Dzudzuana: an Upper Palaeolithic cave site in the Caucasus foothills (Georgia).

Bar-Yosef, Ofer ; Belfer-Cohen, Anna ; Mesheviliani, Tengiz 等

[ILLUSTRATION OMITTED]

Introduction

Upper Palaeolithic occupations have been reported from caves and rockshelters in the western Caucasus foothills since the late nineteenth century (Zamiamin 1957; Berdzenishvili 1972; Tushabramashvili & Vekua 1982; Bader 1984; Tushabramishvili 1984; Liubin 1989; Meshveliani et al. 2004). The earliest phase of the Upper Palaeolithic sequence was believed to be characterised by the lingering presence of Mousterian tools together with the appearance of distinct Upper Palaeolithic types such as endscrapers, burins and retouched blades. This kind of assemblage was thought to represent the cultural transition from the Middle to Upper Palaeolithic. The prevailing historical-evolutionary concept was that such transition was a normative process with ancestral Middle Palaeolithic human groups evolving unto those of the Upper Palaeolithic (but see Cohen & Stepanchuk 1999).

[FIGURE 1 OMITTED]

Since 1996 a joint team of Georgian, American and Israeli researchers has been involved in systematic excavations of two Caucasian sites: Ortvale Klde Rockshelter (Tushabramishvili et al. 1999; Adler & Tushabramishvili 2004, Adler et al. 2006a) and Dzudzuana Cave (Meshveliani et al. 1999, 2004). These sites lie 5km apart at about 560m asl in the Chiatura region, in the foothills of the Caucasus Mountains (Figure 1). Ortvale Klde contains a sequence of Middle Palaeolithic layers capped by several Upper Palaeolithic occurrences and is fully dated by TL, ESR, and radiocarbon readings (Adler et al. 2008 and references therein).

Here we present an account of the excavated sequence at Dzudzuana, together with its rich assemblages (summarised in Table 1). Details of the sediments and materials are marshalled in an online open access supplementary file, to which reference should be made (http://antiquity.ac.uk/projgall/baryosef328/).

Excavation

Excavations in Dzdudzuana Cave have been conducted in two campaigns. The first in 1966-75 was directed by D. Tushabramishvili, and covered an area of c. 40[m.sup.2] near the cave entrance which was excavated down to bedrock (Figure 2). The excavations were carried out in units 0.1m deep in 1[m.sup.2] squares, and artefacts were recovered by hand, without wet-sieving. The stratification was subdivided into two major units: Layer I--the Upper Eneolithic and Layer II--the Upper Palaeolithic deposits (Liubin 1989). Tushabramishvili further subdivided Layer II into eight sub-layers designated as II-1 to II-8. All except II-8, a sterile deposit immediately above bedrock, contained artefacts and bones.

A second campaign took place in 1996-2008 (Figures 2 & 3) exposing a similar sequence of Upper Palaeolithic units capped by an Eneolithic one, dated by 36 radiocarbon readings (Table 2). Two areas were excavated: the first, an extension of Tushabramishvili's excavations near the entrance of the cave (squares F-I 9-7 and J-K 12-11), which we call hereafter the 'lower area' (LoAr in Table 2, east), the second an 'upper area' comprising squares G-H 24-21, 19-15 (UpAr in Table 2, west). The total excavated surface was c. 24[m.sup.2]. In the lower area (Figure 3, east), the depth of excavated deposits was 4.5m while in the upper area (Figure 3, west) it was 3.25m. The basic units of excavation were 50mm-thick quadrants of 0.5 x 0.5m, within a 1 x 1m grid. The excavated deposits were wet-sieved, dried and later hand-picked in order to retrieve the smallest archaeological components (lithics, bones, etc.).

On the basis of geoarchaeological observations, we divided the sequence in the lower area into four main stratigraphic units (A-D) (Figure 3). In the upper area we uncovered only part of the sequence, namely Units D and C with rare residues of B at the top. As the archaeological sequence was exposed to its full depth only in squares G-H 18-19, the sample representing the upper area in the present account is the one derived from those squares. We subdivided Unit C on the basis of changes in the nature of the clayey deposit and the inclusions of small limestone fragments, into five sub-units numbered from top to bottom as C1 to C5. A special effort was made to correlate the results from the two campaigns of excavations. We used the same grid as Tushabramishvili, and confronted the differences that arose through different techniques of recovery and scale of excavation.

Overall, the sediments in the cave consist of differing proportions of silty clay that vary locally in bedding, colour or rock content (see also supplement 2). There was a general shift in the type of sediment from Units D and C (mudflow deposits) into Unit B (secondary clay accumulations due to dripping water or groundwater). Studies suggest that the deposit was derived from both the inside of the cave (some rockfall and clay) and the outside, from streams (rounded gravel and sandy 'foreign' lithologies, as well as some clay). The lithics and bones in Unit B were found interspersed within lenses of gravels and clay indicating post-depositional processes that most probably interfered with the original spatial distribution of the human occupation. Unit A had been largely removed (as Layer I) in the first campaign.

[FIGURE 2 OMITTED]

Dating

Thirty radiocarbon dates are reported from samples of animal bone and charcoal (Table 2). These show that there is a hiatus of several thousand years between each of the main stratigraphic units, D-A. Dates from Unit D extend from 34.5 to 32.2 ka cal BP and from Unit C to a period from 27 to 24 ka cal BE The three dates from Unit B lie between 16.5 and 13.2 ka cal BP and assign it to the Terminal Pleistocene. The dates from Unit A in the seventh millennium BP attribute it to the Late Neolithic/Eneolithic period, but some sherds dating to the Classical and medieval periods reflect later ephemeral occupations of the cave.

