Allometric scaling of strength measurements to body size

For comparative purposes, normalisation of strength measures to body size using allometric scaling is recommended. A wide range of scaling exponents have been suggested, typically utilising body mass, although a comprehensive evaluation of different body size variables has not been documented. Differences between force (F) and torque (T) measurements of strength, and the velocity of measurement might also explain some of the variability in the scaling exponents proposed. Knee extensor strength of 86 young men was assessed with measurement of torque at four velocities (0-4.19 rad s(-1)) and force measured isometrically. Body size variables included body mass, height and fat-free mass. Scaling exponents for torque were consistently higher than for force, but the velocity of torque measurement had no influence. As the confounding effects of fat mass were restricted, scaling exponents and the strength of the power-function relationships progressively increased. Fat-free mass determined a surprisingly high proportion of the variance in measured strength (F, 31%; T, 52-58%). Absolute force and torque measurements, and even torque normalised for body mass, were significantly influenced by height, although strength measures normalised to fat-free mass were not. To normalise strength measurements to body mass, for relatively homogenous lean populations (body fat < 20%), exponents of 0.66 (F) and 1.0 (T) are appropriate. For more adipose populations (body fat > 20%) lower body mass exponents appear more suitable (F, 0.45; T, 0.68). Nevertheless, fat-free mass is the recommended index for scaling strength to body size, and higher exponents (F, 0.76; T, 1.12) are advocated in this case.