Detection of Plio-Quaternary large mammal communities of Italy. An integration of fossil faunas biochronology and similarity

The application of a new method for the analysis of the large mammal Plio-Quaternary communities of mainland Italy is presented. This method, named “bootstrapped cluster analysis”, avoids, by mean of successive, ever more inclusive cluster analyses, the subjectivity in positioning boundaries of paleocommunities subdivisions. The results of this procedure indicate a good correspondence with previous qualitative analyses, albeit with some differences. Nevertheless, some local assemblages (e.g. Slivia) are grouped in significantly different positions than previously. The principal cause of these differences (as for unbalanced carnivores/herbivores ratio) is taphonomic bias. The bootstrapped cluster analysis approach is not affected by single species positioning as biochronological markers, grouping local assemblages upon their entire record with a statistical treatment. Moreover in this way any single paleocommunity is recognizable as an ecological/evolutionary unit.