[FIGURE 3 OMITTED]

The only aberrant date is RTT-5741 of 25 300 [+ or -] 570 (31-29.5 ka cal BP), which derives from the very top of Unit C yet is older than the remaining C dates in both the upper and lower areas. The sequence of C in the lower area is thicker and its lower levels are somewhat older than those of the upper area which is further inside the cave. Given the permanent presence of flowing water in the cave it is conceivable that further excavations may reveal that some cave deposits were eroded away, thus the chronological gaps and anomalies were in part due to erosion or redeposition.

The pollen sequence

The palynological analysis shows clear differences between the pollen spectra of the various units. During the accumulation of Unit D, the climate was constantly changing, as appears from the contents of the three pollen spectra. The lowermost part of the unit is characterised by warm climatic conditions. Pollen of wild grape (Vitis silvestris), hazel (Corylus) and oak (Quercus) are present. Later on, the level of humidity grew, and that in turn changed to more dry conditions. In the upper part of this unit, at the contact with Unit C, the pollen reflects a deterioration of the climatic conditions. It is characterised by the presence of birch (Betula), a component of highland forests missing from the layers above and a significant amount of wormwood (Artemisia) pollen that indicates a dry and cold climate. The pollen of Chenopodiaceae and Poacae may reflect the mountainous steppe as indicated also by the presence of pine (Pinus).

The palynological spectra (14 samples) of Unit C are dominated by a large amount of weeds that usually grow on trampled ground attributed to human activities (Behre 1981). These comprise plantain (Plantago), nettle (Urtica) and cocklebur (Xanthium). The presence of chicory (Cichoriaideae) is high in all the studied samples. The tree pollen is characterised by species adapted to more warm and wet conditions than those observed in Unit D. They include wingnut (Pterocarya pterocarpa), walnut (Juglans regia), oak (Quercus), linden (Tilia), alder (Alnus barbata), hazel (Carylus), vine (Vitis silvestris) and pine (Pinus). The presence of parenchyma cells of various kinds of wood reflects the existence of hearths in the cave. Spores of forest ferns such as Asplenium ruta-murraria, Ophioglossum vulgatum, Adiathum and Polystichum confirm the suggestion that a warm and wet climate prevailed during the time of Unit C. Chlamydospores of arbuscular mycorrhizal fungi (Glomus), as well as zygospores of the algae Zygnema and Spirogira found in the cave also support the contention of wet and mild climatic conditions (Navarro et al. 2000).

The palynological spectra (five samples) from Unit B demonstrate a major change from bottom to top. The lower part is characterised by large amounts of rhododendron (Rhododendron caucasica) pollen and other highland elements that suggest the presence of an alpine belt in the vicinity of the cave. This 'colder conditions event' was prior to the Younger Dryas, but additional dates are needed for establishing a firmer climatic correlation. The change from the cold conditions to an increase in humidity is indicated by the presence of a highland dark coniferous forest of spruce, fir and pine with rare occurrences of beech.

In the pollen spectra from Unit A, domesticated cereals, grapes and field weeds make their first appearance. Signs of human impact on forest vegetation are clearly visible. After clearing the forest, the secondary landscapes are usually dominated by light-loving species, such as hazel, alder and bracken (Garylus, Alnus, Pteridium aquilinum) (Behre 1981). The quantities of pollen and spores of all the species mentioned approach very high values and there are also high contents of spores of coprophilic fungi. Those are fungi such as Sordaria, Podospora and Sporormiella which grow in dung of grazing animals. There are also micro-remains of mites and eggs of helminths parasites of both animals and humans (e.g. Trichuris).

Comparing the dates and the pollen sequence, it appears that the cave was abandoned during times when either high precipitation (as between units D and C) or colder periods, such as the Late Glacial Maximum (as between Unit C and B) or during the Younger Dryas (post Unit B), made it uninhabitable. Interestingly, during the proceeding Preboreal and Boreal periods the cave was not in use in spite of the ameliorated climatic conditions.

The artefact assemblages

The assemblages comprised lithics, obsidian, bone objects, animal bone, pollen and flax. The relative quantities found in each of the four main stratigraphic units are tabulated in supplement 3, Tables 3-8. Most of the chipped-stone assemblage consisted of a local chert variety (radiolarite) which is easily obtained either from the plateau above the cave or down in the river bed. The quality of the raw material is mediocre as reflected in the amount of rejects among the items in the debitage category, but the artefacts are predominantly in mint condition. There is a very limited use of obsidian, which was brought from e. 80-100km away (Adler 2002). The ratios of debitage-item per tool in all units indicate that the latter were probably brought over to the cave as finished products. In all the layers, use is made of bone and to a lesser degree, antler, for modifying bone artefacts and ornaments.

Lithics (see also supplement 4)

Unit D

The assemblage from Unit D includes tools, cores and debitage, obsidian and bone tools (Figure 4). A unique type of endscraper, of the rounded variety is found in small numbers yet it is quite distinct and absent from the following Unit C assemblages (Figure 4. 7-8). As always, the endscraper category is prominent and outnumbers the burin category which mostly comprises the dihedral forms. Though obsidian tools account for 3.3% and 3.2% of the tool assemblages of the lower and upper areas respectively, the absolute numbers are quite small, and the same is true for the obsidian debitage items and cores. The bone industry comprises mostly awls and points (9 out of 12). There is a single 'decorative' element, a bone fragment with two triangles cut into it and a single bi-point made of antler.

