Bear in Mind: Bear Hunting in the Mesolithic of the Southern Caucasus

We present new faunal data from Kotias Republic of Georgia, where a substantial part of the faunal assemblage consists of brown bear remains ( Ursus arctos ) found in clear association with Mesolithic artifacts. Bear remains are unusually well represented compared to other faunal assemblages from the Caucasus and Eurasia in general. Species diversity, dominance of young individuals, full representation of skeletal elements and skinning butchery marks indicate that bears were actively hunted. Such a hunting endeavor denotes the complex network of relationships that linked the Mesolithic hunting societies with the animal world surrounding them.


Introduction
The regular hunting episodes throughout the Middle, Upper Palaeolithic and Mesolithic across Eurasia involved the acquisition of large prime-age herbivorous prey. Skilled hunting requires detailed knowledge on 'when' and 'where' to ambush or intercept specific animal game, and how to track, trap or capture it. A successful hunt ends in killing, butchering, and consuming the prey. The modes of hunting and butchering of certain species can vary culturally, chronologically and geographically, but they may also share some fundamental similarities in treatment and processing of carcasses.
The daily livelihood of hunters often required frequent encounters with wild animals, some of which displayed fierce defense tactics. However, it is only in later prehistoric periods that hunters started to pursue dangerous carnivore species (e.g., Klein, 2000). Such endeavors involved the hunting of the brown bear (Ursus arctos), apparently an activity imbued with special meaning for hunters of all cultures in the northern boreal zone of the globe. Ethnographic literature enumerates a broad range of ceremonial activities designed to honor and respect the bear (e.g., Hallowell, 1926).
These studies signify that bears have never been hunted for their meat or fur; conversely, the importance of bear hunting lies primarily and foremost in its symbolism.
The great reverence with which brown bears are treated is deeply rooted in the ritual tradition of indigenous societies. One of the most elaborate rituals for many native Siberian people was the feast that followed a successful bear hunt. Carried in procession back to the village, the dead bear was offered food and drink and entertained with songs and dances for several days. Following the celebration the bear body was respectfully butchered and eaten, and its cleaned skull was set on a 4 pole as a guardian totem (Zolotarev, 1937). Also, the word for a Siberian womanshaman is the same as the word for bear (Reid, 2002). Another notable ceremony that involves cultic sacrifice of bears is found among the Ainu people of Japan. For the Ainu the bear symbolized the god of mountains and forests and its hunting was followed by a ritual feast that involved eating the bear's meat and drinking its blood (Hiroshi, 1992). The bear feasts of native Lapps proclaim its hunting as partly ceremonious and Lapp shamans transform themselves into bears (dressed up and wearing masks) when they drum during religious ceremonies (Gjessing, 1947).
Archaeological data supply ample evidence for the particular importance of bear to Palaeolithic people and demonstrate that bears received special attention in spiritual beliefs and ritual behaviors in prehistoric times. Bears are always part of the carnivore scenery in Palaeolithic paintings located in seemingly pre-ordained places in caves, and most bear representations are associated with apparent hunting scenes (e.g., Chauvet, Deschamps, Hillaire, 1996;Morel, Garcia, 2002;Rouzaud, 2002 and references therein). A bear figurine made of mammoth ivory was also found at the Aurignacian of the Swabian Jura in southwest Germany (Conard, 2003). One of the most remarkable artistic representations in this connection is a headless clay (lifesize) modeling of a bear found in the interior part of the Magdalenian cave of Montespan in the French Pyrenees (Begouën, Casteret, Capitan, 1923;reviewed in Kurtén, 1976). It represents a reclining bear with a real bear skull between its forepaws that most probably had once been attached to the sculpture itself. The sculpture is punctured by spear marks, probably thrown during ritual ceremonies. The peculiar engraving of a bear in the Magdalenian cave of Trois-Freres, Ariege, France (Morel, Garcia, 2002), represents another example of the ritual significance of bears 5 to Palaeolithic people. This bear seems to be vomiting blood and there are various markings on its body, perhaps representing spear and projectile wounds.

