Strain magnitude may be more important than the number of loading cycles in controlling bone adaptation to loading. To test this hypothesis, we performed a 12 month longitudinal cohort study comparing bone mass and bone turnover in elite and subelite track and field athletes and less active controls. The cohort comprised 50 power athletes (sprinters, jumpers, hurdlers, multievent athletes; 23 women, 27 men), 61 endurance athletes (middle-distance runners, distance runners; 30 women, 31 men), and 55 nonathlete controls (28 women, 27 men) aged 17-26 years. Total bone mineral content (BMC), regional bone mineral density (BMD), and soft tissue composition were measured by dual-energy X-ray absorptiometry. Bone turnover was assessed by serum osteocalcin (human immunoradiometric assay) indicative of bone formation, and urinary pyridinium crosslinks (high-performance liquid chromatography) indicative of bone resorption. Questionnaires quantified menstrual, dietary and physical activity characteristics. Baseline results showed that power athletes had higher regional BMD at lower limb, lumbar spine, and upper limb sites compared with controls (p < 0.05). Endurance athletes had higher BMD than controls in lower limb sites only (p < 0.05). Maximal differences in BMD between athletes and controls were noted at sites loaded by exercise. Male and female power athletes had greater bone density at the lumbar spine than endurance athletes. Over the 12 months, both athletes and controls showed modest but significant increases in total body BMC and femur BMD (p < 0.001). Changes in bone density were independent of exercise status except at the lumbar spine. At this site, power athletes gained significantly more bone density than the other groups. Levels of bone formation were not elevated in athletes and levels of bone turnover were not predictive of subsequent changes in bone mass. Our results provide further support for the concept that bone response to mechanical loading depends upon the bone site and the mode of exercise.