Unit C

Although Unit C was divided into five sub-units, their assemblages were found to share the same characteristics. The lithic assemblage is dominated by the production of small blades and bladelets detached predominantly from carinated narrow cores, defined as 'rabot' in the pioneering days of the prehistoric research (Bourlon & Bouyssonie 1912) (Figure 5). In the absence of typical Aurignacian tool types this category of carinated cores cannot serve as a cultural marker for the presence of this mostly west European culture (Belfer-Cohen & Grosman 2007). The largest category in both the lower (37.1%) and upper (48.3%) areas is that of retouched bladelets (Figure 5.9-17). A unique characteristic of Unit C (mainly sub-units C2-4) in the upper area is the Sakajia point (defined by Bader 1965; Figure 5.68). Though it appears in small numbers in the sample discussed herewith (0.6%; N = 20), more specimens were recovered from the excavations of Unit C than from the upper area in total. These are arched/curved pointed blades with abrupt retouch (but never bipolar) along the straight edge and a proximal retouched truncation. The blanks were removed from unidirectional cores. Although their shape is reminiscent of Gravette points they differ technologically. Endscrapers are numerous and outnumber burins (Figure 5.18-19). They are diverse in form and type, from the smallest thumbnail scrapers to the bigger varieties, on the end of a flake or a blade. Of all units and areas, the 'lower' C is the richest in obsidian tools, comprising 3.7% of all the tools, while in the 'upper' C they comprise less than half of that percentage (1.2%). The numbers of obsidian debitage items and cores are similar in both areas.

[FIGURE 4 OMITTED]

Two stone pendants were recovered in the Unit C upper area sample presented here. One of them has 31 striations along its circumference (Figures 5.2 & 6.12). Other decorations are made on bone and teeth. A bone pendant was recovered from the lower area while in the upper area there were two tooth pendants, one an incisor of capra, with a drilled suspension hole and the other, a deer tooth with the enamel removed and polished all over (Figure 6.10-11). Other decorated items include incised pieces, sometimes with elaborate patterns, either of long bone splinters or ribs (Figures 6.6 & 6.14). Most of the bone tools belong to the awl/point varieties made by a shaving technique, mostly on splinters, with polishing as the last phase of shaping. Of interest are the antler points which are much thicker and robust than the ones made on bone. Other categories are represented by single specimens, such as a bone needle with an 'eye', a polisher made of a bovid rib, a rib spatula, and a 'retoucher' made on a bone splinter 130mm long and 30mm wide. Overall, Unit C is relatively rich in bone artefacts as there are at least another 100 worked items from the previous excavations and the material not included in the present report, encompassing decorated items as well as points and awls.

[FIGURE 5 OMITTED]

Unit B

Unit B comprises an assemblage characterised by the presence of microgravettes and backed and retouched bladelets (Figure 7). The frequencies of microgravettes decrease from bottom to top. The local character of these assemblages is marked by the consistently high percentages of endscrapers, which always outnumber the burins, through all the archaeological occupations on site. There are elongated blades, c. 80mm long, but most of the debitage comprises flakes, small blades, bladelets and debris. Obsidian tools comprise 1.2% (N = 11) of the total tool count; there are no obsidian cores and only c. 50 debitage artefacts. The worked bone items retrieved during current excavations consist of only eight items, of which one is a decorated, incised rib fragment.

Thus, all through the Palaeolithic sequence there are high frequencies of endscrapers, and to a lesser degree of burins, in conjunction with the particular morphotypes and technologies of each unit. Another type of tool which appears consistently, though in much smaller numbers, is the piece esquillee (Figures 4.15 & 7.11). In addition, we can actually envision the transformation of the Sakajia point into the microgravettes, to be replaced later by asymmetric triangles--all three morphotypes portraying a straight edge shaped through varieties of abrupt retouch combined with basal truncations.

Unit A

The Unit A assemblage comprised some bone tools, pottery and faunal remains, but most of the finds were lithics. Those include polished celts, bifacially retouched knives, spearheads, arrowheads, borers and a few sickle blades, as well as cores and some debitage which may indicate in situ production or modification of the lithic material; some material was probably residual from Unit B.

The faunal remains

Bison (Bison priscus), aurochs (Bos primogenius) and Caucasian tur (goat) (Capra caucasica) are the most common taxa in all occupation levels. The few bison and aurochs remains identifiable to species appear to include both species in equal numbers (Bar-Oz et al. 2008: tab. 2). Other ungulate species are represented in small frequencies and include primarily red deer (Cervus elaphus). It appears that the earliest occupation at the site (Unit D) contains higher proportions of Caucasian tur, while Unit C contains higher proportions of steppe bison and aurochs. In Unit B the percentages of steppe bison and aurochs are similar to those of Caucasian tur. It could be that the differcnces in species abundance between the units reflect differences in the season of occupation. The high frequency of Caucasian tur in Unit D may indicate hunting activities that occurred during the late autumn or winter while the herds descended into the higher part of the forests. On the other hand, the high frequency of steppe bison in Unit C may result from hunting in early spring or summer when bison herds climbed to the woodland in the mountainous areas (see Vereshchagin 1967 and Heptner et al. 1989 for detailed accounts of the behavioural ecology and seasonal migration of Caucasian tur and steppe bison).

