractured surfaces display smooth edges typical of intentional breakage, as well as impact scars, while no such traces are observed on the surface inds. Evidence of human activity, notably fractures on fresh bones, is frequently observed on shafts, bone splinters and determinable epiphysis. Such traces are observed on 35 % of the determinable bones and 41 % of the bone splinters. A signiicantly high proportion of Aepyceros shungurae bones (46 %) also show this kind of intentional breakage (see eChaSSOux et al. 2004 for a detailed taphonomical description). The Fejej FJ-1a site represents a seemingly undisturbed archeological level with knapped stone artefacts associated with a faunal assemblage presenting traces of human intervention. 100 91 % 90 80 70 60 50 40 39 % 30 % 30 20 Note that the locality FJ-1e (Unit 3), located about 100 m from FJ-1a, has yielded three hominin dental remains attributed to Homo aff. H. habilis (lumley & marChal 2004). The Fejej FJ-1a stone assemblage and reits The stone assemblage from Fejej FJ-1a (lumley et al. 2004b; barSKy et al. 2006) was knapped from local raw materials collected from the alluvial deposits of a small river near the site. Quartz pebbles are dominant (91 %) compared to basalt (7 %) and only a few artefacts were knapped from other types of rocks (2;%). By comparison, sampling in the FJ-1 conglomerate (Unit 2), located just below archaeological Unit 3 and probably deposited by the same luvial system, revealed that quartz and basalt pebbles each represent about 35 % of available rock types (Fig. 3). So, at Fejej FJ-1a, the quartz pebbles seem to have been preferentially used. This characteristic is not unique to the Fejej FJ-1a assemblage and is observed elsewhere in the Lower Omo Valley (Shungura Formation, Ethiopia) where quartz often exceeds 90 % of Oldowan assemblage composition even though it is not numerically more frequent than other petrographic groups in the pebble sources (delagneS et al. 2011). The overwhelming dominance of quartz in the archeological assemblage shows that Fejej hominins intentionally sought this rock type out for their tools. The quartz pebbles, although jointed, have few inclusions and their suitability for knapping is underlined by the scarcity of angular fragments (debris) relative to well struck lakes, as well as by the numerous refits of flakes onto cores (Figs. 4–10). However, almost half of the flakes do show transverse or longitudinal fractures which probably occurred during their extraction. The probability that knapping occurred on-site is supported by numerous reits of broken lakes: 10 Siret accidents, 5 transversal breaks, 1 ventral face fracture. Preservation in the silty sandy Fejej FJ-1a Unit 3 deposits was optimal and, apart from the 15 % 9% 7% 10 3% 1% 1% 4% 0 quartz basalte gneiss Fejej FJ-1 archeological level C1 pegmatites amphibolites other rocks Conglomerate (sedimentological Unit 2) Journal of African Archaeology Vol. 9 (2), 2011 Fig. 3. Frequencies of different rock types in the Fejej FJ-1a stone industry and in the nearby Unit 2 conglomerate (Unit 2 sample size=254 pebbles); the most likely raw material source. Note the overwhelming abundance of quartz in the FJ-1a assemblage. 211 D. Barsky et al. basalt industries whose surfaces are slightly altered, all elements of the assemblage are well preserved. Cutting edges are sharp and irregular retouch, perhaps resulting from use, is clearly visible on many pieces. Worked pebbles, lakes and angular fragments in quartz show no traces of rolling, lustre or alteration. The apparent relationship between pebble shape and subsequent pebble use suggests that hominins were capable of purposeful thought processes. Whole pebbles may have been brought to the site to serve as a raw material supply (manuports). Thick pebbles with oval sections were often employed as percussion instruments (61 % of hammerstones) while lat pebbles and pebbles with oval sections were shaped by one or a few blows into chopper-like instruments (56 % pebble tools). Flake production was often carried out on cubeshaped pebbles (50 % of the cores) that did not require initializing knapping stages for successful lake extraction. Unifacial and bifacial knapping methods were performed on lat pebbles using bipolar on an anvil or hand held percussion (Figs. 4–6). Orthogonal (Figs. 7 and 8) and multifacial reduction methods (Figs. 9 and 10) were carried out on thick, cube shaped pebbles. Pebble size also appears as a criterion for hominins when they selected rocks for speciic elements of their toolkit. The industry is quite small with pebble length rarely exceeding 10 cm: few cores bear witness to lengthy knapping episodes and whole and worked pebbles are close in size. The majority of the lakes measure from 2 to 3.5 cm (average non oriented lake length= 3.9 cm). Assemblage composition relects a primary occurrence. Reits between broken and knapped products confirm this hypothesis. Numerous knapped flakes and fragments were reitted onto cores and all stages of pebble reduction are represented. Small lakes and fragments (L= <2 cm) were collected by sieving (526 pieces). The FJ-1 assemblage is composed of 2610 pieces, of which 1114 come from archeological level C1 (Tab. 2). The industry is predominantly composed of lakes and angular fragments with quite a few cores and worked pebbles (Tab. 3). Only eleven pieces have edges that may have been shaped by intentional retouch but numerous others display