Objective The purpose of this study is to test the hypothesis that conarticular surfaces areas and curvatures are correlates of mobility at the hominoid talocalcaneal and talonavicular joints. Materials and Methods Articular surface areas and curvatures of the talonavicular, anterior talocalcaneal, and posterior talocalcaneal joints were quantified using a total of 425 three‐dimensional surface models of extant hominoid and fossil hominin tali, calcanei, and naviculars. Quadric surface fitting was used to calculate curvatures, pairwise comparisons were used to evaluate statistical differences between taxa, and regression was used to test for the effects of allometry. Results Pairwise comparisons show that the distributions of values for joint curvature indices follow the predicted arboreal‐terrestrial morphocline in hominoid primates with no effect of body mass (PGLS p > 0.05). OH 8 (Homo habilis) and LB 1 (Homo floresiensis) can be accommodated within the range of human variation for the talonavicular joint, whereas MH2 (Australopithecus sediba) falls within the ranges of variation for Pan troglodytes and Gorilla gorilla in measures of posterior talocalcaneal joint congruity. Conclusions Joint curvature indices are better discriminators than joint surface area indices, which may reflect a greater contribution of rotation, rather than translation, to joint movement in plantigrade taxa due to discrepancies in conarticular congruence and the “convex‐concave” rule. The pattern of joint congruence in Au. sediba contributes to other data on the foot and ankle suggesting that the lateral side of the foot was more mobile than the medial side, which is consistent with suggestions of increased medial weight transfer associated with hyperpronation. Am J Phys Anthropol 160:446–457, 2016. © 2016 Wiley Periodicals, Inc.