[FIGURE 6 OMITTED]

[FIGURE 7 OMITTED]

The taphonomic history of Dzudzuana Cave indicates that the bone assemblage of each unit was accumulated by the same subsistence strategy. This is indicated by the similarity of species composition and its demographic profile, as well as the similarity of carcass processing and marrow extraction (supplement 5). It is worth noting that the same species were hunted at the nearby Palaeolithic site of Ortvale Klde, but in each of the late Middle and early Upper Palaeolithic layers the Caucasian tur is the most abundant prey species (Bar-Oz & Adler 2005; Adler et al. 2006a; Adler & Bar-Oz 2009). It seems reasonable to conclude that the difference in the taxonomic composition reflects differences in the season of occupation and that the lifeways of Palaeolithic hunters were largely determined by the availability of animal resources on a seasonal basis and by the hunting technology (e.g. Bar-Yosef 2004).

Plant and microzoological remains

Although we conducted systematic flotation, no prehistoric plant remains were recovered. However, during the course of pollen analysis numerous non-pollen polymorphs were discovered. Among these were unique finds of wild flax fibres, including spun and dyed ones (Figure 8; Kvavadze et al. 2009). Fibres were recovered from all units, the richest being Unit C (Unit A = 30; Unit B = 48; Unit C = 787; Unit D = 488). The amount of the flax fibres increases from the lowermost sub-unit C5 to the uppermost C1, while the maximal rates (60-90 fibres) are documented in sub-unit C3. It is interesting that besides spun fibres, there are remains of knitted string with numerous knots (Kvavadze et al. 2009: figs. 1 & 2). Dyed fibres are also more numerous in C3. The colours identified are blue, green and pink. Zoological fossils include microscopic remains of mites (Acari) and hair of abdominal segments of larva of beetle (Coleoptera, family Dermestidae). There are also tur hairs, spun and dyed (in grey and green). The remains of fur, micro-remains of skin beetles and moth can be interpreted as evidence for working hide and flax. The samples with the highest content of flax also contained spores of the fungus Chaetomium, which usually grows on clothes and textiles and destroys them.

The regional context

In the early 1970s J. Kozlowski (1970, 1972) restudied most of the Georgian Upper Palaeolithic lithic assemblages and expressed his doubts concerning the validity of the previously advocated tri-partite sequence. He proposed to date the earliest assemblage at Sagvardjile (Layer V) to 34-30 ka BP (which, when calibrated, would fall today in the range of c. 39-34 ka BP), on the basis of long-range comparisons with the Baradostian assemblages from Shanidar Cave, Iran (e.g. Solecki 1958; Olszewski & Dibble 1993). Later, one of us (Meshveliani 1989) demonstrated that the presence of both Middle and Upper Palaeolithic artefacts in the same assemblage is the outcome of mechanical admixture, an observation confirmed in recent excavations (e.g. Meshveliani et al. 2004; Adler et al. 2008) thus indicating that the local early Upper Palaeolithic was brought in by foreign foragers.

Dzudzuana Cave provides us, for the first time, with a chronological framework for the Upper Palaeolithic cultural manifestations uncovered during many years of excavations in caves and rockshelters in the southern foothills of the Caucasus. We are lucky to have clear chronological gaps, indicating the various periods during which the cave was occupied. At the same time, one can see continuity of a local tradition in the lithic techno-typological properties.

The best match for Dzudzuana Unit D is Layer 4 in Ortvale Klde (Adler et al. 2006b: fig. 10). The dates (e. 37.7-33.7 ka cal BP) may indicate a slightly earlier occupation by the bearers of this industry (Adler et al. 2008: tabs. 8 & 9). A somewhat similar industry was recorded in Apiancha Cave, situated in the lowlands near the Black Sea shore and dated to >32.8 ka uncal BP (Korkia 1998). The industry of Unit D also resembles the early Upper Palaeolithic assemblage, radiocarbon dated to c. 38.2-36.8 ka cal BP, reported from Mezmaiskaya Cave (Golovanova et al. 2006; Adler et al. 2008). The latter is dominated by backed blades and bladelets and does not contain any carinated cores. Given the location of the two sites, separated geographically by the peaks of the Caucasus ridges, the apparent contemporaneity raises the possibility that the assemblages represent the same social entity of foragers, indicating that the dispersal of modern humans across a geographic barrier was relatively fast.

[FIGURE 8 OMITTED]

Two Georgian assemblages traditionally assigned to the early Upper Palaeolithic are Sagvardjile Layer V and the assemblage from Samerzkhle Klde (Liubin 1989: 124). The lithics at the latter site comprised numerous long blades, blade cores and rare bladelet cores. Other components include simple endscrapers on blades, dihedral burins and several rabots, i.e. carinated cores, which to us indicate that the assemblages comprise a mixture of earlier and later material. The recently published bone radiocarbon date of 20 160 [+ or -] 160 BP (calibrated 24 093 [+ or -] 328, OXA-7854, Nioradze & Otte 2000), supports our contention of possible contemporaneity with Unit C at Dzudzuana. Moreover, it is of interest to note that there is a similarity between cores from this site (Nioradze & Otte 2000: 280, fig. 13) and cores derived from Unit B in Dzudzuana (Figure 7.12-13).

Unit C at Dzudzuana (27-24 ka cal BP) is characterised by its reduction sequence based on carinated cores. These are often seen as an Aurignacian characteristic; however, there is no justification (either technological or typological) for associating any of the Georgian assemblages (e.g. Dzudzuana, Svanta Savane, Samerzkhle Klde or Togon Klde) with the Aurignacian tradition of western Europe. The presence of the carinated cores may indicate a general contemporaneity among sites in western Georgia, as with the site of Gubs (Amirkhanov 1986) located on the northern slopes of the Caucasus. The bone and antler implements associated with this industry do not comprise typical Aurignacian artefacts (such as the split-base point). Bone awls, needles, points and the like were recovered from Upper Palaeolithic contexts all over the Old World and the same is true for the rare bone beads and decorations.

