To gain insight into past human physical activity, anthropologists often infer functional loading history from the morphology of limb bone remains. It is assumed that, during life, loading had a positive, dose-dependent effect on bone structure that can be identified despite other effects. Here, we investigate the effects of genetic background and functional loading on limb bones using mice from an artificial selection experiment for high levels of voluntary wheel running. Growing males from four replicate high runner (HR) lines and four replicate nonselected control (C) lines were either allowed or denied wheel access for 2 months. Using μCT, femoral morphology was assessed at two cortical sites (mid-diaphysis, distal metaphysis) and one trabecular site (distal metaphysis). We found that genetic differences between the linetypes (HR vs. C), between the replicate lines within linetype, and between individuals with and without the so-called "mini-muscle" phenotype (caused by a Mendelian recessive gene that halves limb muscle mass) gave rise to significant variation in nearly all morphological indices examined. Wheel access also influenced femoral morphology, although the functional response did not generally result in enhanced structure. Exercise caused moderate periosteal enlargement, but relatively greater endocortical expansion, resulting in significantly thinner cortices and reduced bone area in the metaphysis. The magnitude of the response was independent of distance run. Mid-diaphyseal bone area and area moments, and trabecular morphology, were unaffected by exercise. These results underscore the strong influence of genetics on bone structure and the complexity by which mechanical stimuli may cause alterations in it.