While bear skeletal remains and stone artifacts often occur in Eurasian
Palaeolithic cave deposits from the Middle Palaeolithic onward, their association in many cases is not clear-cut and in most cases appear to represent separate accumulation episodes (Kurtén, 1958(Kurtén, , 1976Chase, 1987;Stiner, 1994;Stiner, Arsebük, Howell, 1996;Baryshnikov, 1997;Stiner, 1998;Tillet, 2002). Such a case was also reported in western Georgia where several partial (only the upper body and forelimbs) cave bear skeletons were found in a cave devoid of anthropogenic remains.
Yet the cave is located in close proximity (as part of a cavernous system) to Middle Palaeolithic cave sites (Bronze Cave;Tushabramishvili, 1978). Also at Djruchula Cave the dominant species represented are cave bears (Tushabramishvili, 1978; see also Adler, Tushabramashvili, 2004). While there are distinct cases where bear (predominantly the extinct European cave bear, Ursus spelaeus and Ursus deningeri) skulls and perhaps other skeletal parts were intentionally collected by Palaeolithic people, direct archaeological evidence for bear hunting remains a rare phenomenon.
In most cases the mortality profiles of bear remains found in archaeological cave sites represent immature and aged individuals which died during hibernation (e.g., Kurtén, 1958;Gargett, 1996;Stiner, 1998;Lord, O'Connor, Siebrandt, Jacobi, 2007). Another conceivable cause for the accumulation of predominantly young and old adult bear specimens in cave deposits might be due to transportation of bones by large carnivores or scavengers (e.g., Gargett, 1996;Niven, 2006;Argenti, Mazza, 2006). In other cases the occasional appearance of young adult bears might suggest that bears died from falling into natural traps in caves (Wolverton, 2001(Wolverton, , 2006. It is only in rare cases that Upper Palaeolithic and Mesolithic faunal assemblages contain single 6 specimens of brown bears that were most likely hunted. These bear remains were found in clear association with anthropogenic refuse, and some were bearing butchery marks (Bárta, 1989;Stiner, 1994;see also Chaix Bridault, Picavet, 1997). Such finds most probably derived either from attacking hibernating bears or from hunting encounters. Evidence for the killing of bear outside the hibernating season is crucial to recognizing active hunting activities of past foragers.

Faunal analysis procedure
The bone sample analyzed includes only faunal remains that originated from the well defined Mesolithic contexts. Zooarchaeological and taphonomic coding procedures used to collect and evaluate the faunal data are published elsewhere (Bar-Oz, Adler, 2005; see also Bar-Oz, 2004 andBar-Oz, Munro, 2004). Each excavation unit was treated separately and recorded stratigraphically. Skeletal elements and broad taxonomic identifications were carried out in the field using a virtual comparative collection of recent skeletons and osteological catalogues (Schmid, 1972;Hilson, 1999). Identifiable bones included articular ends and shafts of long bones, teeth, cranial fragments, carpals, tarsals, appendicular elements and vertebrae which were 8 assigned to the most discriminating taxonomic level. Finer taxonomic identifications of closely related species were achieved with the assistance of A. Vekua from the Georgian State Museum and the comparative collection of the Georgian State Museum.
The separation of brown bear (Ursus arctos) from cave bear (Ursus spelaeus and Ursus deningeri) was based on morphological and size criteria of selected bones following Kurtén (1958) andStiner (1998;Stiner et al., 1998). Skeletal elements that could not be assigned to species were grouped according to body-size classes. This applies to many of the bear and wild boar remains that were combined in a Sus/Ursus category. This category was easily distinguishable from red deer (Cervus elaphus) specimens which were much larger than those of either bear or boar.
Shaft fragments were coded according to the presence of specific zones (i.e., proximal shaft, mid-shaft or distal shaft) or diagnostic features (e.g., foramen muscle attachment, following the protocol of Stiner, 2004) and other morphological criteria of the shaft fragments (e.g., Barba, Dominguez-Rodrigo, 2005). In most cases identified specimens were coded according to their fraction of completeness (following Marean, 1991). Frequencies of element portions were used to calculate the minimum number of skeletal elements (MNE) and the minimum number of individuals (MNI) (following Grayson, 1984;Klein, Cruz-Uribe, 1984;Lyman, 1994;O'Connor, 2000). The number of identified specimens (NISP) was used as a basic measure of taxonomic abundance (Grayson, 1984).
