Growing bone responds to low or moderate exercise through significant additions of new bone in both cortical and trabecular moieties and results in adaptation through periosteal expansion and endocortical contraction. Intracortical activation frequency declines in growing bone in response to exercise, reducing porosity and the remodelling space. These adaptations can be maintained into and throughout adulthood. Young bones have a greater potential for periosteal expansion than aging bone, allowing them to adapt more rapidly and efficiently to an acute need for increased strength, but a threshold level of activity exists above which some bones respond negatively by suppressing normal growth and modelling activity, reducing geometric, mechanical and material properties in cortical and trabecular bone. From cross-sectional studies, differences in bone mass between exercising and non-exercising adults are generally less than 10%, but do not account for exercise history which may be very important, and often fail to consider important confounding variables. There is sufficient longitudinal data to demonstrate that moderate to intensive training can bring about modest increases of about 1-3% in bone mineral content (BMC) of men and premenopausal women. In young adults very strenuous training may increase BMC of the tibia up to 11% and its bone density (BD) by 7%, but may represent periosteal woven bone formation in response to excessive strain. Some evidence shows that exercise can also add bone mass to the post-menopausal skeleton, although the amounts are site-specific and relatively modest. Increases as high as 5-8% can be found after 1-2 years of intensive exercise, but additions of bone to the femur and radius are generally less than 2%, well within the range of the remodelling space and measurement precision. Although increases in bone mass of the post-menopausal skeleton may be extremely modest, physical activity is important to preserve bone mass and muscle function. Detraining reduces any bone mass increase to pre-existing values so that long-term benefits are only retained with continuing exercise. Most importantly, the amount of bone gain that can be achieved appears dependent primarily on the initial bone mass suggesting that individuals with extremely low initial bone mass may have more to gain from exercise than those with moderately reduced bone mass.