This cross-sectional study of adult female athletes assessed whether the apparent loading-related differences in bone structure are primarily associated with the loading type or the muscle performance-related joint moments. Several structural variables at shaft sites of the tibia, radius and humerus, and distal sites of the tibia and radius were measured with peripheral quantitative computed tomography (pQCT) among 113 female national level athletes (representing hurdling, volleyball, soccer, racket-sports and swimming) and their 30 nonathletic referents. For the weight-bearing lower extremities, the loading modalities of the above sports were classified into high-impact (hurdling, volleyball), odd-impact (soccer, racket-sports) and repetitive, nonimpact (swimming) loadings; and for the nonweight-bearing upper extremities into high magnitude (functional weightlifting in hurdling and soccer), impact (volleyball, racket-sports) and repetitive, nonimpact (swimming) loadings. As expected, athletes' bone mass was substantially higher at loaded bone sites compared with the nonathletic referents, but more pertinently to the locomotive perspective, the loading-induced additional bone mass seemed to be used to build mechanically strong and appropriate bone structures. Compared with controls, the weight-bearing bone structures of female athletes (swimmers excluded) were characterized by larger diaphysis, thicker cortices and somewhat denser trabecular bone. The athletes' bones at the nonweight-bearing upper extremity were generally larger in cross-sectional area. The estimated indices of joint moment (muscle force x estimated lever arm) were explained from 29% to 50%, and the loading modalities from 8% to 25%, of the variance in most bone variables (P < 0.05) of the tibia (shaft and distal site). In contrast to the weight-bearing tibia, only the estimated joint moment was positively associated (P < 0.05) with the structural characteristics of the radius and humerus, accounting for 6% to 26% of the variance in bone variables of the shafts of these bones. Such association was not observed at the distal radius. In conclusion, at the weight-bearing lower extremity, the strong bone structure of the female athletes was attributable to muscle performance-related estimated joint moments and impact loading modality. At the shaft sites of the nonweight-bearing upper extremity, the strong bone structure was mainly attributable to the estimated joint moments. Thus, different loading history and other features of loading seemed to govern the skeletal adaptation at the upper and lower extremity.