Deep-time history of primate behavior and ecology as revealed by ancestral state reconstructions

Abstract

Primates exhibit many key behavioral traits that differentiate them from their mammalian relatives. The origin and evolution of these traits is a major focus of primate paleontology. Here, we perform a formal ancestral state reconstruction of extant primate species to generate hypotheses about when these behaviors originated in primate evolutionary history. We compiled a large dataset of primary source data on substrate use, activity pattern, group size and structure, mating system, natal dispersal, litter size, and diet for 196 extant species. We also include data on body size, sexual dimorphism, and encephalization quotient. We performed ancestral character estimation of continuous characters using a Bayesian model and discrete or binned characters using stochastic character mapping of a k-state Markov model (Mk model). We reconstruct the ancestral crown primates as highly arboreal, nocturnal, small bodied, small brained, and eating a diet predominantly of fruit and invertebrates with the possible addition of other plant foods such as leaves and flowers. Social systems are poorly estimated at deep nodes in the tree, but the best supported states involve small to medium-sized groups. Larger, more complex social groups evolve later and emerge alongside diurnality and larger body size. In general, reconstructions at key nodes are consistent with known fossils. Exceptions include ancestral strepsirrhines, which are not reconstructed as being similar to well-known adapoids, and the brain size of ancestral anthropoids, which is reconstructed as larger than most early anthropoid taxa.