Some other comparisons

Here we make three more comparisons with published bone damage information, starting in the present and working back to the Upper Paleolithic. There are many such studies and our choices were purely random. We will have a brief discussion for each assemblage that is compared at this point. We have organized our comparisons on the basis of Old World and New World assemblages on the assumption that bone damage will be less in the New World because late Pleistocene cave hyenas seemingly did not expand into North America on account of the frigid conditions of Beringia. Our study found no evidence of hyenas in Siberia north of 55° latitude.

A modern assemblage

Kevin T. Jones (1983) studied the way of life of the South American Ache foraging in eastern Paraguay. One aspect of his interesting work was an analysis of the frequency of damage due to marrow extraction in long bones he found in various camp sites. Based on

231 bones, he did not find marrow extraction damage, or only butchering marks of bones with less than one milliliter volume. All the bones that had 3 ml to more than 5 ml volume had damage caused by marrow extraction. In other words, bone breakage correlated highly with bone size. In our study, which considered breakage in all types of bone (Table A1.10), breakage was very common, even in the sites such as Neolithic Boisman II, where there was no evidence of hyenas. Thus, it would appear that bone breaking for marrow extraction was more commonly practiced by the Siberians, along with additional breakage by large carnivores. We can add that the Siberian breakage was not also done to make bone tools, since almost no bone tools have turned up in the Siberian sites. All this hints at a greater scarcity and difficulty of obtaining food for the Siberians than for the tropical forest-dwelling Ache. The absence of abundant tropical plant foods, including nuts, fruits, and tubers, probably forced the Siberians to waste very little of their largely animal-based nutrients.

A Neolithic assemblage

Thomas F. Kehoe (1983) studied Bos (cattle) bone damage in a German Neolithic site called Schreckensee, which dates about 3000 BCE. The cattle bones were recovered fTom lake deposits. A total of 519 pieces were studied and analyzed for cut and gnaw marks. Presumably dogs were the principle gnawers. A total of 14% showed gnaw marks. There were about three times more bones with cut marks (40%).

Kehoe did not break down gnawing into the various categories that we did, but if we use only our tooth scratches (Table A1.14) for comparison it is obvious that sites we have identified as having a hyena presence have much more gnawing - for example, Denisova (20.9%), Dvuglaska (38.5%), Okladnikov (57.0%). Some of our sites with hyena presence have less gnawing than the German Neolithic site - Kara Bom (10.9%), Ust-Kan (9.0%). However, if we add in our other carnivore damage types such as dints, notching, and hollowing, all our sites with hyena presence exceed the German Neolithic gnawing frequency. Our sites that have no clear-cut evidence of a hyena presence but do exhibit some carnivore damage, such as Bolshoi Yakor I, Kamenka, and Varvarina Gora, were probably damaged by wolves. In this comparison, the gnawing by presumed dogs and presumed wolves is considerably less than that attributed to the jaws of the more powerful hyenas. Because of the low occurrence of bear and wolf bones in our assemblages, we attribute most of the Siberian damage to hyenas. This would have to be the case for the Neolithic Boisman II (1.6% tooth scratches), where dog bones were recovered by Alexander Popov (personal communication, June 9, 2000).

An Upper Paleolithic assemblage

Jean Bouchud (1975) studied the faunal remains fTom the famous French site called Abri Pataud, excavated by Hallam L. Movius, Jr. Abri Pataud carbon-14 dates between roughly 22 000 and 33 000 BP, with various cultural manifestations (Movius 1975, site profile, no page number). Compared to our sites of comparable antiquity such as Denisova, Kaminnaya, and Okladnikov Caves, the difference in animal representation

Is dramatic. The Abri Pataud Perigonian IV horizon had 21119 (97.4%) pieces of reindeer, and only three pieces of cave hyena (<1.0%). Two things stand out: (1) the Abri Pataud hunters were concentrating on reindeer, whereas the Siberians were hunting on a much more opportunistic basis; (2) there is almost no hyena presence at Abri Pataud, whereas hyena presence was common in the Siberian sites. This suggests much more discontinuous occupation of the Siberian archaeological sites because humans would not have been present during hyena utilization of the archaeological sites.

