The remarkable diversity of coloration and species present in hummingbirds has been considered the result of sexual selection. I evaluate if color differences among species in the genus Coeligena are consistent with expectations from sexual selection theory. If sexual selection on color is important for speciation, closely related species should be markedly different in the colors of feather patches associated with aggression and breeding. I evaluate this prediction through a statistical assessment of the phylogenetic signal of colors from five feather patches: crown, gorget, belly, upper back, and rump. The first two are associated with aggressive and courtship displays and are expected to be under sexual selection, whereas the others are not. Contrary to expectations, the crown and gorget were the only patches with significant phylogenetic signal. Furthermore, I assess if populations of dichromatic species are more divergent in coloration and therefore have reduced gene flow. Color distances among dichromatic subspecies were larger than among monochromatic subspecies, but the magnitude of phenotypic differentiation was not related to levels of gene flow. These results support a role for sexual selection in shaping color variation among populations, but these differences alone are not sufficient to explain speciation.