We evaluated anthropometric and performance measures which best predict bone mineral density (BMD) in 54 adolescent girls (14.6��0.5y; 22.7��14.0 months past menarche). BMD for the whole body, femoral neck, greater trochanter, lumbar spine (L2-L4), and mid-femoral shaft was assessed using dual-energy x-ray absorptiometry (DXA) (Hologic QDR 1000/W). Whole body lean mass and fat mass were derived from the whole body scan. Knee extensor strength and leg power were assessed by isokinetic dynamometry and the Wingate Anaerobic Power Test, respectively. Using simple regression, lean mass was significantly correlated with BMD at all bone sites r=.45-.77; p<0.001), and was more highly correlated with BMD at all sites than was body weight. Maximum leg power was also associated with bone mass at all sites (r=.41-.67; p<0.001) while leg strength correlated significantly with all sites (r=.41-.53; p<0.001) except the lumbar spine. Stepwise regression analyses revealed that 59% of the variance in whole body BMD was predicted by lean mass alone. No other variables, including fat mass, height, months past menarche, leg power, or leg strength, contributed additionally to the regression model. Similarly, lean mass was the only predictor of lumbar spine and femoral shaft BMD (R��=.25, R��=.37, respectively), while femoral neck and trochanteric BMD were best predicted by leg power (R��=.38, R��=.36, respectively). In this group of adolescent girls, lean body mass and leg power independently predicted bone mineral density of the whole body, lumbar spine, femoral shaft, and hip, which may suggest an important role for muscle mass development during growth to maximize peak bone density. / Graduation date: 1997