The signals that allow bone to adapt to its mechanical environment most likely involve strain-mediated fluid flow through the canalicular channels. Fluid can only be moved through bone by cyclic loading, and the shear stresses generated on bone cells are proportional to the rate of loading. The proportional relation between fluid shear stresses on cells and loading rate predicts that the magnitude of bone's adaptive response to loading should be proportional to strain rate. For lower loading frequencies within the physiologic range, experimental evidence shows this is true. It is also true that the mechanical sensitivity of bone cells saturates quickly, and that a period of recovery either between loading cycles or between periods of exercise can optimize adaptive response. Together, these concepts suggest that short periods of exercise, with a 4-8 h rest period between them, are a more effective osteogenic stimulus than a single sustained session of exercise. The data also suggest that activities involving higher loading rates are more effective for increasing bone formation, even if the duration of the activity is short.