Findings

1 Provenience. In 1999 and 2000 Olga Pavlova and the senior author studied 840 pieces from collections made throughout most of the midden areas of Boisman II, excavated prior to

Those two years. These materials were curated in the Natural History Museum of the Far Eastern State University, Vladivostok. We found no meaningful stratigraphic or spatial differences except those related to pockets of shell, localized burials, and other minor cultural-deposifional events well known for coastal shell middens in the north Pacific, if not worldwide. The pooled Boisman II assemblage total is 9.5% of our grand total of 8813 pieces (see Table A1.1, site 2). Numerically, Boisman II overwhelms some of our smaller Pleistocene assemblages. We try to compensate formally and informally for these dissimilar sample sizes in our various interpretations and inferences.

2  Species. The most common classes identified among 840 Boisman II pieces include indeterminable (59.9%), big mammal (10.4%), Canis (8.2%), and small mammal (4.4%) (see Table A1.2, site 2). Compared with the pooled assemblage averages, Boisman II has more big mammal and indeterminable, and fewer goat-sheep and all mammals that went extinct at the end of the Pleistocene. A faunal analysis of Boisman II that will include birds, fish, and more specific identifications is in progress by David Yesner and students.

3  Skeletal elements. Of 1242 Boisman II pieces classified, the most common are unknown (50.7%), long bone (8.9%), vertebrae (6.7%), foot (5.4%), and mandible (4.5%) (see Table A1.3, site 2). Compared with the pooled assemblage, Boisman II has fewer ribs, long bones, and toes, and more unknown pieces. All in all, there are many more similarities than significant differences between Boisman II and the pooled assemblage. Given the marked ecological difference between these two data sets, should we expect to see more or less general difference in skeletal elements? And if we do not see much difference, is this because of identification potential, preservation potential, or utilization?

4  Age. Boisman II has 7.6% sub-adults (39/513) (Table A1.4, site 2), which is the same as the pooled assemblage.

5  Completeness. Out of 501 Boisman II pieces, there are 13.8% whole bones, 37.5% with one anatomical end, and 48.7% with no anatomical ends (Table A1.5, site 2). The Boisman II pieces are slightly more complete than those of the pooled assemblage.

6  Maximum size. The mean maximum Boisman II piece size is 7.4 cm, and the range is 2.5-22.5 cm (Table A1.6, site 2). These values are slightly less than those of the pooled assemblage. Taken together with completeness, maximum size suggests that average animal size was smaller in Boisman II than in most of the other assemblages, even excluding the very big forms such as mammoth, rhinoceros, and others. Comparing the maximum size values with undamaged long bone lengths provided by Vera Gromova (1950: table 27) shows that the Boisman II upper range limit is similar to that of boars and wolves, but less than the long bone lengths of larger forms similarly expected in the Primorsky region (bear, elk, others).

7  Damage shape. This variable was not scored in Boisman II as it was still being formulated at the time of our studies in Vladivostok.

8  Color. There is only a small number of black pieces (burning) in the Boisman II assemblage (7 /1247; 0.6%) (Table A1.8, site 2). Most pieces are ivory colored (99.1%). Compared with the pooled assemblage averages, burning is nearly the same, but there are fewer ivory colored pieces in the pooled assemblage. Elsewhere it will be shown that brown colored pieces are fairly common in open sites. Since Boisman II is an open site, the difference in the color ratio is attributable to antiquity, soil composition, or both.

9  Preservation. Out of the 513 Boisman II pieces scored for quality, 98.0% were ivory hard and only 1.9% were chalky (Table A1.9, site 2). For an open site, this high-quality condition may be due to a combination of modest site antiquity and the buffering of soil acid by the presence of calcium carbonate-rich invertebrate shells in the midden. Also, rapid burial aids preservation. Compared with the pooled assemblage average for ivory quality, Boisman II is better.

10  Perimortem breakage. Boisman II has 81.8% perimortem breakage in 1022 pieces (Table A1.10, site 2). This is only slightly less than the pooled assemblage average, suggesting substantial similarity in bone reduction behavior by humans and carnivores through time and under widely different environmental conditions.

11  Postmortem breakage. Out of 1099 Boisman II pieces, postmortem breakage occurs in 8.2% of the assemblage (Table A1.11, site 2). This is about half of the pooled assemblage average. We have no idea what lies behind this difference except that rapid burial may be involved.

12  End-hollowing. Inasmuch as domestic dogs were found at Boisman II, and they are ethnographically ubiquitous in Siberia, it is not surprising that end-hollowing occurs in Boisman II. It does so in 3.2% of the 1232 scorable pieces (Table A1.12, site 2). Compared with the pooled assemblage average, Boisman II has about three times less end-hollowing. Again, rapid burial comes to mind.

Boisman II matches the hyena-free open site of Ust-Kova (site 26, 3.5%) for end-hollowing frequency. Because of the dog-wolf size similarity, in considerable contrast to the larger body weights and jaw sizes of hyenas, bears, and lions, we are strongly tempted to suggest that the frequency of end-hollowing might help identify relative size, and thus the species of the carnivore(s) involved. We have no idea how to factor in considerations like hunger and play.

13  Notching. Notching was scored in 541 Boisman II pieces. It occurs in only 2.6% of the assemblage (Table A1.13, site 2). Most of these 14 pieces have only one notch. Compared with the pooled assemblage average, Boisman II has fewer notched pieces. Again, the dog-wolf size similarity comes to mind, especially when much more notching is associated with the more powerful hyenas, as identified on multiple grounds at Razboinich’ya (site 21, 23.6%) and Maly Yaloman (site 16, 23.0%) caves.

