Findings

1 Provenience. Bone was initially saved as coming from upper, lower, and mixed levels. Analysis showed no meaningful differences between upper and lower levels, so all three are presented here as a single assemblage of205 pieces. This number is 2.3% of our study total (Table A1.1, site 6). We ignored the bat, bird, and rodent pieces, but measured the maximal lengths of an additional 220 pieces whose only perimortem information was size.

2  Species. Out of the 205 Dvuglaska pieces, only big mammals (4.9%), bison (2.4%), horse (0.5%), and rhinoceros (0.1%) could be identified (Table A1.2, site 6). While we are confident that Dvuglaska Cave was mainly occupied by hyenas (chewing damage, stomach bones), we recovered no identifiable bones or teeth of this species. Compared with the pooled assemblage averages, Dvuglaska is similar except for the absence of bear, goat-sheep, hyena (bones/teeth), mammoth, and roe deer, and it has a lower frequency of horse.

3  Skeletal elements. Our Dvuglaska sample of 205 pieces contained mainly long bones (47.8%), unknown elements (25.4%), ribs (7.3%), and metapodials (5.9%) (Table A1.3, site 6). Compared with the pooled assemblage averages, Dvuglaska is similar except that it has fewer mandibles, vertebrae, humeri, toes, and more long bones and unknown elements.

4  Age. Of the 205 Dvuglaska pieces, there are almost no identifiable sub-adults (0.5%) (Table A1.4, site 6). Slightly more than half (56.1%) were assigned adult status. Compared with the pooled assemblage averages, both sub-adult and adult pieces are fewer in the Dvuglaska sample.

5  Completeness. Our Dvuglaska sample of205 pieces has only 3.4% whole bones, and a very large number of pieces with no anatomical ends (91.7%) (Table A1.5, site 6). Compared with the pooled assemblage, Dvuglaska has three times fewer whole bones, and 32.5% more pieces with no anatomical ends. The large amount of incompleteness underlies the difficulty in species, skeletal element, and age identification. Dvuglaska even exceeds Razboinich’ya Cave in its amount and degree of incompleteness.

6  Maximum size. Overall, 429 Dvuglaska pieces were measured (Table A1.6, site 6). Incompleteness also influences size. The average maximum dimension is only 5.2 cm, whereas the few whole bones and large pieces extend the range fTom 1.7 cm to 28.2 cm. All the Dvuglaska statistics are lower than the pooled assemblage averages. Comparing the upper range limit with the lengths of whole long bones of bison and horse provided by Vera Gromova (1950) shows that Dvuglaska pieces are considerably smaller.

7  Damage shape. Out of 205 Dvuglaska pieces (Table A1.7, site 6), the most common damage forms are flakes (62.9%), splinters (15.1%), fragments (6.4%), and medial rib pieces (5.9%). Viewed against the pooled assemblage averages, Dvuglaska has more flakes and splinters, and fewer fragments, undamaged, and mostly whole pieces.

8  Color. Dvuglaska has 205 pieces, their color being almost entirely ivory (98.0%) (Table A1.8, site 6). There is only one black piece (0.5%), which represents burning, and one brown piece that may have been burned. Compared with the pooled assemblage averages, Dvuglaska has considerably more ivory colored pieces, and many fewer brown pieces.

9  Preservation. Table A1.9 (site 6) shows that Dvuglaska’s 205 pieces are mostly ivory hard (94.1%), with only 5.9% being chalky (Table A1.9, site 6). Compared with the pooled assemblage averages, Dvuglaska has more ivory, and fewer chalky pieces. The quality of the Dvuglaska assemblage is almost exactly the same as that of Razboinich’ya Cave, suggesting similar preservational conditions shared by both caves, even though they are geomorphically very different, as well as having very different temperatures.

