The Lower Pleistocene lithic assemblage from Dursunlu (Konya), central Anatolia, Turkey.

Gulec, Erksin ; White, Tim ; Kuhn, Steven 等

Introduction

Some of the most enduring questions in palaeoanthropology concern evidence for the repeated expansions of genes, populations and/or cultural practices from sub-Saharan Africa into Eurasia. The initial dispersals of Pliocene hominins and the dispersal of anatomically modern humans during the Upper Pleistocene are the best known examples, but not the only ones. For instance, some researchers hypothesise that a distinct hominin dispersal event was associated with the spread of developed Mode 2/Acheulean technologies with symmetrical handaxes and flake cleavers into Eurasia (Carbonell et al. 1999; Goren-Inbar et al. 2000), whereas other investigators have suggested that H. erectus evolved biologically and technologically in Eurasia and subsequently dispersed back into Africa (Clarke 2000; Roebroeks & Dennell 2005:1100-01).

Our understanding of the timing, mechanisms and archaeological signatures of major dispersal events is limited by large geographic gaps in the database. An increasing amount is known about hominin biological and cultural evolution within Africa, and we have a great deal of information about Western Europe. The Palaeolithic record of the Mediterranean Levant is also extensively documented, particularly given the small size of the area. However, the Pleistocene archaeological and fossil records of much of the rest of the territory that lies between Africa and western Eurasia are considerably less well understood.

Anatolia, or Asian Turkey, represents an especially conspicuous lacuna in most maps of Pleistocene hominin geography. It is certain that the region was traversed by hominin populations during various dispersal events. Anatolia is the most direct land bridge between Africa, central Asia and Europe. Moreover, two of the oldest sites in western Eurasia, 'Ubeidiya (Bar-Yosef & Goren-Inbar 1993) and Dmanisi (Gabunia et al. 2000; Vekua et al. 2002) are situated a short distance to the south and north of Anatolia, respectively. Yet direct, chronologically-controlled archaeological or fossil evidence for early human presence in Anatolia is extremely sparse. For example, numerous surface finds of handaxes and other potential Lower Palaeolithic artefacts have been reported from north-central and especially eastern Turkey (Harmankaya & Tanindi 1996; Kuhn 2002; Takaran 2008). However, Lower Palaeolithic industries and faunas have been documented in geological context at only four sites (including the locality that is the subject of this paper). Two of these, Karain E (Yalcinkaya et al. 1992; Otte et al. 1998) and Yarimburgaz (Kuhn et al. 1996), are caves situated on or near the Mediterranean coast. Dating of the earliest layers at both sites is problematic, but it is unlikely that either exceeds 500 mya. A single locality in central Anatolia, Kaletepe Deresi 3, contains Lower and Middle Palaeolithic assemblages in good context (Slimak et al. 2004, 2008) but the Lower Palaeolithic strata have yet to be dated.

This paper presents findings concerning the lithic artefacts from the Palaeolithic locality of Dursunlu, in south-central Anatolia, initially reported in Gulec et al. 1999. Investigations within this now-disused lignite quarry have brought to light a diverse fauna as well as a small assemblage of chipped stone artefacts. Palaeomagnetic stratigraphy of the sediments as well as faunal indicators provide strong evidence that the artefact-bearing levels at Dursunlu predate the Brunhes/Matuyama magnetic reversal. As such, Dursunlu currently represents the first unambiguous evidence of a hominin presence in this pivotal region during the early Pleistocene.

The site and sediments

The Dursunlu locality is situated in south-central Anatolia, roughly 60km north-west of the city of Konya (Figure 1). Deposits yielding fauna and stone artefacts are exposed within a now-disused and partially-flooded lignite quarry a short distance from the village of Durnslu. Lignite deposits at the locality were exploited in a commercial mining operation that has been discontinued for several years. Although the quarry had long been known as a source of Pleistocene-aged vertebrate fossils there was no systematic study of the locality until it was visited in 1993/1994 by researchers from the University of California at Berkeley, Ankara University and the Turkish geological service (MTA). Subsequent work at the site included intensive surface collection of tip heaps and dumps associated with the quarry, mechanised excavations below the current water table, and careful excavation of large (up to 1.5m diameter) blocks with in situ Pleistocene sedimentary layers preserved within them. These activities resulted in recovery of a large sample of vertebrate fossils along with a small assemblage of modified stone artefacts.

