Abstract:
In the dissertation presented, "Biometric investigation of the stylopodium, zygopodium and metapodium of pleistocene ursids with regard to cave-bear evolution and classification of material from the Einhornhöhle (Unicorn Cave) / Harz", data were evaluated from the upper pliocene Ursus etruscus, the pleistocene Ursus deningeri and also from Ursus spelaeus. The data were gathered using 249 particular measurements on the skeletal elements humerus, ulna, radius, metacarpalia I-V, femur, tibia, fibula and metatarsalia I-V. The 2,977 fossil samples examined were from 21 sites in central Europe and exclusively from adult individuals. Because of the small number of samples the fossil material from Ursus etruscus has not been taken into consideration in the presented statistical analysis. The data from 2,890 deningeri und spelaeus specimens were analysed using uni- and multivariate methods. The analyses have two goals. One is to discover species-defining characteristics. The other is to classify the ursid finds from the Unicorn Cave / Scharzfeld / Harz as belonging to both species, Ursus deningeri or Ursus spelaeus. A lack of accurate detailed stratigraphic information regarding the larger part of the 595 findings has lead to disagreement concerning individual species affiliation. This is especially true of the material from the older excavations. It has not so far been possible to ascertain the absolute geological age of the fossils found.
For each of the 16 postcranial skeletal elements to be analysed between 8 and 25 variables and up to 56 cases per site are present (altogether 249 individual measurements). Width and depth of the joint surfaces were introduced as new sections of measurement. The statistical evaluation was partly univariate (t-test, Welch’s t-test and Mann-Whitney U-test ) and partly multivariate (regression analysis, discriminant analysis, correlation analysis, factor analysis). The data from each species were examined in their entirety without subgroups such as gender separation – no particular measurement or combination thereof allowed a conclusive gender determination. Significant differences between Ursus deningeri and Ursus spelaeus were found univariate in 26 of the 249 variables. The diagnostic characteristics thus found were tested for their correlation with geological age. In 36 variables from Ursus deningeri and 31 from Ursus spelaeus a positive linear regression was found; eight of the sections of measurement showed congruence in the parameters. In the examined timeframe from 900,000 to 30,000 years ago these values decreased. This is so far remarkable, in that the mean values of the data sets analysed show that Ursus spelaeus was generally larger than Ursus deningeri. There is an apparent negative correlation between the values measured and the relative size of the ursids: an allometric phenomenon. Ursus deningeri for example shows a shortening in the largest length values combined with further diminishments of characteristics mainly of the metapodial bones, whereas Ursus spelaeus shows similar changes mainly of the stylopodium. A concurrence of the trends found by regression analysis with the 26 significant species-separating attributes could be determined in two characteristics of Ursus deningeri; one the distal width of the diaphysis of the tibia (BDd), the other the smallest epicondylus width of the metatarsalia V (SBdE). The larger size of Ursus spelaeus compared with Ursus deningeri is generally considered to be associated with an increased herbivory. This is supported by the generally different proportions of proximal and distal skeletal elements of the two ursid species. The longer metapodial bones of Ursus spelaeus combined with the shortened stylopodium appear to be associated with increased wandering activity over large distances. The lengthening of the zygopodium of the front extremity in relation to the humerus, observed during the course of this investigation, also indicates an increased cursoriality. This could be connected with the herbivorous diet, possibly also with particular plant species.
Forty-three specific combinations of variables were identified as capable of being used to distinguish Ursus deningeri from Ursus spelaeus. Twelve of these exhibit a statistical probability of 100% for correct classification. These and other combinations of variables were tested by factor analysis for their discriminatory power, which allowed factors to be traced which may have been responsible for the differences measured, in effect for the evolutionary development of the ursid species. Here the correlation analysis functioned as a preliminary screening, to as far as possible eliminate variables with identical information content. Fifteen of the sixteen skeletal elements proved eminently suitable, in the discriminant analysis. Five sets of variables on humerus, ulna, metacarpalia V and metatarsalia V, already recognized as outstanding classification characteristics, show the species-separating factor. These methods of species differentiation were tested using the findings from the Unicorn Cave (Einhornhöhle).
Of the 595 postcranial findings from the Unicorn Cave 108 could be allocated according to species. 55% of the ulna, metacarpalia III and IV, femur, metatarsalia II, IV, and V findings were classified with 100% probability as Ursus deningeri. 75% of the metatarsalia I with a probability > 95%, 83% of the radius, metacarpalia I, II and V and metatarsalia III with probability > 90%, all as Ursus deningeri. Altogether, most of the postcranial skeletal elements with complete data for the given combinations of characteristics were classified as the older Ursus deningeri. Findings from newer excavations proved mostly too fragmentary and lacking in the necessary species-distinguishing characteristics to be used in the analysis. Specimens from both species have however been found in old and new diggings. Apparently the cave was chiefly inhabited by the geologically older Ursus deningeri, fossils of which became rearranged and mixed with those of the later Ursus spelaeus during the course of sedimentation.