To specify the exercise-induced changes on different skeletal sites, the effect of a 5-week endurance swim training was studied in rats. Eighteen Lyon strain (Sprague-Dawley) 5-week old female rats were divided into nine sedentary and nine swimming rats. Each swim training session was increased by 15 minutes from 2-6 hours per day. A histomorphometric study was performed at the primary and secondary spongiosa of the distal femur and at the secondary spongiosa of lumbar and thoracic vertebral bodies. After training, bone loss was observed in the secondary spongiosa of lumbar vertebral bodies (24.7%) and in the primary spongiosa of distal femur (15.2%). A tendency to bone loss was also detected in the secondary spongiosa of distal femur (10.8%), whereas no change was detected in thoracic vertebral bodies. In secondary spongiosa, bone loss was accompanied with a thinning of trabeculae. Total eroded surfaces and osteoid surfaces were significantly decreased in the three studied skeletal sites, suggesting a decreased bone turnover. The decreased thickness of osteoid seams in both lumbar vertebrae and distal femur could mean that the osteoblastic activity has also been altered at the cell level, leading to thinning of trabeculae. Five-week swim training with such duration and intensity of exercise appears unable to increase bone volume in rats and, therefore, causes adverse effects. The three studied bones seemed to adapt differently to experimental conditions. The lack of ground reaction forces induced by water immersion might have contributed to the observed bone loss. "Normal" gravity would be an important cofactor in the osteogenic effects of exercise.