Divje babe I - poskus uporabe statistične analize množičnih živalskih ostankov v paleolitski arheologiji. I. Določljivi skeletni ostanki jamskega medveda

Povzetek

For the purpose of the analysis of skeletal elements, samples have been gathered from the sediments of specified surfaces (10 m2) and cubic contents (depending on the thickness of arbitrary stratigraphic units, with a range of 0.15 m) which, with minor aberrations, run in succession, one under another, to the unit (14) inclusive. Downward from this unit there comes a stratigraphic sequence of samples of almost the same surfaces and cubic contents but that it had been dislocated (Fig. 1). The sterile units (15-21) between the two sequences form a blank. A sample consisting exclusively of 70.5 m3 of gravel, sand and silt deposits has been investigated in its entirety by the present authors alone. In order to give the treatment as much uniformity as possible, all basic statistical forms comprising units of 1 m- each (Fig. 2) were filled in by one author alone according to previously fixed criteria. The samples include all taxonomically and anatomically determinable remains of bones, both integral and damaged, that have been found independently of the screening of sediments. All skeletal elements have been divided with morphological reliability into two groups of remains, one of juvenile and the other of adult specimens, yet with no division into the left and the right (Tt. 10-12). All anatomically determinable skeletal elements (labelled »n« in Tt. 10-11) have been loaded subsequently on account of the irregularity of their representation within the skeleton. The 'oaded values have not been entered. By use of different loadings (as designated in the leftmost column of Tt. 10 and 11) individual bones, as well as the remains of both age profiles, have been balanced out. The loaded data have then been sorted out into five classes (labelled »r« in Tt- 10-11) after D.R. Spennemann's method (Spennemann D R., 1985). In this way an adequate general account of skeletal remains has been achieved (Figs. 3; 5). The absolute figures from Tt. 10 and 11 (»n« values) were submitted to the two-sided non-parametric tests used for independent samples regarding the differences of the samples. Wi'coxon and Kruskal-Wallis's tests were applied for the purpose of comparing excavation units, and Mann-Whitney's U-test for the purpose of comparing individual skeletal elements from both archaeologically sterile and archaeologically fertile units (Mosteller F., Rourke R. K K., 1973; Weber E., 1972). The correlations have been calculated on the basis of the coefficient of rank correlation after Spearmann, and then put to the t-Test. By reason of the great significance of the relative relations between the arbitrary stratigraphic units and both age profiles, the overlooking diagram of the sums total of all skeletal remains as found in each of the units, with juvenile and adult specimens shown separately, has been produced in the ogarithmic scale (Fig. 4). The correlation between the units (1-30), cultural levels (A-H), and 'ay«rs (2-21) is presented in T. 1. There are two main factors apt to affect the outcome of the statistical analysis, namely the manner of sampling (excavation) and the size of samples. Since it is impossible to make an entry of each skeletal fragment in the coordinate system for practical reasons, the vertical - representing the dimension of time - may have a tendency to exhibit remarkable divergencies between the curve representing the actual (yet unknown) distribution of all skeletal remains and the curve representing the empirical distribution that had been arrived at in consequence of our adopted manner of sampling by sq. metres and by excavation units. In order to avoid further departure from the actual distribution of remains in the vertical, the analysis of samples was performed in strict adherence to the units, with deliberate omission of combining the units to form sedimentary and cultural units. Skeletal remains have been found to be of a fairly regular distribution throughout all the plan-view units, whether the latter be 1 or 10 m2 in surface. There are no major groupings, either anatomical or chaotic. Occasioned by the presence of individual rocks and a number of unpredictable modifications in the configuration of the cave, there is no complete overlap between the plane views of all the units. The percentage of overlapping between the units is clearly observable from T. 1. Allowing that the different thicknesses of the units (see T. 1) and the different intervals of time during which the samples had originated should not be taken into account, and judging from the flora, fauna, and the archaeological finds (all yet unpublished), the units conjointly represent a relatively homogeneous whole, which has been dated in absolute terms for a range of approximately 25.000 years (data not yet published). A discernible secondary perturbation of sediments had occurred only in units (2)—(5) which had been affected by cryoturbation. Bioturbation is traceable in all units. Since it has more or less synchronized with sedimentation, it has no substantial post-sedimentary effect on the homogeneity of the samples. These are taken to be statistically representative