Comparison of dental topography of marmosets and tamarins (Callitrichidae) to other platyrrhine primates using a novel freeware pipeline

Dental topographic metrics (DTMs), which quantify different aspects of the shape of teeth, are powerful tools for studying dietary adaptation and evolution in mammals. However, comparative samples of scanned mammal teeth suitable for analysis with DTMs remain limited in size and scope, with little or no representation of some major lineages, even within well-studied clades such as primates. In addition, current DTM protocols usually rely on proprietary software, which may be unavailable to many researchers for reasons of cost. We address these issues in the context of a DTM analysis of the primate clade Platyrrhini ("New World monkeys") by: 1) presenting a large comparative sample of scanned second lower molars (m2s) of callitrichids (marmosets and tamarins), which were previously underrepresented in publicly available platyrrhine datasets; and 2) giving full details of an entirely freeware pipeline for DTM analysis. We also present an updated discrete dietary classification scheme for extant platyrrhines, based on cluster analysis of dietary data extracted from 98 primary studies. Our freeware pipeline performs equally well in dietary classification accuracy of an existing sample of platyrrhine m2s (excluding callitrichids) as a published protocol that uses proprietary software, at least when multiple DTMs are combined. Individual DTMs however, sometimes showed very different results in classification accuracies between that of our freeware pipeline and that of the proprietary protocol, most likely due to the differences in the smoothing functions used. The addition of callitrichids still resulted in high classification accuracy in predicting diet with combined DTMs, although accuracy was considerably higher when molar size was included (90%) than excluded (73%). We conclude that our new freeware DTM pipeline is capable of accurately predicting diet in platyrrhines based on tooth shape and size, and so is suitable for inferring probable diet of taxa for which direct dietary information is unavailable, such as fossil species.