Debate continues on whether aging or estrogen deficiency is the more important determinant of postmenopausal bone loss. We compared 14 women who had undergone oophorectomy during young adulthood, 14 normal perimenopausal women, and 14 normal postmenopausal women (mean ages, 54, 52, and 73 years, respectively; mean duration of estrogen deficiency, 22, 0.3, and 22 years, respectively). Bone mineral density was assessed by single-photon and dual-photon absorptiometry. As compared with the perimenopausal group, the other two groups had significantly lower bone mineral density at the midradius, femoral neck, femoral intertrochanteric area, and lumbar spine (-15 per cent, -25 per cent, -16 per cent, and -23 per cent, respectively, in the oophorectomized group and -18 per cent, -28 per cent, -26 per cent, and -23 per cent, respectively, in the postmenopausal group). Because bone loss in the oophorectomized group (differing from the perimenopausal group in menopausal status but not in age) was almost as great as in the postmenopausal group (differing in both characteristics), we suggest that estrogen deficiency, and not aging, may be the predominant cause of bone loss occurring during the first two decades after natural menopause.