This cross-sectional study examined whether reduced hip bone mineral density (BMD) is better explained by isokinetic knee extensor strength (KES), lower limb lean body mass (L-LBM), or activity level (PASE). Through population-based recruitment, 1543 adults without knee osteoarthritis were recruited. For men and women respectively, means ± SD were: Age 60.8±8.0 and 61.1±7.9 years; BMI 29.6±4.6 and 29.1±5.4 kg/m2; hip BMD 1.025±0.138 and 0.895±0.128 g/cm 2 ; KES 124.9 ±41 and 72.7±22.9 N•m; L-LBM 10.3±1.5 and 7.0±1.2 kg; and PASE 206.4±99.7 and 163.8±77.0. The relationship between BMD and KES in men (r=.21, p<.002) and women (r=.23, p<.001) was significant before adjustment. However, this association was no longer significant after controlling for L-LBM. Even after controlling for age, race, and sex, the association between BMD and KES was better explained by L-LBM (partial R 2 =.14, p<.001) than by PASE (partial R 2 =.00). Allometric scaling of KES to body size attenuated the association of BMD with KES (Std Beta=.03). The significant association between BMD and L-LBM (Std Beta=.36) remained stronger than that between BMD and weight (Std Beta=.21). Therefore, muscle mass accounted for a greater proportion of the variance in hip BMD than KES or activity level and explained a significant proportion of the association between weight and BMD.