irregular retouch that could result from use. Many whole or broken pebbles display percussion scars and/or accidentally detached lake negatives. These are often situated on pebble extremities or on the edges of their fractured surfaces. Some larger pebbles with percussion scars on their plane surfaces might have been used for anvils. Traces of percussion on worked pebbles and cores suggest their use as multi-purpose tools. Worked pebbles, almost always in quartz (80;%), show little formal standardisation and the distinction between cores and tools is unclear. Pebbles worked by a few unidirectional removals may indeed be called ‘choppers’ although it is generally agreed that, in early Mode 1 or Oldowan assemblages, lakes rather than intentionally shaped pebble tools were the knapper’s ultimate goal (TOTh 1985; SChiCK & TOTh 1993; rOChe et al. 2003). There are however some chopper-like pieces with apparent traces of crushing on worked edges that appear to be intentionally shaped tools. These are mainly “primary” choppers with only one removal (46 %) or choppers with a few adjacent removals; generally between 2 and 5. Sometimes the tools are situated on the lateral edges of a pebble but they are mostly on pebble extremities. Bifacial reduction strategies are rarely observed on either worked pebbles or cores and there are only four bifacially worked pebbles (chopping-tools). There are a few heavy duty scrapers (rostro-carénés) with convex edges “shaped” by abrupt, unidirectional removals and retouch. It is unclear whether or not the retouch on these pieces is the result of use wear or intentional shaping — hominins may simply have exploited the abrupt cutting edge of some unidirectional cores (Tab. 4). Cores were mostly knapped on quartz (86 %); less often on basalt or granite. Comparatively longer and more complex knapping episodes were carried out on basalt. A precise description of the laking techniques used at Fejej is provided by core and lake analysis, as well as by reits of lakes onto cores, revealing that hominins mastered a wide range of relatively standardized reduction modes (Tab. 5). Direct, hard hammer percussion is most frequently observed, although some of the cores were laked using controlled bipolar on an anvil percussion (pebble slicing, as at the Hadar sites of AL 666 and AL 894, 2.3 Ma, Kimbel et al. 1996; gOldman-neuman & hOverS 2009, as opposed Sectors of Fejej FJ-1 FJ-1a FJ-1b FJ-1c FJ-1e Total Surface inds Level CO (colluvium above archeological level) Level C1 (level in place) Total stone artefacts 508 853 1114 2475 1 114 20 1 114 20 643 853 1114 2610 212 Journal of African Archaeology Vol. 9 (2), 2011 Tab. 2. Number of stone artefacts from the different sectors of Fejej FJ-1 (lumley et al. 2004b). The Early Oldowan Stone-Tool Assemblage from Fejej FJ-1a Tool type quartz Whole pebbles and hammerstones Broken pebbles Worked pebbles Cores Flakes (> 2 cm) Small lakes (< 2 cm) Angular fragments Retouched pieces Total basalt granite sandstone-quartzite Total N % N % N % N % N % 100 21 68 78 1001 345 757 11 2381 54.6 47.8 80.0 84.7 95.5 97.5 95.5 100 91.2 63 17 13 10 42 8 32 185 34.4 38.6 15.3 10.9 4.0 2.2 4.0 7.1 14 3 1 3 2 1 4 28 7.7 6.8 1.2 3.3 0.2 0.3 0.5 1.1 6 3 3 1 3 16 3.3 6.8 3.5 1.1 0.3 0.60 183 44 85 92 1048 354 793 11 2610 100 100 100 100 100 100 100 100 100 Tab. 3. Rock type distribution according to the different tool types from Fejej FJ-1a (lumley et al. 2004b). Worked pebble type Surface single concave removal chopper double chopper heavy duty scraper (rostro-caréné) chopper with heavy duty scraper (rostro-caréné) chopping-tool chopping-tool with a chopper chopping-tool with heavy duty scraper (rostro-caréné) Total bed CO N % N 18 23 2 1 1 2 38.3 48.9 4.3 2.1 2.1 4.3 6 6 2 47 55.3 bed C1 % 1 1 16 18.8 Total N % N % 15 4 68.2 18.2 3 13.6 22 25.9 39 33 4 4 1 2 1 1 85 45.9 38.8 4.7 4.7 1.2 2.3 1.2 1.2 100 Tab. 4. Frequency of worked pebble types in the each bed level of the Fejej FJ-1 site (lumley et al. 2004b). Striking platform Technology % Unifacial N unipolar prismatic (unipolar) bipolar intersecting peripheral centripetal multipolar orthogonal sub-total unifacial cores Bifacial removal negative unipolar bipolar centripetal sub-total bifacial cores Multifacial cortex and/or cortex and/or breakage plane removal negative orthogonal surface 1 15 1 17 18.5 multiplatform - 10 - 10 10.9 sub-total multifacial cores 1 25 1 27 29.3 53 37 2 92 100 Total 29 1 6 4 9 1 50 1 1 2 4 1 1 30 1 8 4 11 1 55 32.6 1.1 8.7 4.3 12 1.1 59.8 1 1 2 3 3 2 8 - 3 4 3 10 3.3 4.3 3.3 10.9 Tab. 5. Technological distribution of cores from the Fejej FJ-1 site (lumley et al. 2004b). Journal of African Archaeology Vol. 9 (2), 2011 213 214 D. Barsky et al. Journal of African Archaeology Vol. 9 (2), 2011 Fig. 4. Reit of ive lakes onto a quartz core (58 x 57 x 45 mm). An initial series of lakes was obtained from the surface of an ancient fracture using bipolar laking on an anvil. Another striking platform, oriented orthogonally from the irst, was also used. Finally, a cortical striking platform was exploited to produce lakes from the pebble’s proile in a “slicing” manner, by applying controlled bipolar percussion (drawn by D. Cauche, V. Celiberti, M. Montesinos). The Early Oldowan Stone-Tool Assemblage from Fejej FJ-1a Fig. 5. Reit of two quartz lakes onto a core (82 x 61 x 40 mm). A series of unidirectional lakes was struck from the cortical surface of a lat, oval shaped pebble. The lattest side of the