Bone damage, site definitions, literature comparisons, food ways
At the beginning of this study it was assumed that at least seven perimortem damage stigmata result from human butchering and cooking practices. The combination or signature was first identified in cannibalized prehistoric Native American skeletal remains. They include: cut marks, perimortem breakage, anvil abrasions, burning, missing vertebrae, fragment end-polishing (pot polishing), and chop marks. The first five were identified by the senior author and associates (Turner and Turner 1999). Tim White (1992) proposed the end-polishing following a cannibalized assemblage of prehistoric Anasazi Indians found in Mancos Canyon, southwestern Colorado. The seventh trait was added following the work of Carmen Pijoan (1997) who, along with her co-workers, differentiated chop marks from cut marks in cannibalized prehistoric Mexican Indian assemblages. The combination of damage types has been found repeatedly as a multivariant signature of cannibalistic body processing of prehistoric Indians of the southwestern United States and Mesoamerica. Several of those assemblages were discovered in archaeological contexts that largely ruled out the possibility that the damage was caused by carnivores or any other identifiable non-human agency. In the prehistoric American Southwest and Mesoamerica, humans alone were responsible for this distinctive multi-feature processing signature. Comparisons between the human butchering and that of large and small game animals (Dice 1993, Turner and Turner 1999, White 1992) showed remarkably similar kinds and amounts of damage. It was on the basis of this overall similarity of butchering that cannibalism was cautiously proposed.
Operating on the principle of uniformitarianism, we assumed that these same stigmata would also identify butchering and carcass processing of non-human animals carried out by late Pleistocene Siberian hunters. However, since this assumption had never been tested in Siberia, we added to our multi-site survey paleontological assemblages whose context and content indicated perimortem bone damage caused only or mainly by carnivores, in particular, hyenas. It was quickly learned that missing vertebrae would not serve to identify late Pleistocene human activity since vertebrae were rare in both the archaeological and paleontological contexts (Table A1.3; on average <5.0%).
Since nearly all of the sites in this study are pre-ceramic, there is no chance that polishing could have occurred by stirring bone bits in a pottery cooking vessel. If polishing were caused mainly by human cooking practices, then the vessels must have been made of wood, leather, or tightly woven basketry. Boiling crushed bone fragments and adhering soft tissue in wood, hide, or basket containers to extract calorie-rich marrow grease would have required the addition of heated stones, whose abrasive surfaces would have caused much the same sort of polishing as the abrasive interior walls of pottery cooking jars. However, there would be a difference because hot stones in bag-, wood bowl-, or basket-boiling would also polish the middle portions of bone fragments, not just the ends, as would occur when stirred in a geometrically spherical pottery cooking vessel. Hence, polishing was adjusted from end-polishing to an end-middle polished condition. We continued to assume that polishing, in this modified condition, was solely or mainly the product of humans, even though we had a few examples of bones chewed and polished all over by domestic urban and ranch dogs. Analysis at the end of our study showed end-middle polishing would have to be abandoned, along with missing vertebrae, as a marker trait of human presence. As Table A1.27 shows, end-middle fragment polishing is more frequent in the hyena sites than in the archaeological sites without hyenas, and in the archaeological sites with mixed human and hyena presence. While these results do not rule out boiling in non-ceramic cooking vessels, they do show that identical end-middle polishing had happened as a result of carnivore chewing, and probably some other agency as well, possibly geochemical.
As for abrasions (caused in the American Southwest by slippage during the cracking open of marrow bones with the use of a hammer stone and stone anvil), the findings herein proved disappointing as an indicator of human presence. The mainly hyena sites have an unweighted mean of 1.7% abraded pieces (weighted mean, 0.6%). The mixed archaeology and hyena sites’ unweighted mean is 2.0% (weighted, 2.3%). The archaeology-no hyena sites’ unweighted mean is 0.6% (weighted, 0.5%). In other words, the stigmata that we define as abrasions, and attribute to humans alone, occur almost with the same frequency in the three types of contextual settings. The difference between mainly hyena and mixed is statistically significant (X = 18.9; 1 d. f. ; P <0 .001), whereas it is not in the mainly hyena/no hyena comparison (x2 = 0.16; 1 d. f. ; P >0 .05). By themselves, abrasions do not distinguish between carcass and bone processing by carnivores and humans. In the above-mentioned cannibalism studies, the kinds of perimortem damage that might possibly be found in fully paleontological faunal assemblages were not systematically studied due to the lack of appropriate comparative paleontological collections with perimortem damage. While there are a few well-known Pleistocene animal traps in North America, such as the La Brea tar pits in southern California, these are inappropriate for identifying a carnivore damage signature. Indeed, much of the reason for the present study was to take advantage of the ideal kind of comparative faunal assemblages that the Siberian hyena cave sites provide for taphonomy research. This is especially true for perimortem breakage.