The area around Dursunlu is situated on the Lycaonian plateau, at the northern foreland reaches of the Taurus Mountains. To the immediate north lies the Aksehir basin, a large graben (rift valley) containing a series of extant lakes (Lahn 1948); substantial lakes also existed within the area during the Pleistocene (Luttig & Steffens 1976). The surrounding uplands consist mainly of Palaeozoic and Mesozoic rocks. The stratigraphy of the upper Cenozoic infilling of the intermontane basins in south-central Anatolia has recently been reevaluated (e.g. Unay et al. 1997), producing the following series (from oldest to most recent): (a) Asagi Cigil formation (early Miocene, at least in part); (b) Argithana formation (lacustrine limestones and claystones); (c) Doganhisar formation (alluvial fan conglomerates, sandstones and mudstones). All three formations are known to be represented in the Dursunlu area, although their clearest manifestations are differentially distributed around the basin and adjoining uplands.

[FIGURE 1 OMITTED]

The sedimentary sequence in the Dursunlu quarry locality, documented in two cores drilled in association with the study of the archaeological and palaeontological remains, is formationally distinct. Core 2, situated some distance from the quarry, revealed 38m of sedimentary infill underlain by limnic limestones, probably of Neogene age. Core 1, adjacent to the quarry, extended to a depth of 50m and provided a detailed stratigraphic sequence as well as a preliminary palaeomagnetic record. The basal unit (> 5m in thickness) consists of red muds of alluvial origin. The remainder of the sequence (almost 45m thick) consists of a lacustrine sequence that incorporates multiple episodes of lake-level fluctuation, including several episodes of partial drying. Lignite members occur between -23.6 and -21.25m, and again between -12.5 and -9.45m in Core 1. These massive, blocky, lignite units consist largely, if not entirely, of altered organic matter, representing accumulations of plant material in a waterlogged context.

Vertebrate fossils occur in at least two horizons within the upper part of the Dursunlu sequence. The main occurrence is within the upper lignite member (-12.5 and -9.45m) as well as in the underlying limonitic gravelly sands. Vertebrate remains are also recovered from the base of a deposit of massive, bedded grey-to-yellow silty-clays at roughly 4m below surface. The main source of archaeological material and vertebrate fauna discussed here is the upper lignite.

Palaeomagnetic and palaeontological evidence for the age of the Dursunlu deposits

The palaeomagnetics of Core 1 were assessed at the Berkeley Geochronology Laboratory. The results are reported in greater detail in an earlier paper (Gulec et al. 1999). The sedimentary sequence is incomplete but provides a general notion of the palaeomagnetic sequence. The highest measurements are two readings of reversed magnetic stratigraphy from -7m, indicating the Matuyama (R) chron. If the Brunhes/Matuyama boundary is represented somewhere above this point in the sequence it has not been documented. Between -14 and -25.5m there are 16 more determinations of reversed magnetic polarity. By extrapolation, there are roughly 18m of sediments with reversed polarity in the upper part of the core. Two readings of normal polarity, with weak magnetisation, occur at -27 and -28.25m. Below this, between -28 and -35.7m, are seven additional determinations of reversed polarity. Two additional pairs of readings with normal polarity occur at -37 and -40m, and there is another single reading of normal polarity at -44.5m. A series of six determinations again records reversed polarity between -46.5 and -51.5m.

Although the magnetic record is fragmentary, a preliminary interpretation can be suggested. The two episodes of normal polarity within a long palaeomagnetic record dominated by reversed polarity can be interpreted as representing the Jaramillo (at -27 to -28.5m) and Olduvai (at 28 and -35.7m) sub-chrons. The age ranges of these two events are currently estimated to be 0.99 to 1.07 mya and 1.77 to 1.95 mya, respectively. The main fossil- and artefact-bearing upper lignite layer is situated well within the upper interval of reversed magnetic polarity, more than 12m above the uppermost readings of normal polarity. Based on palaeomagnetic evidence alone it is reasonable to suggest that the archaeological layer predates the Brunhes/Matuyama boundary but post-dates the Jaramillo, i.e. that it dates to somewhere between 0.78 and 0.99 mya. Palaeontological evidence described below is consistent with this age estimate.