The number of Abri Pataud pieces of human bone and teeth studied by Ginette Billy (1975) is greater than those found in any of the Siberian sites. This difference adds to our view that the Siberia hyenas were likely responsible for the rarity of human remains in late Pleistocene Siberia.

One way to assess the reliability of the intensity of bone damage indicated in Table A1.27 is to compare our findings with assemblages, as just related. However, there may be a substantial difference in methodology and scoring standards between observers, so at this time we limit our inter-observer comparisons and emphasize intra-observer data (Table A1.31). Larger carnivores such as wolves and bears are rare in the American Southwest sampling area, and there are no species with abilities comparable to the powerful bone-crushing ability of hyenas. Therefore, we would expect the North American carnivore damage to be most like that of the Siberian assemblages that lack evidence of hyenas.

As can be seen in Table A1.29, bone accumulations believed to be the result primarily of hyena activity have the greatest frequency (27.5%) of perimortem bone damage, even when only end-hollowing, tooth dints, tooth scratches, and acid erosion are considered. The frequency would be greater if notching and polishing were included. We have excluded them for this inter-regional comparison because both may have been caused to a small degree by humans as well as carnivores. The assemblages that are known to have a mixture of human and hyena presence have about half the amount of damaged pieces (15.2%). The least amount of Siberian carnivore damage (2.0%) occurs in those assemblages where there is no secure evidence of hyenas, only human and small carnivore activity, the latter identified mainly by the small size of dints and shallow, weak scratches. As expected, the amount ofNorth American carnivore damage (0.9%) is much like that of the Siberian material that is considered to have no evidence of hyenas. Although very similar, nevertheless the difference is statistically significant. While the significance could be due to the more concentrated effort to identify carnivore damage in the present study, it seems better to assume that site context and fewer and probably less aggressive North American carnivores are responsible. Thus, the carnivore signature is easily recognized and, importantly, it varies in intensity. This variation in intensity or severity of damage, as these simple comparisons demonstrate, provides additional confidence in the value and usefulness of perimortem bone damage for certain bioarchaeological problems. Problems that immediately come to mind include considerations of site abandonment, variable discontinuity of site use, predator pressure, distinguishing bone damage caused by carnivores and humans, and other narrowly and broadly defined topics. Because our work focuses on late Pleistocene Siberian perimortem conditions, one topic that we are unable to avoid has to do with the possible stratigraphic blurring and

Mixing of Siberian and possibly many Eurasian archaeological sites because of the use of the same sites by hyenas at different times.

Comparisons between our Siberian assemblages show that archaeological and paleontological sites with no actual hyena remains had the lowest amount of these eight kinds of bone damage. Quite the opposite condition was found in those sites with carnivore bones and no or very little archaeological refuse. These two groups were more-or-less expected, which we initially (and still do) referred to as archaeological assemblages and paleontological assemblages. What was unexpected was the intermediate condition of certain archaeological sites where, as it turned out, a substantial carnivore presence was identified. Tables A1.28-A1.30 showthis threefold contextual grouping. The intermediate group is referred to as “mixed” in Tables A1.27-A1.30. Not all of our sites are shown in these tables. Those left out are either Holocene in age, their faunal assemblage is numerically small, or they are not especially relevant to the concern here of defining a late Pleistocene carnivore processing signature.

Table A1.27 shows the six most distinctive carnivore processing stigmata and their average frequencies in the mainly hyena, mixed, and no-hyena groups. These are end-hollowing, tooth dints, tooth scratches, notches, stomach bones, and end-middle polishing. Tooth dints, scratches, and stomach bones are the three most distinctive of the six. It is noteworthy that a few stomach bones turned up in the no-hyena sites. Although few in number, these acid-eroded pieces fTom Bolshoi Yakor, Kamenka, etc., must represent bone fragments swallowed by wolves, foxes, and possibly even the human inhabitants of these sites.