pebble served as a platform to extract some well struck lakes from its opposite, more convex surface. Flakes obtained show cortical butts and longitudinal negatives on their dorsal surfaces. The irst lake shows a Siret type fracture and the second, larger lake, was obtained by a blow in the same direction as the one preceding its extraction. This knapping sequence clearly demonstrates the unfolding of the recurrent unidirectional technique typical at Fejej FJ-1a. If pursued along the entire periphery of the pebble, this technique would have produced a unifacial discoidal core (Fig. 6). One of the lakes comes from the colluvium (drawn by D. Cauche, V. Celiberti, M. Montesinos). to simple quartz crushing; Fig. 4). Unifacial reduction dominates relative to bifacial or multifacial techniques (60 % of the cores). While unifacial cores most often show a single flaking direction (unipolar) some do display bipolar, intersecting, centripetal (peripheral) or orthogonally oriented removal negatives. Recurrent peripheral flaking produced unifacial discoid cores (Fig. 6). Some cores have two independantly knapped surfaces, with unipolar, bipolar or centripetally oriented removals. Sinuous core edges typical of systematic bipolar exploitation are scarce and we underline the absence of intentional bifacial exploitation. At Fejej FJ-1 hominins most often produced lakes from unprepared surfaces during brief, unidirectional knapping episodes. Rounded pebbles were sometimes sliced or split in order to create suitable knapping surfaces to extract lakes. Flake removal negatives occasionally served as platforms, producing orthogonal cores (CarbOnell et al. 1999). The presence of a few polyhedronshaped multiplatform cores shows that such “prepared” platforms were sometimes used during more lengthy knapping episodes consisting of frequent core rotations (Figs. 9 and 10). However, multiplatform flaking is poorly developed at FJ-1. We may suggest that a more developed multiplatform exploitation was later to lead to the production of spheroid-type tools that are not present in the FJ-1 assemblage (as at Olduvai Gorge, Lower and Middle Bed 1; leaKey 1971). Cores with multiplatformorthogonal removals and two, three or more laking generations are very rare at this site (globular or polyhedron shaped cores =10.9;%, Tab. 5). Journal of African Archaeology Vol. 9 (2), 2011 215 D. Barsky et al. Fig. 6. Reit of four quartz lakes onto a core (101 x 72 x 62 mm). A cortical striking platform was exploited with a recurrent gesture to extract several well struck lakes, four of which were recovered and reitted onto this unifacial discoidal core. All of these elements were found in the same square meter. The lakes all have cortical striking platforms and longitudinally oriented removal negatives on their dorsal surfaces. The reitted core surface shows at least ive more removal negatives for which the lakes were not recovered. The knapping episode was relatively long and attests to organized and systematic lake production (drawn by D. Cauche, V. Celiberti, M. Montesinos). Flake morphology relects a pattern of non-exhaustive unidirectional technology that parallels the patterns seen on the core forms (Figs. 11 and 12). Most of the lakes (90 %) conserve at least some cortex and their striking platforms are almost always cortical (77.5;%, Tab. 6). The overwhelming abundance of lakes with residual cortex and/or cortical butts underlines the use of unprepared surfaces and the overall simplicity of the stone knapping methods employed at FJ-1. While it is true that the abundance of residual cortex on lake butts and dorsal surfaces may also be attributed to the overall small size of the knapped pebbles (TOTh 1982, 1987; braun et al. 2005), non-cortical platforms would be present in higher proportions if orthogonal and multiplatform laking techniques were more frequently used. Thus, the massive representation of 216 Toth lake types I–III (lakes with cortical platforms) underlines the exploitation of unprepared surfaces and the dominance of unidirectional laking strategies at FJ-1 (Tabs. 5 and 6). Non-cortical lakes likely produced during more lengthy knapping episodes are scarce (9 %). In spite of the more extensive knapping observed from the basalt cores, corresponding lakes are relatively rare (3.8 basalt lakes/core, compared with 7.1 quartz lakes/ core and only 0.5 lakes/core for other rock types). However, basalt lakes without any cortex are proportionately more frequent than for quartz. On all of the lakes, removal negatives on dorsal surfaces are most often unidirectional and parallel to the knapping axis (60;%), in conformity with the rest of the assemblage. Journal of African Archaeology Vol. 9 (2), 2011 The Early Oldowan Stone-Tool Assemblage from Fejej FJ-1a Fig. 7. Reit of four basalt lakes onto a core (72 x 55 x 48 mm). This reit illustrates an orthogonal knapping strategy. The core presents three generations of orthogonally oriented removals struck from negatives of previous blows that were apparently used to initiate knapping episodes by providing appropriate lake extraction planes. Flakes show smooth striking platforms and longitudinally oriented removal negatives on their dorsal surfaces. While lat ovate pebbles were reduced using centripetal strategies, rounded or cube shaped supports were laked using this, more appropriate, orthogonal strategy (drawn by D. Cauche, V. Celiberti, M. Montesinos). Fig. 8. Reit of four basalt lakes