The Dursunlu fauna has been described in detail elsewhere, (Louchart 1997; Gulec et al. 1999), and only the most salient features are discussed here. The assemblage identified to date is taxonomically diverse. The infrequent remains of fish, amphibians and tortoises have not yet been studied. The mammalian and avian faunas are larger and more completely analysed. Evidence for hominin modification of bones is scarce. One distal metatarsus of a large bird exhibits several, deep, narrow transversely-oriented incisions which are almost certainly cutmarks, showing some hominin intervention, but it is uncertain how much the fauna as a whole has to do with hominin foraging activities. As a consequence, this discussion focuses mainly on the implications of the faunal remains for palaeo-environment and chronology.

The large assemblage of bird remains, studied by Louchart (1997), includes 13 families, 29 genera and 41 species, only two of which are passerines. Over 80 per cent of taxa represent extant species, whereas seven identified taxa are distinct from modern forms at sub-specific or specific level. Given the nature of the sediments, it is not surprising that most taxa represent organisms that prefer lacustrine or marshy environments. Ducks (Anatidae) are most diverse (17 species), but several species of geese (3), grebes (4), Ardeidae (herons and egrets, 5), rails (5) and waders (Scolopacidae, 3) are also present. Accipiteridae (eagles or falcons), Phasianidae (game birds) and Otididae (bustards) are represented by a single species (or possibly two): the presence of these terrestrial birds suggests open, steppe conditions beyond the margins of the lake or marsh.

Water-sieving sediment from Dursunlu has resulted in the recovery of a substantial and diverse microfaunal assemblage, studied by A. Van der Meulen and E. Unay. Eight families of rodents, comprising at least 12 species, are represented, along with lagomorphs ([greater than or equal to]2 species) and an unidentified insectivore. Many species of small mammal are typical of open, steppe and desert habitat, suggesting that they did not originate in the immediate vicinity but may have been brought in by predatory birds. Chronologically, the most significant taxa are the Arvicolines and Lagurines. Of particular significance are the presence of Mimomys savini, a species with rooted molars, along with Lagurus arankae, and the absence of Arvicola. Overall, the microfaunal assemblage is indicative of a late 'Middle-Biharrian' age, possibly in the range of 0.85 to 0.90 mya, consistent with the palaeomagnetic evidence (Gulec et al. 1999: 355-7).

The large mammal sample from Dursunlu is also diverse. The bones are highly fragmented due to the actions of commercial mining, however, and the incomplete nature of many specimens limits the resolution of taxonomic identifications. In total, four orders, 12 families and more than 20 genera are represented. Megafauna (probiscideans, rhinos and hippos) are present but represented by small numbers of isolated specimens. Cervidae and Bovidae are more numerous, though species-specific identifications are hampered by the fragmentary nature of the bones and the absence of horn cores and other highly-diagnostic elements. Noteworthy is the presence of over two dozen elements attributable to Megalaceros, (species unattributed), consistent with the marshy habitat. Remains of Equidae are the most abundant large animal skeletal elements. At least two taxa are represented. The less common of these is clearly attributable to Equus caballus mosbachensis. More common are remains of a smaller, more gracile equid with hemionine features, most likely Equus altidens. Limonite staining on some bones suggests that they come from the gravelly sands underlying the lignite deposits. Elsewhere in the world this taxon generally occurs within the upper Matuyama (R) chron or Jaramillo (N) subchron (Gulec et al. 1999: 359), again consistent with the palaeomagnetic evidence.

The lithic assemblages

The lithic artefacts from Dursunlu were collected within and around large blocks of consolidated sediments that had been abandoned on the surface after quarrying operations ceased. Many of the artefacts were excavated from the intact sediment blocks found on and in the tip-heaps adjacent to the quarry. As a result many specimens could be attributed to a particular block, although the blocks themselves were not observed in their original positions because the primary deposits are now inaccessible due to the flooding of the quarry. Because we do not know the original spatial relations between the blocks of lignite that were the sources of the artefacts, the entire collection was assessed as a single unit.