Generally speaking, archaeological sites lacking evidence of hyenas have an unweighted average of 8.1% bone damage attributable to carnivores (the weighted average is only slightly higher at 11.2%). These carnivores, we assume, were wolves, small bears, wolverines, and foxes - that is, medium - to small-sized carnivores. Between four and five times (unweighted average, 38.2%; weighted, 30.8%) more damage occurred in the largely non-archaeological, mainly hyena sites. The mixed archaeological-hyena sites have, on average, three times (unweighted 24.4%; weighted 23.2%) more damage than the archaeological no-hyena sites.

Table A1.28 provides comparisons of the three groups using chi-square goodness-of-fit tests. Most of the comparisons are significant at the p > 0.05 and higher confidence levels. The number and size of the differences provides a parallel pattern to the above remarks. There are more (8/12 comparisons) smaller differences between mainly hyena and mixed than between mainly hyena and no-hyena sites. Simply adding up the chi-square values shows clearly the closer relationship between the hyena and mixed groups (Sx2 = 918.2) than between the hyena and no-hyena groups (Sx2 = 2656.2). As would be expected, the mixed and no-hyena groups add up to an intermediate value (2X2 = 2151). Table A1.28 also shows the summed chi-square values for the human perimortem damage stigmata. The mixed and no hyena groups have a low value (2X2 = 160.3) befitting two sorts of archaeological assemblages. The cultural damage is reversed. The hyena and mixed groups have a greater summed chi-square value (2X2 =446.8) than the hyena and no-hyena groups (2X2 = 320.8), which may reflect differences in butchering and cooking practices that have little or nothing to do with the way we have combined sites for the three-group analysis.

One outstanding characteristic of our archaeological faunal assemblages is the enormous amount of bone breakage associated with humans. This is especially noticeable when contrasted with bone deposits from paleontological sites. Examples of such include the frozen mammoths of far northern Siberia, the frozen Alaskan muck deposits near Fairbanks, the La Brea tar pits in Los Angeles, the multiple-mammoth deposit near Waco, Texas (and other such natural die-offs), or natural trap assemblages. In this study we have bone breakage in paleontological contexts that parallels the archaeological situation. Some sense of the degree of breakage can quickly be obtained by comparing the means and ranges in long bone length given in Tables A1.6 and A1.30. While the measurements combine species and skeletal elements, and do not include the few whole bones we excluded from our study because they lacked perimortem damage, nevertheless nearly all of the species in Table A1.2 are medium to large mammals, most of which have been identified in one or another of our assemblages. By referring back to Table A1.2, a crude idea of species content for each assemblage can be gained, and with that in mind, compared with the relevant species provided in Table A1.27. The point of this simple comparison is to emphasize that late Pleistocene carnivores of Siberia were like humans in the great amount of bone breakage and size reduction they were capable of doing. Where it is possible to identify the species from which a given skeletal element was derived, comparing its maximal dimension shows it to be generally less than 15.0% of the lower end of the range provided in Table A1.30.

Comparisons of the maximal size measurements in Table A1.30 show that the mainly hyena and the mixed groups are metrically very similar. Both have mean dimensions that are less than that of the no-hyena group, indicating slightly less human processing reduction in the latter group. However, some of this lesser processing is attributable to the larger means and ranges of the two no-hyena sites with much mammoth bone - Mal’ta and Shestakovo. This particular inter-group difference would be even larger had it been possible for the archaeologists to recover intact all the big mammoth bone pieces we saw in published site photographs and maps. Both size reduction and type of damage are useful for establishing a carnivore damage signature. However, the kinds of damage are considerably more specific to carnivores than is maximal size, which was produced similarly by both humans and carnivores. Examination of Table A1.31 shows that there is a general pattern of similarity in the frequencies of skeletal elements. A chi-square comparison with 18 d. f. of most of the elements from Razboinich’ya and Okladnikov (excluding “teeth,” long bone fragments, and coprolites because of their very large numbers and/or not being a skeletal element), while producing a statistically significant difference (x2>40; p <0.01), is not as large as might be expected given all the possible sources of variation in these two sites. By itself, this comparison is not especially supportive of our view that Okladnikov was more of a paleontological than an archaeological site. However, a chi-square comparison of hyena coprolite occurrence in Razboinich’ya, Logovo Gieny, and Okladnikov shows no statistically significant difference (x2 = 3.61, 2 d. f., 0.05<p<0.20). This comparison suggests that because the archaeological recovery of coprolites was very similar in the two caves, the actual frequency of coprolites must have been about the same in each of the three caves. This implies that the amount of time spent by hyenas in the three caves must have been about