onto a core demonstrating the orthogonal knapping strategy (72 x 55 x 48 mm). The removal negative of a large lake served as a striking platform for a series of recurrent lakes (photo D. Barsky). Flake type (after ToTh 1985) Type I: entirely cortical lakes % 7.2 lakes with Type II: cortical striking platform and residual cortex on the dorsal surface 41.9 cortical platforms: Type III: striking platform only with cortex 28.4 Type IV: non cortical striking platform and cortical dorsal surface Type V: non cortical striking platform and residual cortex on the dorsal surface Type VI: non cortical lake Total 77.5% 2.5 11.5 8.5 100% Tab. 6. Whole lake types from Fejej FJ-1 according to the position of their residual cortex (TOTh 1985). Note the overwhelming predominance of lakes with cortical striking platforms (77.5 %). Journal of African Archaeology Vol. 9 (2), 2011 217 218 Fig. 9. Reit of ive quartz lakes onto a core (65 x 60 x 54 mm). The technique observed from this core approaches the multiplatform knapping method which sometimes produces polyhedron or globular type cores: removal negatives served alternately as striking platforms for subsequent removals with a core rotation following each blow. Natural (cortical) platforms offered by this initially quadrangular pebble were also opportunistically exploited. At least three generations of removals may be observed from this piece and some of the lakes attest to the use of bipolar knapping on an anvil (drawn by D. Cauche, V. Celiberti, M. Montesinos). D. Barsky et al. Journal of African Archaeology Vol. 9 (2), 2011 Fig. 10. Reit of ive quartz lakes onto a core (65 x 60 x 54 mm). This core and corresponding lakes illustrate the multiplatform laking technique that is rarely observed at Fejej FJ-1a. Elsewhere, this technique was applied to knap quartz pebbles and may later have lead to the production of spheroids and bola type objects (photo D. Barsky). The Early Oldowan Stone-Tool Assemblage from Fejej FJ-1a Journal of African Archaeology Vol. 9 (2), 2011 Fig. 11. Quartz knapping products from Fejej FJ-1a level C1. 1–6: cortical lakes; 7–9: lakes with ample cortical surfaces (drawn by D. Cauche, V. Celiberti, M. Montesinos). 219 Fig. 12. Quartz knapping products from Fejej FJ-1a level C1. 1–6: angular fragments; 7: small lake with cortical butt and cortex on the dorsal surface; 8–9: lakes with cortical butts; 10 and 12: lakes with cortical butts and cortex on their dorsal surfaces; 11: lake with cortical butt, cortex on the dorsal surface and mixed opposite impact points; 13: lake with cortical butt, cortex on the dorsal surface and irregular retouch (drawn by D. Cauche, V. Celiberti, M. Montesinos). D. Barsky et al. Primary lakes (with cortical dorsal surfaces and butts) are well represented. Cortical lakes whose dorsal surfaces merge with cortical striking platforms (20 % of the cortical lakes) may have been accidentally produced from hammerstones. Fejej FJ-1a knapping technology and reits Sedimentary conditions, the fresh aspect of the industry and the presence of refits, indicate that the artefacts were rapidly buried and explain the optimal preservation at the site. Among the lithics, 39 broken lake segments were reitted (mainly Siret-type accidents or lakes with transversal fractures). There are also two series of lakes (lacking the core), each of which includes three reitting elements. Two angular quartz fragments were reitted, as were two broken basalt pebbles; one with two and the other with four conjoining elements. Two pieces of a quartz pebble with a percussion negative (hammerstone) and three fragments of a quartz anvil were also reitted. Two angular fragments were reitted onto broken pebbles with isolated removals (basalt, quartz). Of the 92 cores discovered, eleven have from one to ive reitting lakes. Study of the spatial distribution at FJ-1 (FJ-1a sector) suggests that level C1 represents a short term occupation (lumley et al. 2004c). Vertical projections reveal a single archeological level of varying artefact density whose average thickness reaches about 20 cm. Some fossils found in anatomical connection indicate that bones were still connected by ligaments or cartilage when they were buried and that they remained relatively undisturbed since that time. Bone fragments are usually closely associated with the stone industry and both of these show corresponding areas of density. Although no speciic activity areas were deined, two zones do show a greater concentration of archeological material (surface areas: 12 m2 and 24 m2). The smaller of the two, rich in stone artefacts and large bone fragments, has been interpreted as a place where carcasses were “prepared” by activities such as disarticulation, while the cutting of meat from the bones may have occurred a few meters away (lumley et al. 2004c). The close proximity of reitted stone artefacts and the completeness of the lithic sample (hammerstones, anvils, debitage) all indicate in situ knapping. While some of the reitted artefacts were found up to three metres apart or in the colluvium, most of them were discovered within a single square metre or in adjacent squares in archeological level C1. While lakes reitting onto cores were found within a restricted area (nearby or within the same square metre), the cores 220 themselves were often discovered a few meters away. Given the apparently undisturbed nature of the site, this artefact disposition may relect a given stone knapping behaviour