Interpreting hominid behavior on the basis of sexual dimorphism

Numerous studies use estimates of sexual dimorphism in canine tooth size and body weight to support speculation about the behavior of australopithecines. However, the range of mating systems inferred for australopithecines encompasses virtually the entire spectrum of mating systems seen among extant anthropoid primates, from monogamy to polygyny characterized by intense male male competition. This variety of opinion can be attributed partly to the unusual combination of high body size dimorphism and reduced canine dimorphism in australopithecines. Here we provide a joint comparison of recent models for the behavioral correlates of both canine dimorphism and body size dimorphism, and apply this to published estimates of dimorphism in body size and canine tooth size in hominids. Among extant species, body weight dimorphism and canine dimorphism are strongly correlated with estimates of intrasexual competition. Canine crown height dimorphism provides the best discrimination between taxa that show high degrees of male-male competition, and those that do not. Relative male maxillary canine tooth size offers additional evidence about male-male competition. On the other hand, canine occlusal dimorphism offers little discrimination among species of different male-male competition levels. Estimates of canine dimorphism, relative canine size, and body weight dimorphism in australopithecines provide little definitive information about male-male competition or mating systems. Dimorphism of Australopithecus africanus and Australopithecus robustus can be reconciled with a mating system characterized by low-intensity male-male competition. The pattern of dimorphism and relative canine size in Australopithecus afarensis and A. robustus provides contradictory evidence about mating systems and male-male competition. We review a number of hypotheses that may explain the unusual pattern of dimorphism of A. afarensis and Australopithecus boisei, but non-satisfactorily resolves the problem given current data.