The present study was designed to provide data on the effects on femoral bone of endurance training starting only 3 months after orchidectomy in rats. A total of 70 Wistar male rats were used at 8 weeks of age. On day 0 of the experiment, 10 rats were killed by cervical dislocation to be used as first controls. Among the 60 other animals, half was surgically castrated (CX) or sham operated (SH). On day 90, 10 CX and 10 SH were killed and used as intermediary controls (ICX and ISH). Among the other 20 CX and 20 SH, 10 within each group (CXE, SHE) were selected for treadmill running (60% maximal oxygen uptake, 1 h x day(-1), 5 days x week(-1) for 12 weeks). The 20 other rats were used as sedentary controls (CXR, SHR) and killed (as runners) on day 180. On day 90 femoral bone density (BMD) and mineral content (BMC) were lower in ICX than in ISH. On day 180 total femoral BMD was lower in CXR than in CXE. Simultaneously metaphyseal femoral BMD was lower in CXR than in CXE, SHR or SHE. Furthermore, at that time, no significant difference concerning BMD and BMC was observed between SHR and CXE. This would indicate that treadmill running starting only 3 months after orchidectomy is able to restore BMD and BMC to control values, mainly by inhibiting bone resorption (as shown by decreased urinary deoxypyridinoline excretion in CXE) without decreasing osteoblastic activity (evaluated by plasma osteocalcin concentration).