Recorded elements were examined for macroscopic surface modifications using a low-resolution magnifying lens (x2.5). We observed modifications induced by humans (butchery, burning, and bone fragmentation), animals (rodent gnawing or carnivore puncture, scoring and digestion) and post-depositional attrition agents (weathering, trampling, bleaching, abrasion and root activity) (e.g., Behrensmeyer, 1978;Binford, 1981;Villa, Mahieu, 1991;Lyman, 1994;Bar-Oz, Dayan, 2003). The morphology of limb shaft fracture planes was analyzed for bone fragments bearing ancient fractures. Fracture angle and fracture outline were assessed in order to determine the condition which the bone was in when fractured, i.e. fresh versus dry (see Villa, Mahieu, 1991 for a detailed typological description of the fractures). The degree of completeness of the long bone shaft circumference was recorded (i.e., complete, more than half, or less than half of the complete circumference) in order to discern the role of bone marrow extraction (following Bunn, 1983).
The age structure of bear and wild boar was defined on the basis of epiphyseal fusion of long bones and dental eruption and wear of the deciduous fourth premolar (dP4) and the third molar (M3) (data for bear -Stiner, 1998; for wild boar -Bull and Payne, 1982, Grant, 1982. Each bear tooth was assigned to one of nine wear stages using Stiner's (1998:312-313) wear-scale illustrations. Specimens were clustered into four major age categories: 'neonatal', 'juvenile', 'adult' and 'old adult'. The 'neonatal' group includes bones that by size and texture are of near-born fetuses or recently-born young. The 'juvenile' group includes specimens with unfused long bones, in which fusion occurs by the age of 24 months, and deciduous tooth and/or the permanent counterpart with no wear (stages I-III of Stiner's, 1988 wear scheme).
The 'adult' category comprises successive categories of occlusal wear (stage IV-VII) and fused bones, and the 'old adult' category is defined by teeth with heavy wear (stages VIII-IX).

The faunal assemblage
A total of 775 complete and fragmentary bone elements were identified to taxon (including bone elements identified only to body-size group, Table 1 represented. This is also evident by the rate of long bone shaft circumference. The 11 completeness of bear shaft circumferences is higher than that of ungulates in the assemblage. This data could indicate that bear long bones were not extensively broken as those of other ungulates. It could suggest that in contrast to the ungulate the bear bones were not split open for their marrow. Still both taxa represented show relatively equal ratios of long bones with fresh fractures, indicating that the majority of bones were broken while the bones were still green. Butchery marks appear in low numbers; however their anatomic locations on the red deer and boar remains represent virtually all stages of carcass processing, including skinning (distal metapod of red deer) and dismembering of the carcass (single mark on a proximal rib; on a distal femur of boar; distal humerus and scapula glenoid-fossa of red deer), as well as filleting meat from the bones (two marks on rib shafts and one on medial-shaft of boar femur). The wild boar bones, the largest prey category, comprise all skeletal elements ( Figure 2) and exhibit butchery marks from major stages of slaughtering. This could suggest that their carcasses underwent thorough dismemberment and preparation on site. Yet the number of burnt bones is rather small, mostly on limb distal ends (nine phalanges and two distal metapodia), perhaps suggesting that the boar limbs were roasted before filleting.
The bear remains include three butchery marks that were found on three distal metatarsi. Such a butchery mark is consistent with a circular cut to separate the paws from the fur (Binford, 1981), most probably made during skinning of the carcass.
Burnt bones are evident only among feet specimens. These include four phalanges (three first and a single second phalanges) and a single distal metapod. While the assemblage is too small to reconstruct a detailed skeletal elements profile, it appears that all skeletal elements are represented (Figures 2, 3). Thus it seems reasonable to conclude that entire bear carcasses were transported to the site.
While roe and red deer are represented solely by adult individuals (12 and 10 fully fused long bones; Table 3), the wild boar remains are dominated by a high number of young individuals (approximately 50% are under the age of 24 months) and neonatal piglets (about 10%, Table 3). The size of the neonates suggests that some were less than three months old when killed (Amorosi, 1989). Assuming that wild boar in the Caucasus give birth their young in early spring (March-April;Heptner et al., 1989) the Kotias Klde remains represent animals killed during late spring-early summer. Teeth wear and bone fusion of brown bear remains reveal that the majority derive from adult individuals (Table 3). Several unfused bones indicate the presence of at least one young individual.