The Unit B assemblage is rich in blades and bladelets detached from bipolar cores, which differ considerably, in the various steps of core reduction, from those of the carinated cores. This Epi-Gravettian assemblage, characterised by microgravette points is currently dated to c. 16.5-13.2 ka cal BP but could have first appeared at an earlier date. Similar assemblages were excavated many years ago in the Megvimevi Rockshelter, located about 2km south of Dzudzuana. There are at least four additional Epi-Gravettian assemblages, rich in backed bladelets and microgravettes, besides the standard inventory of simple endscrapers, burins and awls, some with increasing percentages of geometric microliths.

Sakajia Cave, located further west from Dzudzuana, recently yielded a radiocarbon date on a bovid bone of 11 700 [+ or -] 80 BP (calibrated 13 583 [+ or -] 148, OxA-7853, Nioradze & Otte 2000). Although the context of the sample is unknown, it should be noted that the original excavators (Schmidt & Koslowski) identified three separate Upper Palaeolithic phases, while later excavators (G. & M. Nioradze) considered the whole Upper Palaeolithic sequence as one unit (Zamiamin 1957; Bader 1984; Liubin 1989; Nioradze & Otte 2000). Therefore, this particular date could represent the uppermost assemblage dominated by microliths.

A similar industry including a local variant of small-shouldered points was reported from Apiancha Cave, further west. It is of interest to note that there are two radiocarbon readings of c. 17.9-17.6 ka cal BP from the site, although their exact context is not reported (Korkia 1998).

Closer to Dzuduzuana, the site of Gvardjilas Klde is known for its rich microlithic component incorporating, besides microgravettes, large numbers of geometric microliths--triangles and lunates. There are also small delicate awls, long borers and high frequencies of small endscrapers (including the thumbnail type). In addition, the rich bone industry includes several ornaments and decorated items. Unfortunately, here also, several Upper Palaeolithic occupations were lumped together. Stratigraphic observations reveal that there were sterile layers interspersed between the archaeological ones, reaching a total thickness of 3.5m (see Bader 1984; Liubin 1989; Nioradze & Otte 2000). Therefore the two radiocarbon dates of 15 960 [+ or -] 120 BP (OxA-7855) and 15 010 [+ or -] 110 BP (OxA-7856) (Nioradze & Otte 2000; calibrated as 19.1-18.2 ka cal BP), on bone artefacts could have come from the lower part of the sequence, prior to the visible dominance of geometric microliths. Our recent excavations at Kotias Klde (Meshveliani et al. 2007) established the dates for the Mesolithic industry rich in triangles as ranging from 12.2 to 10.4 ka cal BP.

There are several other Upper Palaeolithic assemblages that lack any radiocarbon dates and differ in their techno-typological properties from the assemblages enumerated above. One example is the assemblage from Svanta Savane, a site in the lowlands, rich in scrapers with scalariform retouch and high frequencies of burins on truncations. This assemblage resembles similar industries reported from Levantine sites which date to c. 25-20 ka cal BP (Belfer-Cohen & Goring-Morris 2003) and may correlate with the context in Apiancha Cave dated to 25 975 [+ or -] 650 ka BP (30 841 [+ or -] 611 cal BP; LE-3112; Korkia 1998) filling in the first chronological gap in Dzudzuana.

Conclusion

The earliest manifestations of the Upper Palaeolithic in western Georgia are relatively late when compared with the earliest Upper Palaeolithic in the Near East and southeastern Europe. The early assemblages in the Caucasus appear to be already dominated by the production of bladelets. This characteristic of the western Georgian assemblages recalls the Ahmarian blade industries from the Levant (Belfer-Cohen & Goring-Morris 2003). Long-distance comparisons of the local Middle Palaeolithic on both sides of the Caucasus ranges reveal that, while the Mousterian of the southern flanks closely resembles the Mousterian of the Taurus (i.e. Karain Cave) and the Zagros (Adler et al. 2006a), the Late Mousterian of the northern Caucasus is similar to the northern European Micoquian Mousterian.

Hence, while the Caucasus served as a geographic barrier between the two Middle Palaeolithic Neanderthal populations, the early Upper Palaeolithic assemblages on both sides of the Caucasus Mountains demonstrate similarities, indicating the dispersal of modern humans throughout the whole region. The ensuing cultural traditions do not follow the Upper Palaeolithic sequence of western Europe or the Near East as previously claimed. In particular, the 'carinated core' industries found all over the Caucasus region lack any evidence for the presence of the west European 'classical' Aurignacian.

Acknowledgements

This paper is dedicated to the memory of our friend Dr Merab Tvalchrelidze, a geologist, who made the first observations on the cave's deposits. We are grateful to Prof. Abesalom Vekua for his important assistance in the paleontological aspects of our research. We thank Prof. David Lordkipanidze (Director of the Georgian National Museum, Tbilisi) who facilitated our work at the site and shared with us his original unpublished palynological observations. We thank the American school of Prehistoric Research (Peabody Museum, Harvard University) for generously sponsoring this project since 1996, the L.S.B. Leakey Foundation for its support in 2008 (grant to L. Meignen) and the National Geographic Society (grant to O. Bar-Yosef).