As discussed above, there are at least two fossil-bearing levels in the upper part of the Dursunlu sequence, one associated with the upper lignite member between 10 and 12m below the surface, and another associated with silty-clays at around -4m. The upper lignite unit was exposed over a small area via mechanical excavation, but no artefacts were recovered during that operation. In all cases in which the lithics could be provenanced to stratigraphic unit by virtue of being embedded within an excavated block, they were within the upper lignite. Artefacts excavated directly from the blocks came from lignite or from indurated, peaty, silty sediments (sometimes containing gastropods), that occur immediately below the lignite units. Some specimens even exhibit a distinctive bluish-black staining that also indicates their origins in the peat or lignite deposits. Blocks of the grey-to-yellow silty-clay from higher in the sequence that also contain vertebrate fossils were closely inspected in the field. No artefacts were observed within them. Thus, we are confident that the artefact assemblage is derived from the upper lignite and not from the more recent silty-clays.

The most common raw material in the Dursunlu assemblage is milky white vein quartz. It is notoriously difficult to identify technological features on this kind of rock. As a consequence, it was difficult to determine with certainty whether or not a particular object was an artefact, based on its morphology alone (for example, see Mora & de la Torre 2005 and de la Torre & Mora 2005 on the re-evaluation of certain Oldowan artefacts). However, given the nature of the sedimentation, there is no natural agency that could have brought large pieces of vein quartz and flint to this location, so we infer that most or all of these lithics are present because of hominin activity.

Every lithic assemblage, even those made of high-quality raw materials utilising elaborate flaking schemes, includes specimens whose artifactual nature is ambiguous, based on morphology alone. Given the character of the materials from Dursunlu, we decided to confront this ambiguity from the start. During analysis all specimens were assigned a score representing the analyst's certainty that they showed signs of percussion or use. This subjective score varied between one (lowest level of certainty) and five (highest level of certainty). Criteria for scoring artefacts included: 1) the presence of a platform or clear bulb of percussion; 2) well-defined dorsal and ventral surfaces; 3) the number of scars from previous removals on the dorsal face; 4) evidence for secondary modification or edge-damage. Two analysts examined every specimen and scored them independently. Discrepancies in scoring were then discussed and resolved. Specimens that bore no indisputable traces of percussion were excluded from the analysis. The distribution of the remaining specimens among the five scores is shown in Table 1.

The total assemblage of stones originally collected consisted of roughly 175 specimens. A number of the pieces collected during field studies were rejected during analysis because they showed no apparent evidence of human action. That does not exclude the likelihood that these represent manuports or items of debitage with few clear technological features. A total of 135 specimens were identified as showing at least some traces of alteration by hominins. In the tables that follows, counts are shown both for the total assemblage of altered pieces and the subset of specimens most certain to be artefacts (scores 4 and 5 in Table 1) (n = 75).

Ninety-five per cent of the Dursunlu artefacts are made of quartz (Table 2). Flint and an unidentified igneous rock are represented in much smaller quantities. Gravel units within the Dursunlu sequence contain quartz clasts, so this material at least was probably available very close by. However, the majority of the quartz flakes do not possess water-rounded cortex typical of stream gravels. Likewise, occasional small, unmodified quartz clasts found within intact blocks of lower lignite are not rolled. Metamorphic stone with quartz and quartzite outcrops in road cuts a few kilometres from the site (Garniss Curtis, pers. comm.).

The Dursunlu lithic assemblage consists mainly of fakes and flake fragments (Table 3). The great majority (85.3 per cent) of flakes and flake tools are fragmentary. Fewer than 40 per cent preserve platforms, and 36.2 per cent of the total is composed of flake fragments or 'chips' preserving neither the proximal nor the distal ends. High rates of fragmentation are common when vein quartz raw materials and bipolar technology are utilised, as is the case with Dursunlu.

Flake platforms in the Dursunlu assemblage are most often plain (single facet) or else are crushed or collapsed (Table 4). Faceted platforms, typical of Levallois method and some forms of bifacial production, are completely absent. The crushed or collapsed platforms probably reflect the use of bipolar or hammer-on-anvil percussion, a technique frequently associated with exploitation of quartz throughout prehistory. Flakes are relatively short and thick: unbroken flakes average 28.5mm in maximum dimension. None of specimens collected has unmistakable attributes of flakes produced during the shaping or thinning of bifaces. On the other hand, it is unlikely that flakes of quartz raw material like that employed at Dursunlu would resemble classic biface-thinning flakes, regardless of how they were produced. There is little difference in the representation of platform types or raw materials between the whole sample and the sub-sample of specimens that are more clearly artefacts.