Discussion
The Mesolithic faunal assemblage of Kotias Klde rockshelter represents repeated seasonal visits by hunters targeting mainly wild boar and bear. The abundance of piglets and the fact that bear remains were an integral part of the bone refuse indicate that hunting episodes took place during the late spring-early summer. This implies that bears were actively hunted rather than slaughtered while hibernating. This observation is also supported by the demographic profile of the bears killed. The setting of Kotias Klde of the Mesolithic stratum in the middle of the rockshelter's entrance further indicates that the bear assemblage, which includes adult individuals, could not have been created from natural trap capture of bears (Wolverton, 2001(Wolverton, , 2006 13 Bar-Oz, Adler, 2005, Adler et al., 2006, Bar-Oz et al., 2008, the Kotias Klde assemblage displays a different suite of taxa dominated by dangerous prey such as wild boar and bear. The Mesolithic deposit from Darkveti rockshelter (Nebieridze 1978), located a few kilometers away in the Kvirila river gorge, contains the only similar (but much smaller) faunal assemblage within the region.
Brown bear is known from numerous Eurasian Mesolithic sites, normally represented by a few bones in each site (reviewed in Sommer and Benecke, 2005).
However, at Kotias Klde the brown bear is one of the dominant prey taxa.
Though skinning marks may indicate that bear were hunted for their fur (Charles, 1997), it is reasonable to infer that the ultimate reason behind bear hunting was different. Bear hunting offers a challenge for the hunter and enables him to prove his hunting skills. These challenges were most probably interwoven with the mythological and ideational dimensions of bear within Mesolithic societies. The mythological role of bear is apparent in the complex beliefs and practices of indigenous boreal people, some of which were discussed above (Hallowell, 1926;Edsman, 1987;Janhunen, 2003). While the evidence from Kotias Klde may mirror these beliefs and practices, it is difficult to affirm whether the bear remains represent ritual killings such as those documented at bear festivals in various Eurasian societies.
Also, the shift in the assortment of hunted game from the Upper Palaeolithic period to the Mesolithic may indicate the availability of new hunting tools and perhaps a new organizational strategy that allowed the targeting of dangerous prey.
Though we have found no distinct evidence for use of bow or large spear heads, it is possible that the geometric flint tools found (mainly scalene triangles) were used in composite arrows. Ultimately the targeting of dangerous prey at Kotias Klde, such as wild boar and bear, must have been based on considerable skill and courage. Killing 14 young wild boar meant facing their protective, ferocious mothers. The same dangers are heightened when facing hungry and belligerent bears emerging from the forests following the hibernation season. As such encounters likely occurred at close quarters, they tested one's courage and skill; these hunting patterns may represent rites of passage or initiation of young individuals as full fledged members of a hunting society. Ethnographic data support this notion. For example, similar hunting rites were practiced by the Orochi people of Siberia who used to hunt bear with a wooden shaft surmounted by a blade (Hallowell, 1926;Reid, 2002). This was an extremely dangerous undertaking as the spear had to be thrust into the animal's heart from close quarters. Similarly, some other Siberian tribes (e.g., Khant) used to practice hunting rites in which bears were tackled with knives (Reid, 2002). The hunter's left hand was thickly swathed, while his right hand held a long blade.
If the hunting of dangerous prey at Kotias Klde was motivated by ideological considerations it may explain why there was little or no exploitation of bear bone marrow as opposed to other ungulates. Alternatively, the Mesolithic hunters of Kotias Klde, with their intimate knowledge of the animals they exploited, knew that there was little marrow available for extraction following the long hibernation period.
We suggest that the exploitation of bears in the Caucasus Mesolithic can be viewed as part of the complex network of relationships that linked Mesolithic hunting societies with the animal world surrounding them. One may see a link between the activities that took place in Kotias Klde and the beginning of long lasting ceremonial bear hunting and cultic traditions that are strongly rooted and widespread to this very day among the hunting societies of indigenous boreal people.