nearly 2 million years ago. Although it is dificult to interpret this phenomenon, we might suggest that, following the in situ production of a few lakes, cores were transported over a short distance to be used for percussion activities or as cutting tools, before being discarded. This hypothesis is supported by the presence of percussion marks and/or edge damage on the cores themselves. In any case, among the activities that likely took place at Fejej FJ-1, hominins made stone tools, disarticulated carcasses, broke bones and probably fed on meat after cutting it off of the bones with small, sharp cutting instruments. Conclusions The Fejej FJ-1a stone assemblage provides an exceptional example of early stone technology in Africa. Different reduction modes are illustrated by numerous reits of lakes onto cores which clearly demonstrate that hominins simultaneously mastered several organized and relatively standardized stone reduction modes. In spite of the relatively large variety of rocks available from nearby alluvials, the overwhelming dominance of quartz in the assemblage suggests that hominins selected their raw materials with discernment. Also, pebbles naturally presenting angles favourable for flake detachment were preferentially chosen for knapping. This choice is relected by the strong dominance of cortical striking platforms on the lakes. This characteristic may indicate that Fejej FJ-1a hominins did not master platform preparation or it may relect a technological preference. Overall, knapping episodes were short and 90 % of the lakes maintain at least some cortex. The assemblage is dominated by small, non standardized lakes with some cores and worked pebbles. However, intentionally shaped, standardized tools are not a signiicant component of this industry. Unifacial technology largely dominates among the knapping methods used and few pieces display bifacially worked edges. Bifacial technology resulting in cores with extraction negatives originating from a sinuous equatorial edge (typical of later Mode 2 assemblages), are not represented at Fejej FJ-1a. Multifacial technology is incidental and there are very few polyhedron-shaped or globular cores. While many flakes and worked pebble edges show irregular retouch or crushed edges, intentionally retouched pieces are scarce and show no formal standardization. At Fejej FJ-1a, technological variability is expressed by the use of different modes of unidirectional laking (linear recurrent, recurrent Journal of African Archaeology Vol. 9 (2), 2011 The Early Oldowan Stone-Tool Assemblage from Fejej FJ-1a on an anvil, orthogonal and unifacial discoidal, bar2009; CarbOnell et al. 2009) and resulting core morphology depended largely on the pebble’s initial shape. Flake extractions were often achieved using the bipolar on an anvil technique and there are some truncated or sliced pebbles. SKy In spite of its seemingly developed variability, the Fejej FJ-1a assemblage shows afinities to some older African industries. Subtle differences have recently been described in the cultural complex referred to as “Oldowan” or “Mode 1” (delagneS & rOChe 2005; barSKy 2009) and new perspectives for the study of early flake-core assemblages have been proposed (CarbOnell et al. 2009). Indeed, given new discoveries, the “Oldowan” cultural complex is now overexpanded and includes not only the earliest African stone industries, such as those from Kada Gona or the eponymous Oldowan assemblages from Lower and Middle Bed 1 at Olduvai Gorge (leaKey 1971; de la TOrre & mOra 2005), but also Eurasian lake-core assemblages dating to as late as the Early Middle Pleistocene (the term “Oldowan” has been used to describe the industry from Isernia la Pineta in Italy, dating to ca 0.6 Ma; lumley et al. 2009). knapping, existed within the range of potential technological capacities, and that these may have been more or less fully developed at different sites depending on a variety of factors, including raw material quality or site function. Controlled pebble slicing using the bipolar on an anvil technique is present at FJ-1a as at other early Mode 1 sites, particularly those where quartz was exploited (2.3-2.4 Ma, Omo 71, 57, 123, FtJi 1, 2 and 5, Shungura Formation, Members E and F, ChavaillOn 1970, 1975, 1976; hOwell et al. 1987). This technique was also used to knap volcanic rocks at A. L. 666 and A. L. 894 (2.3 Ma, Hadar, Ethiopia; Kimbel et al. 1996; gOldman-neuman & hOverS 2009). Fejej hominins employed longer, slightly more complex direct knapping methods when knapping basalt, a iner raw material than quartz. At Fejej, as at other early African sites, hominins demonstrate selective behaviour by choosing core reduction schemes in accordance to a given raw material’s quality. Such discernment is also noted at Kada Gona and at the Lokalalei localities LA 2C and LA 1 (Kibunjia 1994; rOChe et al. 1999; rOChe 2000; delagneS & rOChe 2005; STOuT et al. 2005). The industries from Kada Gona localities EG10 and EG 12, OGS6 and OGS7 from the Hadar region of Ethiopia demonstrate that hominins possessed systematic and eficient stone reduction modes very early on (2.6 Ma, Semaw et al. 1997, 2009; Semaw 2000, 2005) and the FJ1-a assemblage relects how these methods were developed and adapted to other raw materials, notably quartz. Recent evidence from the FwJj20 site (1.95 Ma, Upper Burgi Member, Koobi Fora Formation, braun et al. 2010), with in situ artefacts and fossil bone fragments, some