14  Tooth scratches. Out of 1235 Boisman II pieces, tooth scratching is present in only 1.6% (Table A1.14, site 2). The number of scratches ranges between one and six per piece. Compared with the pooled assemblage average, Boisman II has fewer scratched

Pieces. To turn again to the dog-wolf vs. hyena comparison, Razboinich’ya and Maly Yaloman have 37.1% and 28.6% of pieces with scratching, respectively.

15  Tooth dints. There are 8.2% of 512 Boisman II pieces with one to more than seven dints per dinted piece (Table A1.15, site 2). Compared with the pooled assemblage average, this shows three times less dinting in the Boisman II assemblage, as well as much less than in Razboinich’ya and Maly Yaloman.

16  Pseudo-cuts. There are only two Boisman II pieces with pseudo-cuts (0.4%) (Table A1.16, site 2). This is several times less than the pooled assemblage average, as well as less than what Razboinich’ya (3.1%) and Maly Yaloman possess (6.3%). These sites were hyena dens.

17  Abrasions. The Boisman II assemblage has 1.8% of pieces with abrasions (9/512 pieces) (Table A1.17, site 2). This value is slightly greater than the pooled assemblage average. Since abrasions are uncommon in our study, this difference is probably meaningless, although in all assemblages bone breaking seems to have been done far less often with abrasive stone than with bone, horn, or wood implements.

18  Polishing. Boisman II possesses 14.2% end, 0.8% middle, and 1.6% both end and middle polishing in 485 scored pieces (Table A1.18, site 2). Compared with the pooled assemblage averages, it is clear that there is significantly less polishing in Boisman II. Polishing is much more common in Razboinich’ ya and Maly Yaloman caves.

19  Embedded fragments. Out of the 513 Boisman II pieces, 4.7% possess 1-6 embedded fragments (Table A1.19, site 2). This frequency is nearly the same as the pooled assemblage average. Razboinich’ya, but not Maly Yaloman, has more pieces with embedded fragments.

20  Tooth wear. Boisman II has the largest number (38) of teeth for assessing wear in our study (Table A1.20, site 2). Using grades 0 and 0-1 (no wear and wear present but dentine not exposed) as criteria for defining “young” individuals, the Boisman II dental wear group suggests that three-quarters (76.3%) were young. This is twice the number of young individuals in the pooled assemblage. We are generally disappointed with the tooth wear variable because of its overall small sample size (293 pieces), which means statistical inference is weak. Nevertheless, young animals suggest summer or fall kills. In Aleutian shell mounds, sea mammal tooth wear is rare except for sea otters.

21  Acid erosion. Boisman II has no examples of acid erosion in 553 scored pieces (Table A1.21, site 2). Since we have long suspected acid erosion to be associated primarily with hyenas, Boisman II offers a useful comparative sample to support this view due to the lack of hyenas and the presence of middle-sized carnivores (dogs) in the site. The amount of acid erosion that we could possibly attribute to wolf digestion at Bolshoi Yakor I, Kamenka, Varvarina Gora, and other sites, is very low to absent. Hence, Boisman II, despite its Holocene age, nicely provides comparative information for differentiating between Siberian wolf and hyena perimortem taphonomy where acid erosion is a critical variable.

22  Rodent gnawing. The largest amount of rodent gnawing in our study occurs in the 553 scored Boisman II pieces (Table A1.22, site 2). This value is much greater than the pooled assemblage average. The only other site with this magnitude of occurrence is found in Krasny Yar (4.9%), a naturally deposited riverside accumulation of large and small mammal remains. We have no satisfactory explanation why these two sites stand out so against a background of general absence of rodent gnawing in our study. Perhaps in both localities the bones remained upon the ground surface longer than at sites with no rodent gnawing, or possibly both site localities had larger rodent populations, thus increasing the chance for occasional gnawing. Porcupines are well-known gnawers of human bone in tree burials on the northwest coast of British Columbia. Ovodov has identified at least one porcupine in Pleistocene Siberia, so these creatures cannot be excluded from other rodents.

23  Insect damage. Like most of the assemblages we investigated, none of the 513 Boisman II pieces showed any sign of insect damage or unusual damage that might be attributed to insects (Table A1.23, site 2).

24  Human bone. The human skeletal remains from intentional burials were removed from the faunal collection and sent to physical anthropologist Tatiana Chicksheva at IAE, Novosibirsk. There, the senior author examined most of the remains. No perimortem damage was identified. Morphometric analysis of the human remains can be found in Popov et al. (1997).

25  Cut marks (Figs. 3.3-3.6). Cut marks were plentiful (21.3%) on 553 Boisman II pieces (Table A1.25, site 2). While the number of cut marks per piece varied from one to more than seven, most common was one cut per piece (6.1%). Compared with the pooled assemblage average, Boisman II has three times more pieces with cutting. There is nothing much to be made of this difference since some of the pooled assemblage sites are mainly paleontological, not archaeological.

26  Chop marks. Out of 513 Boisman II pieces, 44 have chop marks (8.6%) (Table A1.26, site 2). As with cutting, the number of chop marks per piece varies fTom one to more than seven, slightly skewed toward lower numbers of chops. The frequency of Boisman II chop marks differs only slightly fTom the pooled assemblage average.