The potential and pitfalls of using simple dental metrics to infer the diets of African antelopes (Mammalia: Bovidae)

The use of mesowear to infer diets of extinct species is fast becoming widespread in palaeoecological studies. Nevertheless, traditional mesowear analyses suffer from a specimen number limitation, in that a minimum number of specimens identified to the species level is necessary to make accurate dietary predictions. This is problematic in many fossil African antelope (Mammalia: Bovidae) assemblages, where isolated teeth cannot always be assigned to species. Hereweexplore the possibility of using simple dental metrics to predict diets on the basis of individual teeth as well as gnathic rows using linear discriminant function analyses.We find that browsers are accurately classified at both the individual and species levels, across all models and tooth positions. Mixed feeders and grazers are classified accurately only sometimes, and this is probably a reflection of the more limited sample size of larger bodied species in our study. Body size was a highly significant predictor of the inaccurate classifications obtained in our models, with larger bodied species tending to grazing classifications and smaller bodied species browsing classifications. Nevertheless, the models correctly classify the majority of specimens we examined to their correct trophic group, as determined through stable isotope analyses or as defined through the literature. The methods outlined hold some promise for determining the diets of isolated fossil specimens unassigned to species in a simple manner and, when used in conjunction with other palaeodietary and palaeoecological proxies, may help determine palaeoenvironments more accurately.