of which presenting clear striation marks apparently made by stone tools, provides evidence that diverse resources were made available to hominins practicing stone tool-mediated food accessing. Industries from the more recent sites of Kokiselei 5 and Naiyena Engol 1 (ca 1.6-1.8 Ma, Kibunjia et al. 1992 West Turkana, Kenya, rOChe et al. 1999, 2003) show slightly more advanced morphologies; elements such as multidirectional laking leading to polyhedron or globular shaped cores are more frequent at these sites than at Fejej FJ-1a and knapping sequences are generally longer. The production of large and small intentionally shaped tools (spheroids, scrapers) developed around this time as well, for example in Lower and Middle Bed 1 at Olduvai Gorge (1.8 Ma, Tanzania, leaKey 1971, TamraT et al. 1995; de la TOrre & mOra 2005), once again raising questions about the scope of “Oldowan” or Mode 1” assemblage variability. Like at Fejej FJ-1a, unidirectional, recurrent knapping methods and their variants dominate at Kada Gona EG 10 and EG 12 and at Lokalalei LA2C (Kibunjia 1994; rOChe et al. 1999; rOChe 2000; delagneS & rOChe 2005). The more complex bifacial orthogonal technique (involving striking platform preparation) has also been documented at these sites. Among the earliest sites at Gona, OGS-7 (Ounda Gona) has yielded a stone assemblage displaying relatively sophisticated and varied stone reduction strategies (dominant bifacial and multifacial knapping) and selectivity in raw material collection and use (Semaw et al. 2009). We may propose that some stone reduction methods, in this case bifacial or multifacial Acknowledgements We thank the French Ministry of Foreign Affairs, the Authority for Research and Conservation of Cultural Heritage, the National Museum of Ethiopia, the Ministry of Culture and Tourism and the South Nations, Nationalities and Peoples Region for permits and support; D. Gebre, M. Bekele, A. Dessie, A. Amzaye, T. Yifru, S. Bacha, B. Seyfu, T. Hagos, A. Sebbo, G. Kedir, M. Girma, Sissay, Godana, Amharu, Ubichat and S. Tornay, for ield assistance. We also thank the French Center for Ethiopian Studies, the administration of Omorate town, the members of South Omo Police forces and the Daasanach people. Journal of African Archaeology Vol. 9 (2), 2011 221 D. Barsky et al. References Asfaw, B., Beyene, Y., Semaw, S., Suwa, G., White, T.D. & WoldeGabriel, G. 1991. Fejej: a new paleoanthropological research area in Ethiopia. Journal of Human Evolution 21 (2), 137–143. Chapon, C. 2007. Cadre stratigraphique, sédimentologique, paléoclimatique et géochronologique du site à hominidés fossiles de Fejej FJ-1, sud-Omo, Ethiopie. Unpublished Ph.D. thesis, University of Perpignan, Perpignan. Barsky, D. 2009. An overview of some African and Eurasian Oldowan sites: evaluation of Hominin cognitive levels, technological advancement and adaptive skills. In: Hovers, E. & Braun, D.R. (eds.), Interdisciplinary Approaches to the Oldowan. Springer, Dordrecht, pp. 39–48. Chapon, C., Bahain, J.-J., Beyene, Y., Bilcot, J.-B., Guy, C., Jullien, M., Lumley, H. de & Provitina, O. 2008. Le tuf FJ-1 de Fejej (Sud-Omo, Ethiopie): Nouvelles corrélations téphrostratigraphiques sur la base de la géochimie avec les Formations plio-pléistocènes du bassin de l’Omo-Turkana. Quaternaire 19 (2), 141–146. Barsky, D., Cauche, D., Celiberti, V., Pleurdeau, D., Beyene, Y. & Lumley, H. de 2006. Les industries lithiques archaïques du site de Fejej en Ethiopie. In: XIVe Congrès de l’Union Internationale des Sciences Préhistoriques et Protohistoriques (Septembre 2001). Liège, pp. 1–9. Chapon, C., Bahain, J.-J., Beyene, Y., Bilcot, J.-B., Guy, C., Jullien, M., Lumely, H. de & Provitina, O. 2011. Geochemistry of the Fejej Tuffs (South-Omo, Ethiopia), their tephrostratigraphical correlation with Plio-Pleistocene formations in the Omo-Turkana Basin. C.-R. Palevol 10 (4), 251–258. Black, C.C. & Krishtalka, L. 1986. Rodents, bats and insectivores from the Plio-Pleistocene sediments to the east of Lake Turkana, Kenya. Contributions in Science 372, 1–15. Chavaillon, J. 1970. Découverte d´un niveau Oldowayen dans la basse vallée de l´Omo (Ethiopie). Bulletin de la Société Préhistorique Française 67 (1), 7–11. Braun, D.R., Harris, J.W.K., Levin, N.E., McCoy, J.T., Herries, A.I.R., Bamford, M.K., Bishop, L.C., Richmond, B.G. & Kibunjia, M. 2010. Early hominin diet included diverse terrestrial and aquatic animals 1.95 Ma in East Turkana, Kenya. Proceedings of the National Academy of Sciences 107 (22), 10002–10007. Chavaillon, J. 1975. Le site Paléolithique ancien d´Omo 84 (Ethiopie). Documents pour servir à l´Histoire des civilisations Ethiopiennes 6, 9–18. Braun, D.R., Tactikos, J.C., Ferraro, J.V. & Harris, J.W.K. 2005. Flake recovery rates and inferences of Oldowan hominin behavior: a response to Kimura 1999, 2002. Journal of Human Evolution 48 (5), 525–531. Chavaillon, J. 1976. Evidence for the technical practices of early Pleistocene hominids. In: Coppens, Y., Howell, F.C., Isaac, G.L. & Leakey, R.E. (eds.), Earliest Man and Environments in the Lake Rudolf Basin: Stratigraphy, Paleoecology, and Evolution. Chicago University Press, Chicago, pp. 565–573. Brown, F.H. & Feibel, C.S. 1986. Revision of lithostratigraphic nomenclature in the Koobi Fora region, Kenya. Journal of the Geological Society 143 (2), 297–310. Delagnes, A., Boisserie, J.R., Beyene, Y., Chuniaud, K., Guillemot, C. & Schuster, M. 2011. Archaeological investigations in the Lower