In a population-based sample of 348 men (age 22-90 years) and 351 women (age 21-93 years), we evaluated the relationship of bone density assessed at a variety of skeletal sites by dual-energy X-ray absorptiometry (DXA) with various muscle mass estimates obtained also from the DXA scan and with physical activity by interview and strength assessed both subjectively and objectively. All these parameters declined with age as judged from these cross-sectional data. All estimates of total skeletal muscle mass were strongly correlated with bone density at different skeletal sites. Muscle mass, in turn, was correlated with physical activity and hand strength. In multivariate models including these variables, muscle mass was the strongest determinant of bone density, accounting for 6-53% (mean 27%) of the variance at the different skeletal sites. Physical activity (and/or a physical activity x age interaction) was an independent predictor of bone mass in 48% of the site-specific models and accounted for 0.03-39% (mean 10%) of the variance, while hand strength (and/or a hand strength x age interaction) accounted for up to 4% (mean 1%) of the variance as an independent predictor of bone density in a third of the models. Although these variables together accounted for a large proportion of the variance in bone density, other potential predictors were not assessed in these analyses. The dramatic decline in physical activity over life seemed unable to completely explain the age-related loss of bone mass, and additional research is needed to determine whether the relationship of muscle mass with bone density is a direct one or due instead to other factors such as circulating hormone levels.