Published online by Cambridge University Press: 01 March 2000
The visual predation hypothesis of primate origins was introduced by M. Cartmill in the 1970s. In outlining a series of predictions regarding changes in orbital orientation, he further posits that the formation of a bony postorbital bar is correlated with greater orbital convergence, which in turn appears to be linked to selection for increased stereoscopic visual acuity in a nocturnal milieu. A series of predictions related to this model is tested in living analogues that vary in postorbital bar formation: felid and herpestid carnivorans and the pteropodid megabats. Bivariate correlations and regressions, ANOVA and ANCOVA are used to assess the intrinsic allometric and non-allometric factors thought to affect two aspects of orbital orientation, convergence and frontation. These angles of orbital orientation are then examined in an effort to determine how they are associated with the development of a postorbital bar. Cartmill suggested that convergent orbits develop from changes in relative orbit diameter and relative interorbital breadth. This study indicates that felids show only weak support for these predictions, while herpestids do not provide support and pteropodids show some support. Instead, in response to natural selection, convergence seems to develop to improve stereoscopic vision and depth perception in nocturnal predators. As predicted, felids show a significant positive relationship between orbital frontation and relative brain size. Analyses in herpestids, however, do not show significant relationships between frontation and relative palate length or relative brain mass. Pteropodids also do not support the model regarding intrinsic frontation factors. In both carnivoran groups, convergence and frontation are positively correlated, whereas they are not correlated in Pteropodidae. In partial contrast to Cartmill's predictions, we found that orbital convergence and/or orbital frontation are generally higher in carnivorans and bats with postorbital bars. Therefore, a greater emphasis on nocturnal stereoscopic visual predation and increased relative brain size both move the orbital aperture out of the plane of the temporal fossa and a postorbital bar thus provides adequate rigidity to the lateral orbital margin. In the more encephalized Felidae, orbital frontation is positively correlated with relative brain size; as smaller mammals have relatively larger brains, this explains why postorbital bars tend to be found in smaller felids. In herpestids and pteropodids, postorbital bars tend to occur in larger taxa which are more convergent than smaller forms.