Exercise that imparts rapid, high-magnitude mechanical loading is considered to be advantageous to bone health. Previous rodent studies have suggested that swimming may also be beneficial to bone. We investigated the differential effects of exercise with and without weight bearing on cortical and trabecular bone. Forty female Sprague-Dawley rats (120 days) were weight-stratified and randomized into four groups: swim control (Cs, n = 10), swim (S, n = 10), treadmill control (Ct, n = 10), and treadmill (T, n = 10). Treadmill speed was adjusted to match the average limb loading frequency used for swimming, and all training progressed to 1 hour/day, 5 days/week, for 12 weeks. Femurs and humeri were assessed for cortical morphometry by peripheral quantitative computed tomography, areal bone mineral density (BMD) by peripheral dual-energy X-ray absorptiometry, mineral content by ashing, strength by three-point bending, and trabecular volume (BV/TV) by micro-computed tomography. Swimming was associated with increases in cortical thickness and BMD in the humerus midshaft and trabecular BV/TV in the distal femur and proximal humerus compared with age-matched controls. Compared to swimming, treadmill training was associated with increases in percent ash of the femur and humerus and Young's modulus of the femur. Swimming appears to engender novel bone strains and osteogenic adaptations in the humerus and femur, which are different from those induced by normal cage activity. In summary, our findings suggest that when limb loading frequency is matched, swimming may afford greater benefits to cortical and trabecular bone than uphill treadmill work in rats.