A total of 12 specimens show clear evidence of secondary modification as cores or tools (Tables 2 and 5). These include several modified or used flakes, a polyhedron, a chopper, several polyhedral cores, as well as a piece esquillee or bipolar core. Several artefacts are pictured in Figures 2-5. The single polyhedron (Figure 2) is a classic example of this form. It was produced using an unidentifiable, highly weathered type of igneous or metamorphic rock. A single flake of this same material may have been derived from the polyhedron although it does not refit. The most elaborated of the polyhedral cores, manufactured of high-quality flint, is shown in Figure 3. This core is roughly cubic in form, with negative scars from flake removals on several of its faces. There is no evidence of platform preparation: flakes were simply struck from scars left by previous removals in another plane. The other cores, which are made of quartz, present less obvious technological features.

The pieces with retouch or secondary modification are few and largely undiagnostic (Figures 4 and 5). They include three notched flakes and two flakes with partial retouch or heavy use-wear. Edge modification appears minimal, though retouch is not easily recognised on the milky, semi-transparent quartz. It is possible that most flakes were used without further modification. One comparatively large, unmodified flint flake does preserve microscopic evidence of use wear on one margin characteristic of light meat cutting or butchering activity (N. Toth, pers. comm.).

[FIGURE 2 OMITTED]

[FIGURE 3 OMITTED]

The significance of the Dursunlu assemblages

The presence of unambiguous artefacts and cut-marked bone in the upper lignite layer at Dursunlu attests to a hominin presence in central Anatolia during the early Pleistocene. Palaeomagnetic evidence places the artefact-bearing deposits within the later part of the Matuyama chron. Microfaunal evidence may indicate an age in the range of 0.85-0.90 my. The existence of much earlier Pleistocene sites such as 'Ubeidiyah (Bar-Yosef & Goren 1993; Shea 1999) to the south and Dmanisi to the north (Gabunia et al. 2000; Vekua et al. 2002) has long implied that hominins must also have occupied or at least passed through central Anatolia at this time. However, Dursunlu presents the first clear material traces of hominin occupation in the region that can be attributed securely to an early Pleistocene context.

At face value, the Dursunlu assemblage suggests a technological system oriented toward production of small, unstandardised flakes, mainly from poor quality raw materials. Both free-hand, hard-hammer percussion and bipolar technique were employed. The unique chopper and polyhedron may have been cores or they may have been end-products in themselves. There is no obvious evidence for bifacial reduction. However, we cannot exclude the possibility that the absence of evidence for bifacial technology is due to the limited size of the existing collection (e.g. Villa 2001) and to the small size and poor quality of the raw material most commonly employed.

Acheulean handaxes have been reported from surface deposits throughout Anatolia, especially along the major river drainages in south-east Turkey (Harmankaya & Tanindi 1996; Kuhn 2002; Takiran 2008). However, handaxes and cleavers have been found in sealed geological context at only one site, Kaletepe Deresi 3, a stratified open-air site near an obsidian source area located roughly 200km east of Dursunlu (Slimak et al. 2004, 2008). It is worth noting that polyhedrons similar to those from Dursunlu are also quite abundant in Acheulean levels at Kaletepe Deresi 3 (Slimak et al. 2008), much more abundant, in fact, than diagnostic large cutting tools such as handaxes and cleavers.

[FIGURE 4 OMITTED]

[FIGURE 5 OMITTED]

Because the Dursunlu assemblage is small, and the regional record so incomplete, we can draw only preliminary inferences from it. The material collected to date would fall within the catch-all category of 'Mode 1' industries, characterised by unstandardised flake production and lacking both bifaces and Levallois. If the absence of bifaces and biface-reduction debris at Dursunlu is not an effect of small sample size, the presence of such an assemblage in deposits pre-dating the B/M boundary would be consistent with results from other Lower Pleistocene localities such as Dmanisi (de Lumley et al. 2005), Pirro Nord in Italy (Arzello et al. 2007), Orce (Palmqvist et al. 2005) and Sima del Elefante (Carbonell et al. 2008) in Spain, and Kozarnika in Bulgaria (Guadelli et al. 2005), indicating that the earliest lithic assemblages in Eurasia are simple core-and-flake technologies. This would also be consistent with the hypothesis that a developed form of Acheulean spread from Africa into Eurasia later, during the Middle Pleistocene (see Carbonell et al. 1999; Goren-Inbar et al. 2000). However, we cannot exclude sample-size effects at present. Moreover, we lack firm evidence about the chronological relationship between Acheulean and non-Acheulean technologies in Anatolia. There are currently no precise estimates for the ages of the lower levels at Kaletepe Deresi 3, which have yielded handaxes and cleavers.

