Study Objective: To assess the effect of weight-bearing exercise training and subsequent detraining on lumbar bone mineral content in postmenopausal women. Design: Non-randomized, controlled, short-term (9 months) trial and long-term (22 months) exercise training and detraining (13 months). Setting: Section of applied physiology at a university school of medicine. Patients: Thirty-five healthy, sedentary postmenopausal women, 55 to 70 years old. All women completed the study. There was 90% compliance with exercise training. Interventions: All women were given calcium, 1500 mg daily. The exercise group did weight-bearing exercise (walking, jogging, stair climbing) at 70% to 90% of maximal oxygen uptake capacity for 50 to 60 min, 3 times weekly. Measurements and Main Results: Bone mineral content increased 5.2% (95% confidence interval [Cl], 2.0% to 8.4%; P = 0.0037) above baseline after short-term training whereas there was no change (-1.4%) in the control group. After 22 months of exercise, bone mineral content was 6.1% (95% Cl, 3.9% to 8.3% above baseline; P = 0.0001) in the long-term training group. After 13 months of decreased activity, bone mass was 1.1% above baseline in the detraining group. Conclusions: Weight-bearing exercise led to significant increases above baseline in bone mineral content which were maintained with continued training in older, postmenopausal women. With reduced weight-bearing exercise, bone mass reverted to baseline levels. Further studies are needed to determine the threshold exercise prescription that will produce significant increase in bone mass.