APPENDIX 1.

Statistical tests of univariate mesowear values calculated by our method on the original mesowear data for modern ungulate species (Fortelius and Solounias, 2000). The hierarchical clustering analysis (A 1.1) yielded similar results to the ones shown by Fortelius and Solounias (2000), clustering the species in relatively consistent and biologically meaningful dietary groups, with extreme browsers at one end and extreme grazers at the other. Typical diets in each cluster are named according to those of the dominant species in each cluster. MABRA = ‘”minute abraded brachydonts”; a special case of small ungulates which feed on fruit seeds and acquire a more abrasion-dominated mesowear signal than other browsers because of cusp tip-crushing wear (none of these were present in the Pleistocene of Europe). Discriminant analysis (A 1.2), following the methodology of Fortelius and Solounias (2000) showed that our univariate mesowear values still classify ca. 65 % of all extant ungulate species (excluding MABRA) and 85 % of extant ungulate species “typical of their dietary class” (see Fortelius and Solounias, 2000) correctly into the broad “traditional” dietary classes of “browsers”, “mixed-feeders” and “grazers”.

APPENDIX 2.

Mean mesowear values of species in localities, with standard errors.

APPENDIX 3.

Mean body mass (kg) of all species in localities, with standard errors.

APPENDIX 4.

Pairwise correlation analyses of mean body mass (kg) and mean mesowear of species in localities. Indications of correlation in brackets refer to R2 values, which do not have significant p-values, and ones without brackets refer to to significant correlations. + = positive, ̶ = negative, none = no correlation.

APPENDIX 5.

Pairwise correlation analyses of mean body mass (kg) between the species (upper rows) and with minimum, maximum and mean NAP % (lower rows) across localities. Count = number of pairs compared. Statistically significant p-values are emboldened.

APPENDIX 6.

Pairwise comparison by Wilcoxon tests of mean body mass of Cervus elaphus in localities. Negative Z values indicate smaller body size and positive values larger body size in the population marked in the first column compared to the one in the second column. Statistically significant p-values are emboldened.

APPENDIX 7.

Pairwise comparison by Wilcoxon tests of mean body mass (kg) of Dama dama in localities.

APPENDIX 8.

Pairwise comparison by Wilcoxon tests of mean body mass (kg) of Capreolus sp. in localities.

APPENDIX 9.

Pairwise comparison by Wilcoxon tests of mean body mass (kg) of Megacerini spp. in localities. The significant differences are mostly due to the smaller size of early Middle Pleistocene Praemegaceros and Megaloceros species compared to late Middle and Late Pleistocene Megaloceros giganteus.

APPENDIX 10.

Pairwise comparison by Wilcoxon tests of mean body mass (kg) of Rangifer tarandus in localities.

APPENDIX 11.

Pairwise comparison by Wilcoxon tests of mean body mass (kg) of Alcini in localities. The significant differences are mostly due to the larger size of early Middle Pleistocene Cervalces latifrons compared to Late Pleistocene and Holocene Alces alces.

APPENDIX 12.

Pairwise comparison by Wilcoxon tests of mean body mass (kg) of Stephanorhinus hundsheimensis in localities.

APPENDIX 13.

Pairwise comparison by Wilcoxon tests of mean body mass (kg) of Stephanorhinus hemitoechus in localities.

APPENDIX 14.

Pairwise comparison by Wilcoxon tests of mean body mass (kg) of Coelodonta antiquitatis in localities.

APPENDIX 15.

Pairwise comparison by Wilcoxon tests of mean body mass (kg) of Bison priscus/schoetensacki in localities. Some of the significant differences are due to the smaller size of early Middle Pleistocene Bison schoetensacki compared to Late Pleistocene B. priscus.

APPENDIX 16.

Pairwise comparison by Wilcoxon tests of mean body mass (kg) of Bos primigenius in localities.

APPENDIX 17.

Pairwise comparison by Wilcoxon tests of mean body mass (kg) of Equus ferus/mosbachensis in localities.