Clearly, much more needs to be learned about Dursunlu and about other sites in the region before any firmer conclusions can be drawn. The great depth at which the archaeological remains occur at Dursunlu confirms what many researchers have suspected, namely that early Pleistocene deposits preserved the basins of central Anatolia are buried beneath substantial accumulations of later Pleistocene and Holocene sediments. Expanding the early Pleistocene record of central Turkey will require a very careful search for exposures of sediments of appropriate age, in addition to fortuitous events such as those associated with the discovery of the archaeological layers at Dursunlu quarry.

Acknowledgements

The authors are grateful to the many colleagues who aided in the fieldwork at Dursunlu, including Susan Anton, Anne Getty and Cesur Pehlevan. We especially want to recognise the contributions of Garniss Curtis, who conducted the palaeomagnetic studies and contributed many insights about the geology of the Dursunlu quarry. Anonymous reviewers and the editor provided useful comments on the manuscript. The authors express their profound sorrow at the passing of their colleague and co-author, Prof. E Clark Howell.

Received: 16 January 2008; Accepted: 15 May 2008; Revised: 16 June 2008

References

ARZELLO, M., F. MARCOLINI, G. PAVIA, M. PAVIA, C. PETRONIO, M. PETRUCCI, L. ROOK & R. SARDA. 2007. Evidence of earliest human occurrence in Europe: the site of Pirro Nord (southern Italy). Naturwissenschaften 94:107-12.

BAR-YOSEF, O. & N. GOREN-INBAR. 1993. The lithic assemblages of 'Ubeidiya, a Lower Paleolithic site in the Jordan Valley (Qedem 34, Monographs ofthe Institute ofArchaeology). Jerusalem: Hebrew University.

CARBONELL E., M. MOSQUERA, X.P. RODRfGUEZ, R. SALA, &J. VAN DER MADE. 1999. Out of Africa: the dispersai of the earliest technical systems reconsidered. Journal of Anthropalogical Archaeology 18: 119-36.

CARBONELL, E., J.M. BERMUDEZ DE CASTRO, J.M. PARES, A. PEREZ-GONZALEZ, G. CUENCA-BESCOS, A. OLLE, M. MOSQUERA, R. HUGUET, J. VAN DER MADE, A. ROSAS, R. SALA, J. VALLDERU, N. GARCIA, D.E. GRANGER, M. MARTINON-TORRES, X.P. RODRIGUEZ, G.M. STOCK, J.M. VERGES, E. ALLUE, F. BURJACHS, I. CACERES, A. CANALS, A. BENITO, C. DIEZ, M. LOZANO, A. MATEOS, M. NAVAZO, J. RODRIGUEZ, J. ROSELL & J.L. ARUSAGA. 2008. The first hominin in Europe. Nature 452(27): 465-70.

CLARKE, R.J. 2000. Out of Africa and back again. International Journal of Anthropology 15: 185-9.

DE LUMLEY, H., M. NIORADZE, D. BARSKY, D. CAUCHE, V. CELIBERTI, G. NIORADZE, O. NOTTER, D. ZVANIA & D. LORDKIPANIDZE. 2005. Les industries lithiques preoldowayennes du debut du Pleistocene inferieur du site de Dmanissi en Georgie. L'Anthropologie 109: 1-182.

GABUNIA, L., A. VEKUA & D. LORDKIPANIDZE. 2000. The environmental context of early human occupation in Georgia (Transcaucasia). Journal of Human Evolution 34: 785-802.

GOREN-INBAR, N., C.S. FEIBEL, K.L. VEROSUB, Y. MELAMED, M.E. K1SLEV, E. TCHERNOV & I. SARAGUSTI. 2000. Pleistocene milestones on the Out-of-Africa corridor at Gesher Benot Ya'aqov, Israel. Science 289/5481: 944-7.

