The increasing availability of noninvasive methods for measuring bone mass raises the issue of whether perimenopausal women should routinely have such measurements to identify those at risk for osteoporotic fractures of the hip, wrist, and spine. Although the mortality and morbidity caused by hip fractures would warrant routine screening, measurement of bone mass has uncertain value in assessing the risk for hip fracture. Wrist fractures generally cause only transient disability, and measurement of bone mass may not reliably predict risk. Measurements of bone density of the spine might be better able to assess a woman's risk for vertebral fractures, but the value of screening depends on whether the findings would affect a woman's decision about using estrogen therapy after menopause. Serial measurements of bone mass to estimate a woman's rate of bone loss are relatively imprecise, increase the cost of screening, and have at best a limited role in screening women to assess risk for osteoporotic fractures.