10  Perimortem breakage. Almost every one (99.5%) of the 204 scorable Dvuglaska pieces has perimortem breakage (Table A1.10, site 6). This amount of perimortem breakage is more than that found in the pooled assemblage average, and even more than what was found in Razboinich’ya (92.4%).

11  Postmortem breakage. Out of 204 Dvuglaska pieces, 10.3% have postmortem breakage (Table A1.U, site 6). Compared with the pooled assemblage average (17.8%), Dvuglaska has somewhat less postmortem breakage. Dvuglaska and Razboinich’ya (10.9%) are almost exactly the same.

12  End-hollowing. Dvuglaska has 4.4% (9/ 205) end-hollowing (Table A1.12, site 6). This value is less than the pooled assemblage average and the Razboinich’ya frequency (8.1%).

13  Notching. Out of205 Dvuglaska pieces, 15.1%havenotching(TableA1.13,site6). Of these, most have one (8.8%) or two (3.4%) notches. The maximum number of notches is five. Compared with the pooled assemblage, Dvuglaska is almost exactly the same in the occurrence of notching, as well as being very similar in terms of the number of notches. Razboinich’ya has more notched pieces (23.6%).

14  Tooth scratches. More than one-third (38.5%) of the 205 Dvuglaska pieces have tooth scratching (Table A1.14, site 6). The largest percentage (10.7%) of these have only one scratch, but 8.0% have more than seven scratches. Compared with the pooled assemblage, Dvuglaska has almost twice as many pieces with tooth scratches. Dvuglaska is extremely similar to Razboinich’ya (37.1%) in its frequency of tooth scratches.

15  Tooth dints. Almost half (43.4%) of the 205 Dvuglaska pieces have one or more tooth dints (Table A1.15, site 6). Of these, they range fTom having one dint (9.8%) to having more than seven dints (9.9%). There are considerably fewer occurrences of tooth dinting in the pooled assemblage. Razboinich’ya Cave has about 10% more cases (52.8%).

16  Pseudo-cuts. There are only a few examples (2.9%) of pseudo-cuts in the 205-piece Dvuglaska assemblage (Table A1.16, site 6). Most of these (2.0%) have only one pseudocut. The maximum number of cuts per piece is five. The occurrence of pseudo-cuts in the pooled assemblage is slightly greater. Razboinich’ya is very similar (3.1%) to Dvuglaska.

17  Abrasions. There are no examples of abrasions on any of the 205 Dvuglaska pieces (Table A1.17, site 6). This finding adds to our belief that abrasions are produced mainly by human bone processing.

18  Polishing. Most of the 203 scorable Dvuglaska pieces have polishing; only 5.4% do not (Table A1.18, site 6). Polishing occurs on one or both ends (9.4%), the middle (0.5%), and most often on both an end and the middle (84.7%). The absence of polishing in the pooled assemblage average is much greater. Razboinich’ya has far more unpolished pieces (40.6%) than Dvuglaska.

19  Embedded fragments. There are very few (1.0%) of the 205 Dvuglaska pieces with an embedded fragment (Table A1.19, site 6). This frequency is less than the pooled assemblage average, and very much lower than the Razboinich’ya frequency (8.8%).

20  There are no teeth in the Dvuglaska assemblage (Table A1.20, site 6).

21  Acid erosion. Pieces with acid erosion are common (22.1%; 45/204) in the Dvuglaska Cave assemblage (Table A1.21, site 6). In addition to other variables indicating intensive carnivore bone processing, the many acid-eroded pieces point to Dvuglaska having been used by hyenas much more often than by humans, as assessed by the very few artifacts found by Abramova (we found two, a polished splinter and an eyed needle). Compared with the pooled assemblage average, Dvuglaska has more than three times the amount of acid erosion. The same can be said for Razboinich’ya (7.3%).

22-26 Rodent gnawing, insect damage, human bone, cut marks, and chop marks. There are no examples ofthese variables in the 205 Dvuglaska pieces (Tables A1.22-A1.26, site 6).