The same, assuming a similar number of hyena occupants for each cave, a not unreasonable assumption judging from hunting pack size in Kalahari spotted hyenas (Mills 1990:77). So, the general pattern of preserved hyena skeletal elements and the frequency of their coprolites, in addition to the kinds of perimortem damage reviewed elsewhere, leave little room for doubting that Okladnikov Cave was used as much if not more by hyenas than by humans. The implications for stratigraphic disturbance are rather obvious.

The lengths of undamaged long bones published by Vera Gromova (1950) are provided in the site descriptions in Chapter 3. Her measurements are included to help the reader appreciate the marked degree to which perimortem bone breakage had reduced whole long bones to very small fractions of their original size. Similar fragmentation and size reduction occurred to all other parts of the thousands of skeletons represented in this study. At the end of each assemblage’s description there is a brief discussion of our main inferences about the perimortem taphonomy of the site. Much of our description may be of interest only to the small number of specialists in Siberian taphonomy. For this reason, the reader can skim or even ignore most of the descriptions, and concentrate on our more general discussions of each site. However, for the reader who focuses on the discussions and is uncomfortable with our inferences, we urge him or her to go back and read the descriptions of the relevant variables.

We propose that the Siberian late Pleistocene no-hyena baseline signature of carnivore activity is an average of about 10% damage, as recorded by the six traits listed in Table A1.27. This baseline signature would represent small - to medium-sized carnivores. When larger carnivores, especially hyenas, were present in a given region, the damage signature increases fourfold to 38.2% (weighted, 30.8%) in paleontological sites, and threefold (24.4%; weighted, 23.2%) in archaeological sites. Said another way, the amount of presumed hyena bone-processing activity in the mixed archaeology and hyena sites is more like that found in hyena sites than in non-hyena archaeological sites. We view this amount of hyena presence in these mixed archaeology sites to be potentially of considerable interpretative importance for three main reasons.

First, as discussed preliminarily in Chapter 3, the presence of hyena remains in an archaeological site should sound an alarm about the likelihood of substantial stratigraphic disturbances caused by the digging behavior of these animals (see Brain’s 1981 discussion of hyena dens, pp. 57-58). The nature of the disturbance will to a large degree be dependent on how the archaeological site was utilized by humans. Ifcave occupation was seasonal, say, in the summer when there was less need for warmth-producing fires, then hyenas could prowl, scavenge, and dig in the human midden for maybe half the year. This would certainly be enough time for even a few animals to annually churn the human refuse into a random mixture, bringing to the surface artifacts and refuse of earlier times and falsely associating older remains with “recent” objects. The stratigraphic effect of carnivore digging would be to blur the boundary separating earlier autochthonous cultural remains and artifacts left later by incoming Upper Paleolithic Europeans, as exemplified by Mal’ta. Movements into Siberia by external alien bands most probably did occur, and probably repeatedly over the millennia due to fluctuations in the more easterly Eurasian population density, pressure, and response to natural catastrophes such as large-scale earthquakes, droughts, forest fires, epidemics,

Conflict, and so forth. Site formation processes are not limited to just within a site itself. Instead of inferring “replacement” or “discontinuity,” undetected hyena disturbance of midden would more likely be interpreted as “gradual local evolution.” Indeed, this seems to be the prevailing Siberian archaeological view (flake industry evolving into blade industry: Derevianko 2001, 2005) based on the same sites we have identified as having a substantial mixture of human and hyena occupation. Without an evaluation of the extent of bioturbation, the argument is potentially unsupportable for local tool type evolution, and cultural continuity between the Middle and Upper Paleolithic.