GUADELLI, J-L., N. SIRAKOV, S. IVANOVA, S. SIRAKOVA, E. ANASTASSOVA, P. COURTAUD, I. DIMITROVA, N. DJABARSKA, P. FERNANDEZ, C. FERRIER, M. FONTUGNE, D. GAMBIER, A. GUADELLI, D. IORDANOVA, N. IORDANOVA, M. KOVATCHEVA, I. KRUMOV, J.C. LEBLANC, J-B. MALLYE, M. MARINSKA, V. MITEVA, V. POPOV, R. SPASSOV, S. TANEVA, N. TISTERAT-LABORDE & T. TSANOVA. 2005. Une sequence du Paleolithique inferieur au Paleolithique recent dans les Balkans: la grotte Kozarnika a Orechets (nord-ouest de la Bulgarie), in N. Molines, M.-H. Moncd & J.-L. Monnier (ed.) Les premiers peuplements en Europe: colloque international, donnees recentes sur les modalites de peuplement et sur le cadre chronostratigraphique, geologique et paleogeographique des industries du Paleolithique ancien et moyen en Europe (British Archaeological Reports International Series 1364): 87-103. Oxford: Archaeopress.

GULEC, E. EC. HOWELL & T. WHITE. 1999. Dursunlu--A new Lower Heistocene artifact-bearing locality in southern Anatolia, in H. Ullrich (ed.)Hominidevolution: lifestyles and survival strategies: 349-64. Gelsenkirchen: Archaea.

HARMANKAYA, S. & O. TANINDI. 1996. Turkiye Arkeolojik Yerlesmeleri, 1. Paleolitik/ Epipaleolitik. Istanbul: Ege Yayinlari.

KUHN, S. 2002. Paleolithic archaeology in Turkey. Evolutionary Anthropology 11: 198-210.

KUHN, S., G. ARSEBUK & EC. HOWELL. 1996. The Middle Pleistocene lithic assemblage from Yarimburgaz Cave, Turkey. Paleorient 22: 31-49.

LAHN, E. 1948. Contribution a l'etude geologique et geomorphologique des lacs de la Turquie. Maden Tetkik ve Arama Enstitusu Yaymlarmdan Seri B, no. 12:1-178 (in Turkish and French).

LOUCHART, A. 1997. Les oiseaux de Pleistocene inferieur de l'ancienne carriere de Dursunlu (Turquie). Unpublished dissertation (Memoire de D.E.A. Paleontologie) University of Montpelier II.

LUTTIG, G. & E STEFFENS. 1976. Paleogeographic atlas of Turkey from the Oligocene to the Pleistocene (explanatory notes). Paleogeographic atlas of Turkey 1:500,000. Hanover: Bundesanstalt fur Geowissenschaften und Rohstoffe.

MORA, R. & I. DE LA TORRE. 2005. Percussion tools in Olduvai Beds I and II (Tanzania): implications for early human activities. Journal of Anthropological Archaeology 24: 179-92.

OTTE, M., I. YALCINKAYA, J. KOZLOWSKI, O. BAR-YOSEF, I. LOPEZ BAYON & H. TAKIRAN. 1998. Long-term technical evolution and human remains in the Anatolian Palaeolithic. Journal of Human Evolution 34: 413-31.

PALMQVIST, P., B. MART|NEZ-NAVARRO, I. TORO, M. PATROCINIO ESPIGARES, S. ROS-MONTOYA, V. TORREGROSA & J.A. PEREZ-CLAROS. 2005. Reevaluation de la presence humaine au Pleistocene inferieur dans le Sud de l'Espagne. L'Anthropologie 109(3): 411-50.

ROEBROEKS, W. & R. DENNELL. 2005. An Asian perspective on early human dispersal from Africa. Nature 438(22/29): 1099-1103.

SHEA, J. 1999. Artifact abrasion, fluvial processes, and 'living floors' from the Early Paleolithic site of Ubeidiya. Geoarchaeology 14(2): 191-207.