Omo Valley (Shungura Formation, Ethiopia): new data and perspectives. Journal of Human Evolution 61 (2), 215–222. Brown, F.H. & Feibel, C.S. 1991. Stratigraphy, depositional environments, and palaeogeography of the Koobi Fora Formation. In: Harris, J.M. (ed.), Koobi Fora Research Project. The Fossil Ungulates: Geology, Fossil Artiodactyls, and Palaeoenvironments, Vol. 3. Clarendon Press, Oxford, pp. 1–30. Delagnes, A. & Roche, H. 2005. Late Pliocene hominid knapping skills: the case of Lokalalei 2C, West Turkana, Kenya. Journal of Human Evolution 48 (5), 435–472. Cande, S.C. & Kent, D.V. 1995. Revised calibration of the geomagnetic polarity timescale for the Late Cretaceous and Cenozoic. Journal of Geophysical Research, B, Solid Earth and Planets 100, 6093–6095. Carbonell, E., García Antón, M.D., Mallol, C., Mosquera, M., Ollé, A., Rodríguez, X.P., Sahnouni, M., Sala, R. & Vergès, J.M. 1999. The TD6 level lithic industry from Gran Dolina, Atapuerca (Burgos, Spain): production and use. Journal of Human Evolution 37 (3/4), 653–693. Carbonell, E., Sala, R., Barsky, D. & Celiberti, V. 2009. From homogeneity to multiplicity: a new approach to the study of archaic stone tools. In: Hovers, E. & Braun, D.R. (eds.), Interdisciplinary Approaches to the Oldowan. Springer, Dordrecht, pp. 25–38. Cerling, T.E. 1994. Lake Turkana and its precursors in the Turkana basin, East Africa (Kenya and Ethiopia). In: GierlowskiKordesch, E. & Kelts, K. (eds.), Global Geological Record of Lake Basins 1. Cambridge University Press, Cambridge, pp. 341–343. 222 de la Torre, I. & Mora, R. 2005. Technological Strategies in the Lower Pleistocene at Olduvai Beds I & II. ERAUL 112, Liège. Echassoux, A., Moullé, P.E., Desclaux, E. & Alemseged, Z. 2004. Les faunes plio-pléistocènes du site de Fejej FJ-1. In: Lumley, H. de & Beyene, Y. (eds.), Les sites préhistoriques de la région de Fejej, Sud-Omo, Ethiopie, dans leur contexte stratigraphique et paléontologique. Association pour la diffusion de la pensée française (ADPF). Editions Recherche sur les civilisations, Paris, pp. 203–340. Feibel, C.S. 1993. Freshwater stingrays from the Plio-Pleistocene of the Turkana Basin, Kenya and Ethiopia. Lethaia 26 (4), 359–366. Feibel, C.S. 1997. A terrestrial auxiliary stratotype point and section for the Plio-Pleistocene boundary in the Turkana Basin, East Africa. Quaternary International 40, 73–79. Feibel, C.S. 1999. Tephrostratigraphy and geological context in paleoanthropology. Evolutionary Anthropology: Issues, News, and Reviews 8 (3), 87–100. Journal of African Archaeology Vol. 9 (2), 2011 The Early Oldowan Stone-Tool Assemblage from Fejej FJ-1a Fleagle, J., Yirga, S., Bown, T.M. & Rasmussen, D.T. 1992. Fejej, Southern Ethiopia. Research and Exploration 8, 119–120. Goldman-Neuman, T. & Hovers, E. 2009. Methodological issues in the study of Oldowan raw material selectivity: insights from A. L. 894 (Hadar, Ethiopia). In: Hovers, E. & Braun, D.R. (eds.), Interdisciplinary Approaches to the Oldowan. Springer, Dordrecht, pp. 71–84. Haileab, B. & Feibel, C.S. 1993. Tephra from Fejej, Southern Ethiopia. Journal of Human Evolution 25 (6), 515–517. Harris, J.M., Brown, F.H. & Leakey, M.G. 1988. Stratigraphy and paleontology of Pliocene and Pleistocene localities west of Lake Turkana, Kenya. Contributions in Science 399, 1–128. Heinzelin, J. de (ed.) 1983. The Omo Group. Archives of the International Omo Research Expedition. Musée Royal de l’Afrique Centrale, Annales, Série 8, Sciences Géologiques 85, Tervuren. Howell, F.C., Haesaerts, P. & Heinzelin, J. de 1987. Depositional environments, archeological occurrences and hominids from Members E and F of the Shungura Formation (Omo Basin, Ethiopia). Journal of Human Evolution 16 (7–8), 665–700. Kibunjia, M. 1994. Pliocene archaeological occurrences in the Lake Turkana Basin. Journal of Human Evolution 27 (1–3), 159–171. Kibunjia, M., Roche, H., Brown, F.H. & Leakey, R.E. 1992. Pliocene and Pleistocene archeological sites west of Lake Turkana, Kenya. Journal of Human Evolution 23 (5), 431–438. Kimbel, W.H., Walter, R.C., Johanson, D.C., Reed, K.E., Aronson, J.L., Assefa, Z., Marean, C.W., Eck, G.G., Bobe, R., Hovers, E., Rak, Y., Vondra, C., Yemane, T., York, D., Chen, Y., Evensen, N.M. & Smith, P.E. 1996. Late Pliocene Homo and Oldowan tools from the Hadar Formation (Kada Hadar Member), Ethiopia. Journal of Human Evolution 31 (6), 549–561. Leakey, M.D. 1971. Olduvai Gorge: Excavations in Bed I and Bed II, 1960–1963. Vol. 3. Cambridge University Press, Cambridge. Lumley, H. de, Bahain, J.J., Beyene, Y., Chapon, C., Durandau, A., Laurent, M., Perrenoud, C. & Saos, T. 2004a. Etudes stratigraphique, sédimentologique, magnétostratigraphique et géochronologique du site de Fejej FJ-1. In: Lumley, H. de & Beyene, Y. (eds.), Les sites préhistoriques de la région de Fejej, Sud-Omo, Ethiopie, dans leur contexte stratigraphique et paléontologique. Association pour la diffusion de la pensée française (ADPF). Editions Recherche sur les civilisations, Paris, pp. 129–172. Lumley, H. de, Barsky, D. & Cauche, D. 2009. Les premières étapes de la colonisation de l’Europe et l’arrivée de l’Homme sur les rives de la Méditerranée. L’Anthropologie 113 (1), 1–46. Lumley, H. de & Beyene, Y. (eds.) 2004. Les sites préhistoriques de la région de Fejej, Sud-Omo, Ethiopie, dans leur contexte stratigraphique et paléontologique. Association pour la diffusion de la pensée française (ADPF). Editions Recherche sur les