Second, the senior author has had six years of excavation experience with highly stratified prehistoric village middens in the Aleutian Islands. Upon visiting the Siberian mixed archaeology-hyena cave sites he could distinguish no Pleistocene stratigraphic levels comparable to the hundreds of easily discerned layers in Aleutian archaeological sites, or, for example, in the very distinct Holocene (Neolithic) layers of Denisova Cave. Hyenas were extinct by the end of the Pleistocene, so the Holocene Denisova strata were unaffected by their digging behavior. In the Aleutians, there were no terrestrial carnivores larger than foxes, and Aleut settlements seem to have been continuously occupied for thousands of years. Hence, foxes did not den in the Aleutian village middens, and even minutely thin strata remained distinct, pristine, and readily recognizable. The excavation of Okladnikov Cave was nearly completed in 1987 when Sergei Markin led the senior author to see the last bit of deposit remaining deep in the cave. Even that small section seemed odd, as mentioned in Chapter 3. At Kaminnaya Cave, he again could see no distinct Pleistocene strata, even with the detailed explanation that Markin provided about the stratigraphy. Ust-Kan was the same story. While Alexander Postnov pointed out where he thought his levels were, clear and distinct banding was not apparent. At the time of these visits, the senior author’s inability to recognize distinct strata was attributed to inexperience with Pleistocene archaeology. Now, there may be a better explanation, which is bioturbation caused by carnivore disturbance and consequential stratigraphic blurring.

Third, the amount of carnivore perimortem damage should correspond positively with the amount of site usage by these animals in the same manner as the amount of human refuse reflects occupation duration and intensity in much of the world. In this respect, then, Denisova, Kaminnaya, Kara-Bom, Okladnikov, and Ust-Kan are more like the six hyena caves than like the eight no-hyena sites. Okladnikov Cave stands out as having as much of a carnivore perimortem bone damage signature (37.3%) as the unweighted average for the six hyena sites combined (38.2%) (Table A1.27). Had Okladnikov Cave produced only a handful of stone or bone artifacts, we would have no reluctance to classify it as a paleontological site on the grounds of perimortem damage. If we choose to ignore the possibility of midden disturbance by hyenas, we cannot ignore the fact that these creatures and probably other species of digging animals occupied these caves in the absence of competing humans. In other words, human use of the cave and open sites was almost certainly seasonal, sporadic, or decades apart. We are not alone in this view. Wrinn (2010) concluded that there was so much “carnivore ravaging of bone assemblages, [which] suggests a competitive situation for human foragers, and the overall pattern is consistent with ephemeral occupations and low population density.”

There had to have been discontinuities in the human occupation of these Siberian sites, the possible reasons for which are numerous - local reduction of food supplies, epidemics in game animals and/or humans, carnivore predation on the humans, competition with other human groups, exceptionally severe winters, forest and steppe wildfires, human infertility, and so forth. These breaks could have been seasonal, decades, centuries, or millennia in duration. Said another way, the longer the periods of absence, the greater the opportunity for short - or long-distance introduction of new cultural and human genetic variation. We suspect that there were many introductions of “novelty” caused by fluctuating human population pressure in southern and eastern Eurasia. We envision late Pleistocene Siberian winter conditions, especially the Late Glacial Maximum, as a significant obstacle to continuous local human occupation. The marked need for winter warmth most likely drove humans into the fuel-rich and wind-protected forests, and out of the frigid, barren caves, and possibly even out of the region by making north to south fall migratory treks of hundreds of kilometers, leaving the caves to the over-wintering hyenas and other animals. One needs look no further for evidence of long-distance migratory treks than ethnographic accounts of annual migrations by Alaskan and Canadian Eskimo bands across vast tracts of land.

In sum, our observations and analyses ofperimortem bone damage have the potential to make significant contributions to the study of archaeological site formation and content. Based on our analyses we proposed the following. (1) Archaeological sites without a hyena presence have a much lower level of carnivore damage. (2) In sites with mixed human and hyena content, there were periods of human site disuse. (3) Disturbance of stratigraphy and probable blurring of strata by hyena and other animal digging can lead to false inferences about site and regional cultural evolution. (4) Comparison of paleontological and archaeological faunal assemblages shows that when the latter indicate the presence of hyenas, these sites have perimortem damage signatures closer to those of paleontological than archaeological sites. (5) By implication, the presence of some animal remains in mixed human-hyena sites may have little to do with human procurement.

We make some of the above remarks in consideration of ethological information in the following literature review of modern carnivores.