SLIMAK, L., H. ROCHE, D. MOURALIS, H. BUITENHUIS, N. BALKAN-ATLI, D. BINDER, K. KUZUCUOLU & M. GRENET. 2004. Kaletepe Deresi 3 (Turquie). Aspects archeologiques, chronologiques et paleontologiques d'une sequence Pleistocene en Anatolie centrale. Comptes Rendus Palevol de l'Academie des Sciences de Paris 3:411-20.

SLIMAK, L., S. KUHN, H. ROCHE, D. MOURALIS, H. BUITENHUIS, N. BALKAN-ATLI, N., BINDER, C. KUZUCUOLU & H. GUILLOU. 2008. Kaletepe Deresi 3 (Turkey): archaeological evidence for early human settlement in Central Anatolia. Journal of Human Evolution 54(1): 99-111.

TASKIRAN H. 2008. Reflexions sur l'Acheuleen d'Anatolie. L'Anthropologie 112(1): 140-52.

DE LA TORRE, I. & R. MORA. 2005. Unmodified lithic material at Olduvai Bed I: manuports or ecofacts? Journal of Archaeological Science 32 (2): 273-85.

UNAY, E., M. KARABIYIKOLU, N. KAZANCI & G. SARAC. 1997. The Dursunlu open cast mine, in S.A.G Leroy & C. Ravazzi (ed.) Volume of abstracts and guidebook for the Inter-INQUA colloquium on Milankovitch and Plio-Pleistocene vegetation succession from 26 to 09 Ma, 29 March-1 April 1997, Ankara, Turkey: 69-74. Ankara.

VEKUA, A., D. LORDKIPANIDZE, G.P. RIGHTMIRE, J. AGUSTI, R. FERRING, G. MAISURADZE,A. MOUSKHELISHVILI, M. NIORADZE, M. PONCE DE LEON, M. TAPPEN, M. TVALCHRELIDZE & CHRISTOPH ZOLLIKOFER. 2002. A new skull of early Homo from Dmanisi, Georgia. Science 297: 85-9.

VILLA, P 2001. Early Italy and the colonization of Western Europe, in O. Bar-Yosef & L. Straus (ed.) Out of Africa in the Pleistocene (Quaternary International 75 special issue): 113-30. Oxford: Pergammon Press.

YALCINKAYA, I., M. OTTE, O. BAR-YOSEF, J. KOZLOWSKI, M. LEOTARD, H. TAKIRAN. 1992. Karain 1991: recherches paleolithiques en Turquie du sud. Paleorient 18/2: 109-22.

Erksin Gulec (1), Tim White (2), Steven Kuhn (3), Ismail Ozer (1), Mehmet Sagir (1), Hakan Yilmaz (1) & F. Clark Howell (2) ([dagger])

(1) Ankara Universitesi, Dil ve Tarih-Cografya Fakultesi, Antropoloji Bolumu, 06100 Sihhiye, Ankara, Turkey

(2) Department of Integrative Biology and Human Evolution Research Center, University of California at Berkeley, Berkeley, CA 97420, USA

(3) Department of Anthropology, Bldg. 30, University of Arizona, Tucson, AZ 85721-0030, USA

Table 1. Scores of analysts' certainty as to evidence for human
action on 135 potential artefacts from Dursunlu.

              Low     Moderate   High

Certainty    1    2    3    4      5
            12   14   35   46     28

Table 2. Raw materials of flakes and flake tools from Dursunlu.

Raw material    All   Scores 4 and 5 only    With secondary
                                              modification

Quartz          120           62                   9
Flint             5            4                   2
Igneous           1            1                   1
Indeterminate     1           --                  --
Total           127           67                  67

Table 3. Blank types for flakes and flake tools from Dursunlu.

Blank type            All   Scores 4 and 5 only

Cortex flake (>75%)     5            3
Plain flake            37           29
Bipolar flake          12           12
Flake fragment         60           23
Chunk                  13           --
Total                 127           67

Table 4. Platform types for flakes and flake tools from Dursunlu.

Platform type   All   Scores 4 and 5 only

Cortex            4           3
Plain            27          20
Dihedral          5           5
Crushed          13          12
Total            49          40

Table 5. Artefacts with secondary
modification, (core tools, cores and flake
tools) from Dursunlu.

Form               Number

Notch                 2
Modifieded flake      3
Tested chunk          1
Chopper               1
Polyhedron            1
Polyhedral core       3
Piece esquille        1
Total                12