To assess the effects of physical activity patterns on trabecular bone density in college women, we studied three groups of nonsmoking eumenorrheic women with different (but chronic) exercise regimens. There were nine sedentary (SED) women exercising less than 1 h/week, nine women who performed aerobic (AER) exercise greater than 2.5 h/week, and nine women who supplemented aerobics with muscle-building activities (MB) for more than 1 h/week. Resting energy expenditure, calorie, protein, and calcium intake, total body weight, and body mass index were not statistically different among the three groups. AER and SED women had similar lumbar bone mineral density (BMD). MB women had significantly greater spinal bone density (p less than 0.007 versus SED, AER). IGF-1 (insulin-like growth factor) concentrations were greatest in MB (p less than 0.01), and hours muscle building per week correlated with IGF-1 (r = 0.86, p less than 0.03). For all 27 women (mean age 24.5 years), body mass index was the single best predictor of lumbar BMD (r = 0.42, p less than 0.03); hours in muscle-building exercise per week conferred an additive effect on lumbar BMD. This cross-sectional study of young women suggests chronic muscle-building exercises may augment lumbar bone mass. The additive effect of anaerobic exercise on bone density may be mediated by both local weight-bearing changes and possible systemic factors.