To elucidate determinants of peak trabecular bone density, we studied the role of androgens, estrogen, and aerobic exercise in 30 women from 18 to 22 years old. The women were divided into three groups: Sedentary, 11 normal women who did not exercise regularly; eumenorrheic, 10 athletes with normal menstrual function; and oligomenorrheic, 9 athletes with exercise-induced oligomenorrhea. All athletes participated in aerobic sports that did not involve selective resistance loading of the back. Serum free and albumin-bound testosterone (fab T), androstenedione (A), and estradiol (E2) were measured on four separate occasions at consecutive 7 day intervals and averaged. Trabecular density was measured by quantitative computed tomography of the lumbar spine. Peak trabecular bone density was related to fab T (r = 0.48, p = 0.007), A (r = 0.40, p = 0.03), and E2 (r = 0.40, p = 0.04). When taken in combination, androgens and estrogen each accounted independently for significant portions of the variance in bone density [fab T and E2 (R2 = 0.38, p = 0.002) and A and E2 (R2 = 0.27, p = 0.01)]. Bone density (mg/ml, mean +/- standard error of the mean, SEM) in the sedentary group (174 +/- 6) was not significantly different from that in the eumenorrheic (183 +/- 12, p = 0.47) or oligomenorrheic (161 +/- 11, p = 0.32) subjects. We conclude that androgens and estrogen function as independent and additive determinants of peak trabecular bone density in young women. The quantitative impact of aerobic exercise (without resistance loading) and exercise-induced menstrual dysfunction appears to be less important than that of the hormones.