Studies of rodents have shown that both forced and voluntary chronic exercise cause increased hindlimb bone diameter, mass, and strength. Among species of mammals, "cursoriality" is generally associated with longer limbs as well as relative lengthening of distal limb segments, resulting in an increased metatarsal/femur (MT/F) ratio. Indeed, we show that phylogenetic analyses of previously published data indicate a positive correlation between body mass-corrected home range area and both hindlimb length and MT/F in a sample of 19 species of Carnivora, although only the former is statistically significant in a multiple regression. Therefore, we used an experimental evolution approach to test for possible adaptive changes (in response to selective breeding and/or chronic exercise) in hindlimb bones of four replicate lines of house mice bred for high voluntary wheel running (S lines) for 21 generations and in four nonselected control (C) lines. We examined femur, tibiafibula, and longest metatarsal of males housed either with or without wheel access for 2 months beginning at 25-28 days of age. As expected from previous studies, mice from S lines ran more than C (primarily because the former ran faster) and were smaller in body size (both mass and length). Wheel access reduced body mass (but not length) of both S and C mice. Analysis of covariance (ANCOVA) revealed that body mass was a statistically significant predictor of all bone measures except MT/F ratio; therefore, all results reported are from ANCOVAs. Bone lengths were not significantly affected by either linetype (S vs. C) or wheel access. However, with body mass as a covariate, S mice had significantly thicker femora and tibiafibulae, and wheel access also significantly increased diameters. Mice from S lines also had heavier feet than C, and wheel access increased both foot and tibiafibula mass. Thus, the directions of evolutionary and phenotypic adaptation are generally consistent. Additionally, S-line individuals with the mini-muscle phenotype (homozygous for a Mendelian recessive allele that halves hindlimb muscle mass [Garland et al., 2002, Evolution 56:1,267-1,275]) exhibited significantly longer and thinner femora and tibiafibulae, with no difference in bone masses. Two results were considered surprising. First, no differences were found in the MT/F ratio (the classic indicator of cursoriality). Second, we did not find a significant interaction between linetype and wheel access for any trait, despite the higher running rate of S mice.