Several meta-analyses confirm that physical exercise can slow down postmenopausal bone loss, but it is not clear whether physical exercise alone can increase bone mass. Our intent was to evaluate high-impact exercises (including jumping) and combined balance and leg-strength training, with and without raloxifene treatment, in three healthy elderly women, age 68-71 years. The 40-week study period consisted of two 17-week exercise periods with a 6-week rest period in between. The jumping exercises were performed both vertically and in different directions. Effects were measured in bone mineral density (BMD), balance, maximal gait speed, and leg extensor strength. BMD (g/cm(2)) was measured with dual-energy X-ray photon absorptiometry (DXA) at the proximal femur, lumbar spine, and total body. After the first exercise period, large losses of trochanteric BMD (8.1%-10.8%) were seen in all subjects. After both 6 weeks of rest and the second exercise period, which included both exercise and raloxifene, BMD increased in all subjects. During both exercise periods, the balance, gait speed, and leg extensor strength increased in all subjects. The results show that this kind of high-impact exercise had limited effects on BMD, but had large positive effects on balance, gait speed, and leg extensor strength. In conclusion, high-impact exercise in elderly women improves their fall risk factors, but, at least without raloxifene treatment, the trochanteric fracture risk might even increase because of reduction in the regional bone mass.