Few data are available on bone density in late adolescence. We studied factors affecting peak bone density in females. Forty-three white girls, aged 13-20 yr, were studied. Integrated estrogen exposure over the pubertal years was obtained by a score based on physiological events known to reflect circulation estrogen levels. The subjects were selected to provide great variation in estrogen exposure. Bone mineral density (BMD) was measured by single photon absorptiometry (midradius) and dual photon absorptiometry (spine and first metatarsal of the foot). Weight, estrogen score, and testosterone levels were highly correlated with BMD of the spine, wrist, and foot (P less than 0.05). Age correlated positively only with the BMD of the wrist. Twenty-four girls reaching ages 18-20 yr in the 2 yr of observation were divided into groups reflecting low (less than 24), medium (25-48), and high (greater than or equal to 49) estrogen exposure. The lowest scoring groups had the lowest spine and wrist BMD (P less than 0.05). This group weighed less and had lower weight to height ratio (P less than 0.05), the lowest weight (P less than 0.05) during adolescence, the highest age of menarche, and the highest amount of fiber in the diet (P less than 0.05). These subjects were separated into low and high BMD groups. Those subjects with the lowest values for spine, wrist, and foot were found to have significantly lower estrogen exposure scores and lower weight/height ratios; in addition, low BMD of the foot was associated with higher activity levels. Thus, wrist and spine BMD are affected by estrogen exposure during adolescence and weight; foot BMD, in addition, was negatively affected by activity, suggesting that bone mass in the active adolescent is affected by the absence of estrogen exposure.