Bone mass and muscular mass show a parallel evolution during growth, and parallel involution with age. However, the bone loss related to the withdrawal of oestrogens is independent of muscular waste. The extensive study of disuse osteoporosis shows that exercise without weightbearing cannot counteract the loss of bone mass provoked by bed rest or weightlessness. Physical training, even at low frequency (30 to 60 min/day, 2 or 3 days/week), can increase bone mass or reduce bone loss associated with age. This effect is even present when exercise is practised by very old people at a seemingly low level of muscular tension on bone. It is not known whether muscular exercise could be helpful in pathological osteopenia. Experiments in animals indicate a short-lived benefit of exercise practised during a definite growth period; the long term effect of physical training in humans, after cessation of such activity, has not been studied extensively. Equal distribution of tension on all parts of the skeleton is probably not mandatory to obtain a general effect of exercise on bone mass. It is assumed that muscular exercise acts through tension exerted on bone, but the exact mechanism is unknown, as are the specifications of effective exercise in terms of site of application, intensity, frequency and duration. Moreover, little is known about the expected synergy between exercise and occupational activity.