North Ossetia is situated in the center of the Caucasian isthmus between the Caspian and the Black Seas. It is very mountainous with well-defined vertical environments situated between 100 and 5000 m AMSL. In the ancient past it attracted people due to the variety of habitats and its diversity of animal and plant life.    It is generally thought that human occupation of the Caucasus began in the Acheulean (700,000 to 400,000 years ago) during the Gunz-Mindel. However, artifacts dating to ca. 1 million years ago have been discovered at Azych Cave, in Azerbaidjan, and the site of Dmanisi in Georgia is dated to ca. 1.9 mya, and contains H. erectus fossils.

   Paleolithic research in North Ossetia began in 1954 with the investigations of V. Lubin. He discovered a few sites with Paleolithic tools from shallow cultural layers. However, his search for Paleolithic sites with intact stratigraphy proved futile. Dr. Nazim Hidjrati discovered Weasel Cave in 1981. It is the first intact, stratified Paleolithic cave situated on the northern slopes of the Central Caucasus. Weasel Cave is a westward facing cascade-type formation in upper Jurassic dolomite, it is found at 1125 meters AMSL, and contains stratified Medieval, Iron Age, Bronze Age, Chalcolithic, Mesolithic and Paleolithic occupations. There is excellent preservation of not only pollen and faunal materials, but also wood charcoal and seeds (plum) from ca. 10 hearths.

Stratigraphy

   Layers 1-3 contain Medieval through Chalcolithic assemblages. The Mesolithic is found in the upper portions of layer 4. The climate is wet and temperate. The middle and lower part of the layer (and Layer 5) are filled with limestone eboulis, representing cold and very cold periods. These lower portions of Layer 4 correspond to the late Wurm, and should contain Upper Paleolithic assemblages. However, there are no Upper Paleolithic sites yet discovered in the north-central Caucasus. Pollen indicates a pine forest close by, and at lower elevations, a birch forest.

   Layer 5 produced a Mousterian assemblage consisting mainly of tools, but there are also hearths with charcoal and broken animal bones. Radiocarbon assays on burned bone from this layer are 32,980 + 1070 and 34,288 + 1235 BP, placing it among the most recent Mousterian deposits known.

   Layer 6 also contains a large hearth, but the recovery of several forest dwelling rodent species that characterize layers 6-11 indicate a warmer climate.

   In Layers 6-11, pollen studies distinguish two stadials and two interstadials. During the stadials, grasses and birch trees increase, while during the interstadial, pine increased. These layers formed during rapid changes from subalpine meadows to birch and fir forests, which is a typical pattern for early Wurm climates. Radiocarbon dating on bone indicated an age greater than 38 thousand years.

   Lower Layer 11 and the surface of Layer 12 represent cold, dry conditions. Layer 12 is covered by large limestone blocks from massive roof collapse.

   During the deposition of layers 12-14, the climate was very warm and damp. The area held broad-leaved birch and chestnut forests mixed with hornbeam, oak, elm and lime-trees. Chestnut, lime and walnut do not grow in the Caucasus Range today, and testify to a very warm, damp climate. Apparently, this was the Riss-Wurm interglacial, which is generally dated to between 75 and 125 thousand years ago. These layers contain hearths, bones and stone-tools.

   The pollen assemblage of Layer 15 reflects an Alpine grassy meadow. The middle of the layer probably represents the Riss glacial maximum. Stone tools are found in the upper and lower horizons.

   Layers 16-18 comprise a single complex during which Pine and Juniper prevail in the region of the cave, layer 17 contains stone tools, while Layer 18 is 70 cm of volcanic ash, representing one eruption episode of Mt. Kazbek, some 15 km to the southeast.

   Layer 19 represents a dry and warm climate with broad-leaved forests of elm, beech, walnut, hornbeam and oak. This is evidently a warm interstadial or interglacial period. Stone tools are present.

   Layers 20-25 are ca. 5 meters thick. They contain volcanic ash and stone tools of indeterminate industry. At present these lower layers are undated, but the pollen profile points to a Middle Pleistocene age. Absolute dating of the volcanic ash layers that is underway will resolve these chronological questions shortly.

