We report associations of anthropometric, lifestyle and menstrual characteristics with peak bone mass in a multiethnic population of premenopausal women, ages 25-34 y. Four bone sites were examined: the spine, calcaneus, and distal and proximal radius. Body mass index (BMI) was positively associated with bone mass at the spine and calcaneus, the two most trabecular sites. The calcaneus, in addition, was positively associated with an index of calf muscle mass. The distal radius was associated with an arm muscle index, grip strength and the intensity of menstrual flow (light, moderate or heavy). Menstrual flow was also associated with spine bone mass in models not adjusted for BMI; light flow was an indicator of low distal radius or spine bone mass. We hypothesize that light flow may indicate a lack of regular ovulation. The cortical proximal radius was positively associated with milk consumption and height. The magnitude of associations varied by exposure and bone site. Among the strongest associations were BMI at the spine and calcaneus (4 and 7% differences in bone mass per standard deviation), and menstrual flow at the distal radius (a 7% difference between light and heavy menstrual flow). Ethnicity was not statistically significant in multivariable models at any of the four bone sites. However, the prevalence of exposures associated with bone mass varied substantially among the ethnic groups. Ethnic differences in bone mass may have derived indirectly through differences in exposure histories. Other influences of skeletal development may also occur indirectly. Sports activity during adolescence, for instance, was not an independent predictor of bone mass, but sports activity was associated with both strength and muscle mass. Our results suggest that many factors influence peak bone mass, and that the factors vary in importance by skeletal region, possibly related to bone composition and the local intensity of exposure.