Perimortem breakage is very common in both archaeological and paleontological assemblages, averaging about 85%. Statistical comparisons provided in Table A1.28 show that the amount of perimortem breakage in mainly hyena sites and non-hyena archaeology sites is about the same as in hyena and mixed hyena-archaeology sites.
Hence, by itself, perimortem breakage is of no value as a perimortem damage indicator of human activity, at least not quantitatively. While we kept our eyes open for potential distinctive qualitative features of perimortem breakage, none was recognized. At this point, by elimination, we are left with only burning, chopping, and cutting as indicators of butchering and cooking activity by humans.
Table A1.29 shows the frequencies of pieces with burning, cutting, and chopping. Burning is rare to absent in most of our assemblages. Only one burned piece was found in 1638 pieces (0.08%) that came fTomthe mainly hyena sites. This is significantly less than the occurrence of burning in the mixed archaeology-hyena type of sites (1.6%; x2 = 41.9; 1 d. f.;p <0.001) and in the archaeology, no hyena sites (1.3%; %2 = 9.4; 1 d. f.;p <0.01). Burned bone pieces occurred around five times more often in the two classes of archaeological sites than in the mainly hyena sites. Burning remains an indicator of human presence if natural burning such as forest or steppe fires can be excluded as the cause. However, based on our study, its frequency is so low that sample size needs to be substantial (100 or more pieces) to expect to find a 1.0-2.0% occurrence of burned fragments.
Whatever ways late Pleistocene Siberians cooked their game, our assemblages (5170 pieces from 13 sites) show without any question that roasting was not the usual or preferred on-site practice. Had it been, many more burned pieces would be expected. Moreover, we would also expect to have some pieces with both burned and unburned areas, dependent on the thickness of overlying protective soft tissue. No such incompletely burned piece was found in any assemblage, not even in the thousands of pieces excluded from study because they were smaller than our minimal size selection criterion of 2.5 cm or because of excessive root damage.
Chop and cut marks are the best indicators in our study of carcass butchering by late Pleistocene Siberians. Chop marks were found on 9/1618 pieces (0.5% unweighted mean) from the mainly hyena sites, the same percentage as cut pieces. In the two types of archaeological sites, pieces with chop marks are more numerous but fewer than those with cuts - 9.2% and 8.7% unweighted means, respectively. In both cases the difference between them and the mainly hyena sites is statistically significant (x2= 124.3; 1 d. f.; p <0.001; x2 = 144.6; 1 d. f.;p <0.001). There are roughly 20 times more chopped pieces in the two types of archaeological sites than in the mainly hyena group. Statistically speaking, chopping is definitely a perimortem taphonomic indicator of human presence in late Pleistocene Siberia.
Cut pieces are very rare in the mainly hyena group. Eight pieces had cut marks (8/1608). Four came fTom Maly Yaloman, and four from Straschnaya. Both sites had small amounts of archaeological materials, so cut pieces here are not totally unexpected. Both unweighted and weighted means indicate less than a 1.0% average occurrence of pieces with cut marks in our mainly hyena group. This frequency is significantly less that the 12.0% (x2 = 247.7; 1 d. f.;p<0.001) and 12.6% (x2= 150.1; 1.d. f.;p<0.001) unweighted averages (15.1%, 10.1% weighted) for our two classes of archaeological sites. Since some of the mainly hyena sites did have small amounts of archaeological refuse, we can say that their extent or degree of occupation, as measured by cutting of bone, was something like 20-30 times less than in the mixed and no-hyena archaeological groups. Cut marks on bone are the
Preeminent indicator of perimortem carcass processing by humans of late Pleistocene Siberia, and reasonable evidence of human occupation, at least in our study area.
World History