   Faunal remains are very well preserved at Weasel Cave. Asiatic deer, Caucasian mountain goat, and cave bear are most frequently represented. Horse, roe deer, wild boar, and bison are less frequently observed. Mountain goat dominate the upper Mousterian Layers 5-11; and Asiatic deer and cave-bear dominate the middle Layers 12-14. Similar faunal patterns are observed in Kudaro Cave, South Ossetia, Georgia, where this pattern is associated with the boundary between the Acheulean and Mousterian, indirectly confirming the provisional chronology of Weasel Cave.

The Lithic Industry of Weasel Cave

   The vast majority of stone artifacts are made of silicified quartzite. Only rarely are artifacts made of andesite, flint, or other raw materials. The lithic industry of the lower Layers 15-21 (Photos: 15-19, 20-21) is represented by a small number of artifacts so far, but they are quite similar to industry from Layers 12-14. (Photos: 12, 13, 14ab, 14cd)

   Layers 12-14 reveal a Denticulate Mousterian of the Levallois type with both Levallois blade and Typical Mousterian tool types. This blade industry is probably related to those early blade containing industries reported from the Levant. In contrast, Layers 5-7 contain a Typical Mousterian of the Non-Levallois type, without blades. This combination of different features of development is one of the most peculiar aspects of the inventory, which has no good analogies in other Caucasian industries. There are certain similar elements in the Mousterian of South Ossetia (Tskhinval culture), and in the Acheulean of the Kudaro Cave, however.

   The process of technological evolution (which is generally assumed to be progressive) appears to change after Layer 12. In Layers 5-7 we find an "archaic"-looking industry. Chronologically these layers are from the early Upper Paleolithic, but technologically they would seem more at home in earlier time periods.

   A detailed analysis shows that the many of the characteristics that are vividly displayed in Layers 5-7 also occur in Layers 12-14, where at least some tools were made using the same technological processes. There are also stylistic similarities in some techno-morphological elements common to the entire inventory.

   An unique attribute-based analysis technique called "combinatory analysis", which was developed by Dr. Hidjrati, shows connections between the assemblages from Weasel Cave and those from other cave and open air sites in the area around Kuardzin Mountain, and makes it possible to define a “Kuardzin Culture” from the Middle Paleolithic.

Detailed Analyses

   Our detailed analyses have focused on the most recent Mousterian in Layer 5.

Spatial Patterning

   Koetje’s analysis of the spatial patterning in layer 5 was conducted using K-means clustering, and diversity measures. An initial examination of layer 5 showed no good evidence for more than a single occupation episode, which was centered around remnants of a large hearth situated towards one edge of the excavated area. There were, for example, no differences in elevation or orientation within any artifact class, nor any significant correlations between elevation or orientation by artifact class and distance from the hearth center. Clustering was therefor conducted using the two dimensional position of each object.

   Five analytically important cluster solutions were found. Diversity was measured for each cluster with objects grouped into debitage, tool, bone and bone splinter classes. Cluster diversity is initially low and decreases as the number of clusters increases compared to the diversity of a simulated assemblage with the same quantitative makeup. Thus the clusters can be considered less diverse than would be expected if the objects were being distributed randomly, and this lower than expected diversity continues to decrease as the spatial resolution (number of clusters) of the analysis increases. In short, there is good evidence for differential grouping of distinct artifact classes, suggesting either a single very structured use of the area, or a series of depositional events that tend to be associated with either a single, or very few artifact classes.

   Similar analyses of much earlier Mousterian occupations at Vaufrey in SW France (Simek 1987, Simek and Rigaud 1989), showed areas that were used both intensely and redundantly, with diversity patterning high and stable as spatial resolution increased. Clearly debris and tools from many different activities overlapped strongly in space. This pattern was linked to Binford’s “foraging” types of subsistence strategies, expected to typify the Mousterian. The pattern found here indicates intensive, though highly structured rather than redundant use of space, and would be associated with “collecting” types of subsistence strategies thought to typify the Upper Paleolithic. These patterns have been found at a number of French Upper Paleolithic sites such as the Aurignacian and Gravettian from Le Flageolet I, and the Magdalenian at Pincevent and several sites in the Isle River valley.

