Inferring social behavior from sexual dimorphism in the fossil record.

Sexual dimorphism is commonly used as evidence of the behavior of extinct species. Even so, few analyses scrutinize whether extant comparative data support inferences of mating systems or behavior in extinct species. This analysis evaluates the relations between measures of dimorphism and several estimates of mating system and intrasexual competition. Dimorphism alone provides poor resolution for reconstructing behavior. Many behavioral inferences based on perceived dimorphism are not supported by extant comparative data. This reflects the large standard errors of relations between dimorphism estimates and behavioral classifications. Used with caution, dimorphism can provide a hint of the behavior of extinct species in some cases. However, in many cases inferred dimorphism allows little more than an inference of polygyny, without any indication of specific types of mating systems.