Physical training may be able to improve bone strength through site-specific changes in the composition, size and structure of the bone without notable increases in volumetric density. To address this possibility specifically, we compared 14 competitive female weightlifters with 14 female physical therapy students. Peripheral quantitative computed tomographic scans (pQCT) were taken from the distal radius, radial shaft, distal femur, and tibial midshaft of the dominant limb. Analysis of covariance (ANCOVA) was used to estimate the intergroup differences, using body weight and age as covariates. Cortical density did not differ between the weightlifters and controls at any site, whereas trabecular density was greater in the weightlifters, the benefit being 10% (P = 0.186) at the distal radius and 11% (P = 0.040) at the distal femur compared with the controls. Weightlifters' cortical cross-sectional area was 38% (P = 0.029) larger at the distal radius, 26% larger (P = 0.001) at the radial shaft, and 9% larger at the tibial midshaft (P = 0.034). Consequently, the weightlifters' forearm bone strength indices were also significantly higher, the intergroup difference being 41% (P = 0.001) at the distal radius and 43% (P = 0.004) at the radial shaft. Thus, the observed intergroup difference at the distal radius was mainly due to enlarged bone, particularly its cortex, rather than higher volumetric bone density. Findings at the radial shaft were similar. In contrast, weightlifters' trabecular tissue at the distal femur was denser but the bone per se was not clearly bigger than that of the controls' (intergroup difference 5%, P = 0.117). We suggest that bones subjected to exceptionally high bending-loading (distal radius and radial shaft) are larger than their normal counterparts while at sites experiencing axial, compressive-loading (e.g., distal femur), a denser trabecular structure (more load-carrying area) may be sufficient and any substantial enlargement in bone size may not be necessary.