Further Away With Dental Microwear Analysis: Food Resource Partitioning Among Plio-Pleistocene Monkeys From The Shungura Formation, Ethiopia

This study aims at improving dietary discrimination among primates through a new analytic approach of dental microwear texture analysis integrating a surface sampling on facets of both phase I and II of mastication. The surface sampling is applied on these two types of molar facets on 104 specimens belonging to four extant cercopithecids, and on 207 Plio-Pleistocene specimens representing four taxa from the Shungura Formation, Ethiopia. The extant species, including Colobus guereza, Theropithecus gelada, Papio hamadryas, Chlorocebus aethiops, compose a model covering a wide range of feeding ecology from leaf-eating through grass-eating habits to opportunistic habits.A principal component analysis built with a selection of the most discriminant variables and extant species as a model displays that the two facets contribute equally to the dispersion of individuals and to the between-species differences. Also, most contributing variables along the components are not central values of texture parameters (mean and median of the parameters measured on subsurfaces or a single value for the whole surface) but rather statistics of their dispersion (percentiles) and distribution (kurtosis, skewness).Once the fossil specimens are inserted into the extant species model, the two large-bodied colobines, Paracolobus and Rhinocolobus, are found to differ from the extant Colobus guereza, one of the most leaf-eating colobines. This highlights ecological diversity among early African colobines. Besides, the analysis tracks betweengenus as well as temporal differences between the extinct species of papionins, attesting of a low ecological overlapping between Theropithecus and non-theropith papionins, notably after 2.4 Ma. No significant difference was found between T. brumpti, T. oswaldi, Theropithecus sp. supporting similar feeding habits with the present-day gelada, at least from 2.9 Ma onwards.