A group of 75 healthy women nearing menopause were entered into a longitudinal study of bone mass and bone cell function. All were at least 46 years, were still menstruating, and had premenopausal levels of follicle stimulating hormone (FSH) and estradiol (E2). Bone density measurements of the spine (dual-photon absorptiometry, DPA) and total-body (DPA) and forearm (single-photon absorptiometry, SPA) were made at 6 month intervals along with measurements of FSH and E2 over an average of 2.04 +/- 0.45 years. The annual rate of change in bone mass was determined for each individual by regressing the bone mass measurement at each site at the time of observation. The slopes were pooled, and the percentage annual change was calculated as the mean slope divided by the initial mean bone mass value. The mean rates of change in bone mass were calculated with and without weighting by length of observation. No meaningful differences in the results were seen in the weighted values compared to the unweighted values, and none of the sites showed a slope significantly different from zero. Demonstrated reproducibility and sample size provided sufficient power to detect annual weighted rates of change of 0.81% for spine BMC, 0.69% for spine BMD, 1.31% for forearm BMC, 1.55% for forearm BMC/BW, and 1.09% for total body bone mineral. We conclude that if changes in bone mineral are occurring in women at this age, they are substantially less than 1%/year for spine and not much more for the forearm or total skeleton.