Online Supplement (see http://antiquity.ac.uk/projgall/baryosef328/)

Contents: 1) Summary; 2) Sediments; 3) Assemblages; 4) Lithics (Tables 3-7); 5) Faunal remains (Table 8). Tables: 3) Tool types of Units B, C and D; 4) Debitage of Units B, C and D; 5) Debitage of Units B, C and D; 6) Obsidian of Units B, C and D; 7) Bone tools and ornaments of Units B, C and D; 8) Measured values of taphonomic and zooarchaeological variables from Units B, C and D.

References

ADLER, D.S. 2002. Late Middle Palaeolithic patterns of lithic reduction, mobility and land use in the southern Caucasus. Unpublished PhD dissertation, Harvard University.

ADLER, D.S. & G. BAR-Oz. 2009. Seasonal patterns of prey acquisition during the Middle and Upper Palaeolithic of the southern Caucasus, in J.J. Hublin & M.P. Richards (ed.) The evolution of hominid diets: integrating approaches to the study ode Palaeolithic subsistence: 127-40. Dordrecht: Springer.

ADLER, D.S. & N. TUSHABRAMISHVILI. 2004. Middle Palaeolithic patterns of settlement and subsistence in the southern Caucasus, in N. Conard (ed.) Middle Palaeolithic settlement dynamics: 91-132. Tubingen: Kerns.

ADLER, D.S., G. BAR-OZ, A. BELFER-COHEN & O. BAR-YOSEF. 2006a. Ahead of the game: Middle and Upper Palaeolithic hunting behaviors in the southern Caucasus. Current Anthropology 47: 89-118.

ADLER, D.S., A. BELFER-COHEN & O. BAR-YOSEF. 2006b. Between a rock and a hard place: Neanderthal-modern human interactions in southern Caucasus, in N. Conard (ed.) Neanderthals and modern human meet: 165-87. Tubingen: Kerns.

ADLER, D.S., O. BAR-YOSEF, A. BELFER-COHEN, N. TUSHABRAMISHVILI, E. BOARETTO, N. MERCIER, H. VALLADAS & W.J. RINK. 2008. Dating the demise: Neandertal extinction and the establishment of modern humans in the southern Caucasus. Journal of Human Evolution 55: 817-33.

AMIRKHANOV, H.A. 1986. Verkhnij paleolit Prikubanja [The Upper Palaeolithic of the Kuban area]. Moscow: Nauka (in Russian).

BADER, N.O. 1965. Cultural variability at the end of the Upper Palaeolithic and Mesolithic of the Caucasus. Sovetskaya Arkheologiya 4: 3-16 (in Russian).

--1984. The Late Paleolithic of the Caucasus, in P.I. Boriskovski (ed.) Paleolit SSSR [The Palaeolithic of the USSR]: 272-301. Moscow: Nauka (in Russian).

BAR-OZ, G. & D.S. ADLER. 2005. Taphonomic history of the Middle and Upper Palaeolithic faunal assemblage from Ortvale Klde, Georgian Republic. Journal of Taphonomy 3: 185-211.

BAR-OZ, G., A. BELFER-COHEN, T. MESHEVILIANI, N. JAKELI & O. BAR-YOSEF. 2008. Taphonomy and zooarchaeology of the Upper Palaeolithic cave of Dzudzuana, Republic of Georgia. International Journal of Osteoarchaeology 18: 131-51.

BAR-YOSEF, O. 2004. Eat what is there: hunting and gathering in the world of Neanderthals and their neighbours. International Journal of Osteoarchaeology 14: 333-42.

BEHRE, K.-E. 1981. The interpretation of anthropogenic indicators in pollen diagrams. Pollen et Spores 23: 225-45.

BELFER-COHEN, A. & A.N. GORING-MORRIS. 2003. Current issues in Levantine Upper Palaeolithic research, in A.N. Goring-Morris & A. Belfer-Cohen (ed.) More than meets the eye: studies on Upper Palaeolithic diversity in the Near East: 1-12. Oxford: Oxbow.

BELFER-COHEN, A. & L. GROSMAN. 2007. Tools or cores? Carinated artefacts in Levantine Late Upper Paleolithie assemblages and why does it matter, in S.P. McPherron (ed.) Tools versus cores: alternative approaches to stone tool analysis: 143-63. Newcastle upon Tyne: Cambridge Scholars Publishing.

BERDZENISHVILI, N.Z. 1972. On the question of the initial stage of the Upper Palaeolithic of Georgia, in V.M. Masson (ed.) Kamennyj vek Srednej Azii i Kazakhstana. Tezisy dokladov soveschanija [The Stone Age of Central Asia and Kazakhstan]: 23-5. Tashkent: FAN (in Russian).

BOURLON, L. & A. BOUYSSONIE. 1912. Grattoirs carenes, rabots et grattoirs nucleiformes, essai de classification des grattoirs. Revue Anthropologique 22: 475-86.

COHEN, V.Y. & V.N. STEPANCHUK. 1999. Late Middle and Early Upper Paleolithic evidence from the East European Plain and Caucasus: a new look at variability, interactions and transitions. Journal of World Prehistory 13: 265-319.

GOLOVANOVA, L.V., N.E. CLEGHORN, V.B. DORONICHEV, J.F. HOFFECKER, G.S. BURR & L.D. SULERGIZKIY. 2006. The Early Upper Paleolithic in the northern Caucasus (new data from Mezmaiskaya Cave, 1997 excavation). Eurasian Prehistory 4(1-2): 43-78.

HEPTNER V.G., A.A. NASIMOVICH & A.G. BANNIKOV. 1989. Mammals of the Soviet Union, Volume 1: ungulates. Leiden: E.J. Brill.