civilisations, Paris. Lumley, H. de, Beyene, Y., Barsky, D., Byrne, L., Camara, A., Cauche, D., Celiberti, V., Fournier, A. & Pleurdeau, D. 2004b. L’industrie lithique préoldowayenne du site de Fejej FJ-1, Ethiopie. In: Lumley, H. de & Beyene, Y. (eds.), Les sites préhistoriques de la région de Fejej, Sud-Omo, Ethiopie, dans leur contexte stratigraphique et paléontologique. Association pour la diffusion de la pensée française (ADPF). Editions Recherche sur les civilisations, Paris, pp. 391–564. Lumley, H. de, Beyene, Y., Bahain, J.J., Barsky, D., Cauche, D., Celiberti, V., Desclaux, E., Echassoux, A., Marchal, F., Moullé, P.E. & Pleurdeau, D. 2004c. Étude de la répartition spatiale du matériel archéologique du site de Fejej FJ-1, dans le secteur FJ-1a. In: Lumley, H. de & Beyene, Y. (eds.), Les sites préhistoriques de la région de Fejej, Sud-Omo, Ethiopie, dans leur contexte stratigraphique et paléontologique. Association pour la diffusion de la pensée française (ADPF). Editions Recherche sur les civilisations, Paris, pp. 565–589. Lumley, M.A. de & Marchal, F. 2004. Les restes d’hominidés du site de FJ-1. In: Lumley, H. de & Beyene, Y. (eds.), Les sites préhistoriques de la région de Fejej, Sud-Omo, Ethiopie, dans leur contexte stratigraphique et paléontologique. Association pour la diffusion de la pensée française (ADPF). Editions Recherche sur les civilisations, Paris, pp. 340–383. McDougall, I. & Brown, F.H. 2006. Precise 40Ar/39Ar Geochronology for the Upper Koobi Fora Formation, Turkana Basin, Northern Kenya. Journal of the Geological Society 163 (1), 205–220. Roche, H. 2000. Variability of Pliocene lithic productions in East Africa. Acta Archaeologica Sinica, sup. vol. 19, 98–103. Roche, H., Brugal, J.P., Delagnes, A., Feibel, C., Harmand, S., Kibunjia, M., Prat, S. & Texier, P.J. 2003. Les sites archéologiques plio-pléistocènes de la formation de Nachukui, Ouest-Turkana, Kenya : bilan synthétique 1997–2001. C.-R. Palevol 2 (8), 663–673. Roche, H., Delagnes, A., Brugal, J.P., Feibel, C., Kibunjia, M., Mourre, V. & Texier, P.-J. 1999. Early Hominid stone tool production and technical skill 2.34 Myr ago in West Turkana, Kenya. Nature 399 (6731), 57–60. Rogers, M.J., Harris, J.W.K. & Feibel, C.S. 1994. Changing patterns of land use by Plio-Pleistocene hominids in the Lake Turkana Basin. Journal of Human Evolution 27 (1–3), 139–158. Schick, K. & Toth, N. 1993. Making Silent Stones Speak: Human Evolution and the Dawn of Technology. Simon & Schuster, New York. Semaw, S. 2000. The world’s oldest stone artefacts from Gona, Ethiopia: their implications for understanding stone technology and patterns of human evolution between 2.6–1.5 million years ago. Journal of Archaeological Science 27 (12), 1197–1214. Semaw, S. 2005. Les plus anciens artefacts lithiques (2,6–2,5 millions d’années) des sites archéologiques du Pliocène inal de EG-10 et EG-12 à Gona Est, Afar, Éthiopie. In: Sahnouni, M. (ed.), Le Paléolithique en Afrique: L’histoire la plus longue. Guides de la Préhistoire mondiale, Collection Paléontologie humaine. Éditions Artcom’/Errance, Paris, pp. 13–52. Journal of African Archaeology Vol. 9 (2), 2011 223 D. Barsky et al. Semaw, S., Renne, P., Harris, J.W.K., Feibel, C.S., Bernor, R.L., Fesseha, N. & Mowbray, K. 1997. 2.5-million-year-old stone tools from Gona, Ethiopia. Nature 385 (6614), 333–336. Toth, N. 1985. The Oldowan reassessed: a close look at early stone artefacts. Journal of Archeological Science 12 (2), 101–120. Semaw, S., Rogers, M.J. & Stout, D. 2009. Insights into Late Pliocene lithic assemblage variability: the East Gona and Ounda Gona South Oldowan Archeology (2.6 million years ago), Afar, Ethiopia. In: Schick, K. & Toth, N. (eds.), The Cutting Edge: New Approaches to the Archaeology of Human Origins. Stone Age Institute Press, Indiana University, Indiana, pp. 211–246. Toth, N. 1987. Behavioral inferences from early stone artifact assemblages: an experimental model. Journal of Human Evolution 16 (7–8), 763–787. Stout, D., Quade, J., Semaw, S., Rogers, M.J. & Levin, N.E. 2005. Raw material selectivity of the earliest stone toolmakers at Gona, Afar, Ethiopia. Journal of Human Evolution 48 (4), 365–380. Tamrat, E., Thouveny, N., Taïeb, M. & Opdyke, N.D. 1995. Revised magnetostratigraphy of the Plio-Pleistocene sedimentary sequence of the Olduvai Formation (Tanzania). Palaeogeography, Palaeoclimatology, Palaeoecology 114 (2–4), 273–283. Toth, N. 1982. The Stone Technologies of Early Hominids at Koobi Fora, Kenya: An Experimental Approach. Ph.D. Thesis, University of California, Berkeley. 224 Umer, M., Kaniewski, D. & Renault-Miskovsky, J. 2004. Étude palynologique du site de Fejej FJ-1. In: Lumley, H. de & Beyene, Y. (eds.), Les sites préhistoriques de la région de Fejej, Sud Omo, Ethiopie, dans leur contexte stratigraphique et paléontologique. Association pour la diffusion de la pensée française (ADPF). Editions Recherche sur les civilisations, Paris, pp. 172–185. Wesselman, H.B. 1984. The Omo micromammals: systematics and paleoecology of early man sites from Ethiopia. In: Hecht, M.K. & Szalay, F.S. (eds.), Contributions to Vertebrate Evolution. Karger, New York, pp. 1–219. White, T.D. 1995. African omnivores: global climatic change and Plio-Pleistocene hominids and suids. In: Vrba, E.S., Denton, G.H., Partridge, T.C. & Burckle, L.H. (eds.), Paleoclimate and Evolution, with Emphasis on Human Origins. Yale University Press, New Haven, pp. 369–384. Journal of African Archaeology Vol. 9 (2), 2011