Most physiological functions vary allometrically with body size; however, few investigators have examined the relationship between strength and body size with allometric scaling. Thus, we hypothesized that allometric analysis would reveal that both the amount and quality of muscle are significant determinants of strength in the elderly. Allometric analyses were used to determine the influence of limb cross-sectional area (CSA), physical activity, demispan (distance between index-middle finger web and the sternal notch), leg length, and sex on grip and plantar flexor strength in men (n = 188) and women (n = 205) age 55–86 years. Physical activity was measured using a self-reporting questionnaire (Taylor et al. [1978] J Chron Dis 31:741–755). Forearm and leg CSA was estimated from anthropometry. There was an age-related decline in grip strength, independent of forearm CSA, demispan, and sex, equal to ≈12% per decade, whereas plantar flexor strength adjusted for leg CSA, physical activity, and sex was reduced at a rate of ≈15% per decade. The allometric models explained 71.4% (r = 0.845) and 38.8% (r = 0.623) of the variance in grip and plantar flexor strength, respectively. Model parameters were identified using multiple linear regression (P < 0.05). Thus, grip strength = forearm CSA0.435 · demispan0.161 · exp(3.905 − 0.012 age + 0.413 sex) and plantar flexor strength = leg CSA0.223 · physical activity 0.115 · exp(5.867 − 0.015 age + 0.366 sex). These findings indicate that age-related reductions in muscle CSA do not fully account for strength declines with age. Physical activity is also important and partially explains these reductions. Am. J. Hum. Biol. 15:48–60, 2003. © 2002 Wiley-Liss, Inc.