KORKIA, L.D. 1998. Zedapaleolituri kultura sakartvelos chrdiloaghmosavlet shavizghvistsirethshi [Upper Palaeolithic culture of the north-eastern Black Sea littoral of Georgia]. Tblisi: Metsniereba (in Georgian with Russian summary).

KOZLOWSKI, J.K. 1970. The Upper Palaeolithic of Transcaucasus and the Near East. Part I. Prace komisyi archeologiczney PAN (oddzial w Krakowie) 9: 14-124 (in Polish).

--1972. The Upper Palaeolithic of Transcaucasus and the Near East. Part II. Swiatowit 33:7-47 (in Polish).

KVAVADZE, E., O. BAR-YOSEF, A. BELFER-COHEN, E. BOARETTO, N. JAKELI, Z. MATSKEVICH & T. MESHVELIANI. 2009. 30 000-year-old wild flax fibers. Science 325(5946): 1359.

LIUBIN, V.P. 1989. Palaeolithic of the Caucasus, in P.I. Boriskovski (ed.) Paleolit Kavkaza i Severnoj Azii [The Palaeolithic of the Caucasus and northern Asia]: 9-142. Leningrad: Nauka (in Russian).

MESHVELIANI, T. 1989. About the early stages of the Upper Palaeolithic cultures in western Georgia. Sakartvelos sakhelmtsipo muzeumis moambe 90(B): 13-31 (in Georgian with Russian summary).

MESHVELIANI, T., O. BAR-YOSEF, A. BELFER-COHEN, N. DJAKELI, A. KRAUS, D. LORDKIPANIDZE, M. TVALCHRELIDZE & A. VEKUA. 1999. Excavations at Dzudzuana Cave, western Georgia (1996-98): preliminary results. Prehistoire Europeenne 15: 79-86.

MESHVELIANI, T., O. BAR-YOSEF & A. BELFER-COHEN. 2004. The Upper Palaeolithic in western Georgia, in P. J. Brantingham, S.L. Kuhn & K.W. Kerry (ed.) The Early Upper Paleolithic beyond western Europe: 129-43. Berkeley (CA): University of California Press.

MESHVELIANI, T., G. BAR-OZ, O. BAR-YOSEF, A. BELFER-COHEN, E. BOARETTO, N. JAKELI, I. KORIDZE & Z. MATSKEVICH. 2007. Mesolithic hunters at Kotias Klde, western Georgia: preliminary results. Paleorient 33(2): 47-58.

NAVARRO C., J.S. CARRION, J. NAVARRO, M. MUNUERA & A.R. PRIETO. 2000. An experimental approach to the palynology of cave deposits. Journal of Quaternary Sciences 15: 603-619.

NIORADZE, M.G. & M. OTTE. 2000. Paleolithique superieur de Georgie. L'Anthropologie 104: 265-300.

OLSZEWSKI, D.I. & H.L. DIBBLE (ed.). 1993. The Paleolithic prehistory of the Zagros-Taurus. Philadelphia (PA): University Museum Press.

SOLECKI, R. 1958. The Baradostian industry and the Upper Palaeolithic in the Near East. Ann Arbor (MI): University Microfilms International.

TUSHABRAMISHVILI, D.M. 1984. The Palaeolithic in Georgia. Sakartvelos sakhelmtsipo museumis moambe 37(B): 5-27 (in Russian).

TUSHABRAMISHVILI, D.M. & A.K. VEKUA. 1982. Paleolit Gruzii. Chetvertichnaja sistema Gruzii [The Palaeolithic of Georgia. The Quaternary system of Georgia]. Tbilisi: Metsniereba (in Russian).

TUSHABRAMISHVILI, N., D. LORDKIPANIDZE, A. VEKUA, M. TVALCHERLIDZE, A. MUSKHELISHVILI & D.S. ADLER. 1999. The Palaeolithic rock-shelter of Ortvale Klde, Imereti region, the Georgian Republic. Prehistoire Europeenne 15: 65-77.

VERESHCHAGIN, N.K. 1967. The mammals of the Caucasus: a history of the evolution of the fauna. Jerusalem: Israel Program for Scientific Translations.

ZAMIATNIN, S.N. 1957. The Palaeolithic of western Transcaucasus I: the Palaeolithic caves of Imeretia. Sbornik Muzeja antropologii i etnografii 17:432-99 (in Russian).

Ofer Bar-Yosef (1), Anna Belfer-Cohen (2), Tengiz Mesheviliani (3), Nino Jakeli (3), Guy Bar-Oz (4), Elisabetta Boaretto (5), Paul Goldberg (6), Eliso Kvavadze (7) & Zinovi Matskevich (1)

(1) Department of Anthropology, Peabody Museum, Harvard University 11 Divinity Avenue, Cambridge, MA 02138, USA

(2) The Institute of Archaeology, The Hebrew University of Jerusalem, Mount Scopus, Jerusalem 91905, Israel

(3) National Museum of Georgia, 3 Rustaveli Avenue, Tbilisi 0105, Georgia

(4) Zinman Institute of Archaeology, University of Haifa, Haifa 31905, Israel

(5) Radiocarbon Dating and Cosmogenic Isotopes Laboratory, Kimmel Center for Archaeological Science, Weizmann Institute of Science, Rehovot 76100, Israel; Department of Land of Israel Studies and Archaeology, Bar-Ilan University, Ramat-Gan 52900, Israel

(6) Department of Archaeology, Boston University, 675 Commonwealth Avenue, Suite 347, Boston, MA 02215, USA

(7) Institute of Paleobiology, National Museum of Georgia, 3 Rustaveli Avenue 0105 Tbilisi, Georgia

Received: 12 April 2010; Accepted: 3 September 2010; Revised: 18 October 2010

Table 1. Summary of stratigraphic divisions and evidence at
Dzudzuana Cave.

