Several well-known Miocene catarrhines, including Proconsul heseloni, have been inferred to combine quadrupedal walking in an arboreal substrate with a significant amount of climbing during locomotion. The degree to which some of these species were adapted to perform these behaviors is not fully understood due to the mosaic of ‘ape-like’ and ‘monkey-like’ traits identified in the forelimb. Given these unique combinations of forelimb features in the fossils, we report on forelimb traits that should be emphasized when investigating skeletal adaptation to quadrupedalism (defined in this manuscript as symmetrical gait movement on horizontal supports, excluding knuckle-walking) and climbing (including both vertical climbing and clambering). We investigate the correspondence between: 1) quadrupedalism and two well-known forelimb traits, humeral torsion and olecranon process length, and 2) climbing and phalangeal curvature. We also test the degree of phylogenetic signal in these relationships using phylogenetic generalized least-squares and branch length transformation methods in order to determine the models of best-fit. We present models that can be used to predict proportions of quadrupedalism and climbing in extant and extinct anthropoid taxa. Each trait–behavior correlation is significant and characterized by an absence of phylogenetic signal. Thus, we employ models assuming a star phylogeny to predict locomotor proportions. The climbing model based on phalangeal curvature and a proxy for size provides the most accurate predictions of behavior across anthropoids. The two quadrupedalism models are less accurate, but distinguish highly quadrupedal species from those that are not. Predictive equations based on these traits support the inference that P. heseloni performed a high proportion of quadrupedalism with a significant climbing component. The degree of phalangeal curvature measured in Pliopithecus vindobonensis predicts that this Miocene catarrhine species performed a proportion of climbing similar to Proconsul, while humeral torsion and olecranon process length provide conflicting inferences of quadrupedal locomotion in this species.