Clinical studies have found that nonimpact non-weight-bearing sports are associated with normal to low bone mineral densities. The objective of this study was to quantify structural bone adaptations of the mouse to swimming using in vivo micro-computed tomography. Prepubertal female mice underwent a 16-wk training program, in which they swam for progressively increasing durations up to 55 min for 5 d·wk(-1). A sham group was subjected to the water, but they did not perform the swimming exercise. Skeletal sites that were assessed included the proximal humerus, lumbar spine, midshaft and distal femur, proximal tibia, and the skull. Normal bone mass accrual was suppressed in the swim group during the initial 4 wk of training, during rapid growth. Swim mice gained 31.8% less (P = 0.001) and 16.7% less (P = 0.003) trabecular bone volume at the proximal tibial and humeral metaphyses, respectively. In the 12 wk that followed, low trabecular bone volume persisted in the swimmers (P < 0.001). Trabecular properties were reduced primarily at the metaphysis but not the epiphysis, and cortical bone was reduced at both compartments. At the femoral midshaft, swim mice had less cortical porosity (SHAM = 32% ± 3% vs SWIM = 22% ± 3%, P = 0.043). Despite the bone microarchitectural adaptations, swimming did not induce detectable changes to macroarchitecture, including bone length and total volume. Tissue mineral density was unaffected by swimming, suggesting that material changes did not occur. In conclusion, the response of prepubertal mice to swimming was adaptation at the bone microarchitectural level and resulted in an altered topology that was sustained until the end of the training program. Low bone mass was systemic throughout the mouse, with the largest depressions in the hindlimbs and spine.