Strata unit D C

Period Upper Upper
 Palaeolithic Palaeolithic
Date range ka c. 34.5-32.2 c. 27-24
 cal BP
Main plants oak, hazel, birch, pine, oak,
 wild grape, wormwood,
 flax walnut, flax
Implied climate warm-humid cold-warm
Lithics blades carinated cores,
 scrapers, retouched
 burins, obsidian bladelets, obsidian
 (Figure 4) (Figure 5)
Main fauna tur bison, aurochs, tur

Strata unit B A

Period Terminal Late Neolithic
 Palaeolithic
Date range ka c.16.5-13.2 c. 6
 cal BP
Main plants rhododendron, woodland
 fir, spruce, clearance,
 beech cereals
Implied climate cold-wet
Lithics microgravettes knives,
 scrapers arrowheads,
 (Figure 7) sickle blades

Main fauna bison, aurochs, tur

Table 2. Radiocarbon dates from Dzudzuana. Note that the lower area
(marked as LoAr) of the cave (squares F-17-9) produced earlier dates.

Layer Material Square & Area Elevation Lab #

A Bone LoAr TB 315a
A Bone LoAr TB 3156
A Bone LoAr TB 316a
A Bone LoAr TB 3166
A Charcoal K11a; LoAr 315-325 RTT-5700
B Bone I9a; LoAr 465-470 RTT-3282
B Bone I9a; LoAr 420-425 RTT-3821
B Bone H7a; LoAr 525-530 RTT-3278
C Bone I9c; LoAr 565-570 RTA-3433
C Bone I8d; LoAr 570-575 RTA-3434
C Bone H9c; LoAr 575-580 RTA-3435
C Bone H8d; LoAr 590-595 RTT-3822
C Bone G8c; LoAr 600-605 RTT-3823
C Charcoal G9a; LoAr 635-640 RTT-4341
C Charcoal H7a; LoAr 635-640 RTT-4339
C Bone G22b; UpAr* 170-175 RTT-5741
C Bone H196; UpAr 255-260 RTT-4334
C Bone I18c; UpAr 285-290 RTT-5742
C Bone I17d; UpAr 295-300 RTT-5743
C Bone I19c; UpAr 320-330 RTT-5744
C Bone I18b,d; UpAr 300-310 RTT-5746
D Bone F7b; LoAr 630-635 RTA-3436
D Bone G9d; LoAr 635-640 RTA-4338
D Bone F7b; LoAr 635-640 RTA-3437
D Bone F7b; LoAr 640-645 RTT-3438
D Charcoal G8b; LoAr 630-635 RTT-4340
D Charcoal G9c; LoAr 645-660 RTT-4336
D Charcoal G6a; LoAr 685-695 RTT-4701
D Bone I19a; LoAr 390-400 RTT-5745
D Bone I186; LoAr 410-420 RTT-4747

 [delta] [sup.13]C %
Layer PDB Age [sup.14]C BP [+ or -] [sigma]

A 5700 [+ or -] 130
A 6300 [+ or -] 170
A 4600 [+ or -] 130
A 5500 [+ or -] 92
A 5560 [+ or -] 40
B 11500 [+ or -] 75
B 13250 [+ or -] 70
B 13860 [+ or -] 90
C -18.9 21 220 [+ or -] 200
C -19.4 20 980 [+ or -] 150
C -18.7 21 930 [+ or -] 190
C -19.0 20 620 [+ or -] 155
C -19.1 23 240 [+ or -] 200
C -25.2 23 125 [+ or -] 175
C -26.5 22 490 [+ or -] 180
C -18.5 25 300 [+ or -] 570
C -18.9 20 333 [+ or -] 155
C -18.5 20 400 [+ or -] 320
C -18.8 21 200 [+ or -] 350
C -19.0 19 920 [+ or -] 300
C -18.4 20 700 [+ or -] 340
D -18.9 27 150 [+ or -] 300
D -23.9 27 450 [+ or -] 275
D -18.6 27 400 [+ or -] 300
D -19.2 30 350 [+ or -] 400
D -25.9 26 990 [+ or -] 260
D -23.4 26 320 [+ or -] 260
D -23.8 32 140 [+ or -] 500
D -18.9 27 260 [+ or -] 775
D -18.6 29 445 [+ or -] 1015

Layer Age cal BP [+ or -] [sigma]

A 6372-6649
A 6992-7373
A 5073-5461
A 6207-6391
A 6318-6393
B 13 272-13 529
B 15 770-16 591
B 16 890-17 300
C 24 997-25 763
C 24 728-25 551
C 25 874-26 734
C 24 290-24 886
C 27 592-28 227
C 27 231-28 124
C 26 725-27 645
C 29 574-30 943
C 23 954-24 598
C 23 898-24 766
C 24 860-25 888
C 23 410-24 285
C 24 201-25 172
D 31 638-32 072
D 31 833-32 354
D 31 798-32 326
D 34 228-34 962
D 31 531-31 937
D 30 759-31 520
D 35 682-37 558
D 31 165-32 689
D 32 784-34 550

RTA and RTT = Weizmann Institute of Science, TB = Tbilisi University.
Dates calibrated with CalPal-online (CalCurve: CalPal2007-HULU). The
exact stratigraphic provenience of the TB dates is unknown. The lower
area (LoAr) is separated from the upper area (UpAr) for clarity. *
Marks a sample that could have been moved from its original position
in the past.