   Although far from a conclusive definition of Caucasian Late Mousterian Subsistence strategy, this analysis supports the idea that late Mousterian groups were using “collector” types of strategies in the high mountain regions. Certainly the patterns found here are more like those found in the Upper Paleolithic than those recovered so far from other Mousterian sites. If these associations hold up, it suggests, at least, that there was not a clean replacement of foraging by collecting strategies across the Middle to Upper Paleolithic transition in the Caucasus and calls into question the idea of a replacement at all.

Traceology

   Kimball’s microwear analysis of the Mousterian assemblage from Level 5 has only begun. So far, however, the typological Mousterian Points (Bordes Type 6, Photos: I & II)of nonlocal flint and one dejete scraper (Type 21) have been analyzed. The high-power, Keeley Method with chemical cleaning was employed. Polishes are well-preserved with little evidence of chemical or mechanical alteration.

   Both implement types were found to have been used hafted and in wood-working tasks. The wood polishes (top photos) are typical of wood polish, in being very bright, and smooth, with an irregularly clear contour, and in being both continuous and concentrated along the edge. The polishes are best developed near the point, which suggests the point itself (and not just the lateral edges) was an important functional element.

   Hafting traces (bottom of photos)are observed on the dorsal and ventral faces. These hafting traces show variation from the “striated” hafting polish reported by Anderson (1981,1990) and Beyries (1987,1988) and which Anderson and Helmer (1987) have experimentally replicated with coarse vegetal cord attached to a wooden haft, allowing tool movement in the haft.

   A striated hafting polish is observed on one Mousterian point but the opposite face shows a different polish. These are brighter,with slightly domed to undulating micro-topographies. Kimball (1989, 1994, 1995) has documented similar traces on Gravettian points from Le Flageolet I and on a variety of Archaic and Early Woodland projectile points from eastern North American. The latter have been replicated in hafted butchery and projection experiments.

   The use of these "typological" Mousterian Points in wood-working actions -- in a low-angle planing or whittling motion, fits the functions of double convergent side scrapers in the microwear studies of Anderson and the Levallois points studied by Beyries.

   The preliminary evidence from Weasel cave also fits well with the results of studies by Shchlenskii (1974; Plisson 1988) of Mousterian assemblages from Monasekaja (Goubskaja), La Gouba, Nosovo I, and Erevan (in the Ukraine, northwestern Caucasus, and Armenia, respectively) which failed to detect any evidence of projectiles. Similarly, the microwear analyses of Gradcja Kararjan (1990) at Erevan I and Lusakert I in Armenia revealed a lack of projectiles and the use of Mousterian points and convergent scrapers in wood-working tasks.

   Shea (1991) and Callow (1986) have found evidence of use as projectile points on Levallois and Mousterian points (respectively), but relied exclusively on impact fractures and striations for their assignment of points as projectiles.

   From the microwear analysis of bifacial projectile points from experiments with atlatl darts by Chris Bergman and Kimball, and similarly organized microwear studies by Fischer et al. (1984) and Geneste and Plisson (1985); it is clear that impact fractures, linear striations, and polishes (fresh hide and spots of bone polish) can be expected for projectiles, none of which were found on the Weasel cave specimens.

   While we obviously have a way to go in the traceological study of the Weasel Cave Mousterian industries, we believe that the results of this research will add an important linkage between the emergent models of Mousterian tool usage in western Europe and the Near East. In particular, we feel that the issue of the functional significance of Levallois and Mousterian points can be addressed with microwear data from the lithic assemblages at Weasel Cave. Unretouched Levallois points, retouched Levallois points, and Mousterian points represent 8-27% ofthe Level 12-14 assemblages and afford an excellent opportunity to address this problem.

Conclusion

   The significance of the research at Weasel cave hinges on the well defined and intact stratigraphic sequence with one of the most recent Mousterian deposits at its top, and at least two datable volcanic ash deposits at lower levels, the excellent preservation of a broad range of faunal and vegetable materials, and the blade based Levallois technology. Four basic questions are being addressed in the continuing project: The lifeways of the Middle Paleolithic peoples adapted to the Caucasus mountains, the origins, organization and functions of the Levallois blade industry, the detailed chronology of the deposits, which are currently being assayed by Alan Deino of the Berkely Geochronology Center, and what these can tell us about the emergence and spread of early modern humans, and the fate of Neanderthals in Eurasia.