when all parts of the skeleton are found to be present in them. This is the case when a sample contains at least 500 skeletal elements. When not, a number of blanks appear, which has the effect of upsetting the statistics (see Tt. 10-11). This is the reason why only rare samples (in semibold type in T. 1) are representative in statistical terms. An increase in the extent of samples will, above all, tend to increase the number of the remains which have been most numerous in the first place, whereas the extremely rare parts of the skeleton will only rarely present themselves, which seems to fit in with the Neyman series (Weber E., 1972). The 8.90 m thick stratigraphic section displays considerable oscillation in the amount of skeletal remains (Fig. 4). The upper section of the stratigraphic series reveals several minor cycles of oscillation within one, or perhaps two periods, while the lower section exhibits a single period with two eminent culminations. Periodic oscillations indicate a slow process of alterations and stable conditions, while the cyclic aberrations in the upper part of the curve may be derived from rapid changes and unstable conditions, which is in perfect agreement with not entirely published results of pollen analyses (Turk I. et al., 1988). Extensive oscillations in amount stand in covariance with the archaeological remains (T. 5). Nevertheless, the correlation has been found to be significant only in the case of the remains of the adult specimens (r, = 0.858; a = 0.017). The amount of the skeletal remains of juvenile specimens has, in all probability, been conditioned by the use of the den - whether it was one of females with their cubs or one of solitary males - which was submitted to constant changes, which again, in the long run, seems to have occurred in connection with ecological changes. The frequency of occurrence of these remains seems to be in no relation with the repeated appearance of middle palaeolithic men in Divje babe I. There are no significant differences between the skeletal elements of the juvenile, and those of the adult specimens in the statistically representative samples (Kruskal- Wallis, P(H 5 12.00) = 0.10) (T. 4). The comparison between individual archaeologically fertile units containing statistically representative samples has yielded no such differences between individual units as it was expected (Tt. 6; 7). What is even more surprising, there are no differences in quality between the units of D- and those of E-level the two of which differ significantly in the amount of archaeological finds, and possibly also in the frequentation of middle palaeolithic hunters and food-gatherers (T. 7). This absence of differences may have resulted from the insufficiency of the amounts of individual skeletal elements in the samples. Due to the occurrence of substantial differences in amount between the archaeologically sterile units (2), (6), (24)-(26), (28), and the archaeologically fertile units (3)-(5), (7)-(14), (22), (23), (27), (29), (30) with their levels A, B, D, E, F, G, H, each part of the skeleton was submitted separately to Mann-Whitney's U-test, which is equivalent to Wilcoxon's test. The results of the two-sided test are exhibited in T. 9, and the results of the one-sided test in footnote 7. Statistically significant disproportions have been established between individual skeletal elements of both juvenile and adult specimens, which seems to be best accounted for by the activity of man and predatory animals. Most of the bones of the juvenile specimens may have been displaced by predators, while most of the long marrow bones and crania of the adult specimens seem to have been reduced to minute fragments until unrecognizable by man, in view of the total absence of the cave hyena. This is one of the possible explanations of the surplus number of some bones and the deficiency of others in units otherwise rich in cultural remains and features (Fig. 6). Since the entire "bone evidence found at Divje babe I, which includes approximately seven times the amount of the one presented here, produced no more than one or two cutmarks on bones due to butchery and the cutting away of flesh (Fig. 7), the only hypothesis that can justifiably be proposed at this point of investigation is one of the utilization of bone marrow for nutritional and other requirements. There is no evidence of whatever utilization of fleshy parts and skins by ancient men, however. In statistical terms, there is strong evidence suggesting a highly probable link between the occurence of middle palaeolithic man and that of the cave bear "at the site of Divje babe I. There are several conjecturable ways of accounting for such a link. For instance: hostility in the form of hunting the effectiveness of which, we believe, is highly improbable. Or: cumulative effects of a large number of separate visits to the cave by one and the other in terms of mutual avoidance, which merged eventually into a homogeneous integral whole. Or: confrontation and mutual tolerance owing to the occasional partial overlap of their trophic and habitat niches. Or: a combination of all three possibilities. The choice between the four admits of a minimal chance to select the correct explanation for the nature of such a link.