Previous studies have referred to the degree of dorsal canting of the base of the proximal phalanx as an indicator of human-like metatarsophalangeal joint function and thus a diagnostic trait of habitual bipedality in the fossil record. Here, we used a simple method to investigate differences in forefoot function on a finer scale. Building on Duncan et al.'s (Am J Phys Anthropol 93 [1994] 67-81) research, we tested whether dorsal canting reflects differences between sexes in locomotor behavior, whether habitual shoe wear influences dorsal canting in humans, and whether proximal joint morphology differs between rays in Pan and humans. Our results corroborate previous research in showing that humans have proximal phalanges with joint orientations that are significantly more dorsal than, but overlap with, those of great apes. We also found that male gorillas have significantly more dorsally canted second proximal phalanges than their female counterparts, while the opposite pattern between the sexes was found in Pan troglodytes. Inter-ray comparisons indicate that Pan have more dorsally canted first proximal phalanges than second proximal phalanges, while the opposite pattern was found in humans. Minimally shod humans have slightly but significantly more dorsally canted second proximal phalanges than those of habitually shod humans, indicating that phalanges of unshod humans provide the most appropriate comparative samples for analyses of early hominins. Overall, our analysis suggests that though the measurement of dorsal canting is limited in its sensitivity to certain intraspecific differences in function, phalangeal joint orientation reflects interspecific differences in joint function, with the caveat that different patterns of forefoot function during gait can involve similar articular sets of metatarsophalangeal joints.