Physical activity and body composition were examined with respect to variation in functional limitation over a 6-year period (four surveys conducted between 1994 and 2000) based on a cohort of 1,655 community-dwelling older women and men living in Sonoma, California. Measures of functional limitation and physical activity were based on standard self-report questions. Measures of body composition (lean mass, fat mass) were estimated from bioelectric impedance by using population-specific prediction equations derived from dual-energy x-ray absorptiometry. For women, a one-unit gain in lean mass:fat mass ratio reduced the report of limitation at all surveys 65.5% (95% confidence interval: 21.8, 87.4). A similar reduction was not observed for men; however, there was a 3% increase in the report of no limitation at any survey. The effect of high levels of physical activity reduced new functional limitation that occurred at the last survey by 36.8% (95% confidence interval: 0.0, 92.2) for men and 52.7% (95% confidence interval: 13.5, 91.9) for women. In summary, higher levels of physical activity appeared to reduce the risk of future functional limitation conditional on the level of functioning established early in the disablement process by lean mass:fat mass ratio.