The universal facial attractiveness (UFA) hypothesis proposes that some facial features are universally preferred because they are reliable signals of mate quality. The primary evidence for this hypothesis comes from cross-cultural studies of perceived attractiveness. However, these studies do not directly address patterns of morphological variation at the population level. An unanswered question is therefore: Are universally preferred facial phenotypes geographically invariant, as the UFA hypothesis implies? The purpose of our study is to evaluate this often overlooked aspect of the UFA hypothesis by examining patterns of geographic variation in chin shape. We collected symphyseal outlines from 180 recent human mandibles (90 male, 90 female) representing nine geographic regions. Elliptical Fourier functions analysis was used to quantify chin shape, and principle components analysis was used to compute shape descriptors. In contrast to the expectations of the UFA hypothesis, we found significant geographic differences in male and female chin shape. These findings are consistent with region-specific sexual selection and/or random genetic drift, but not universal sexual selection. We recommend that future studies of facial attractiveness take into consideration patterns of morphological variation within and between diverse human populations.