Different methods for adjusting muscle strength ( S) to normalise for differences in various estimates of body size [such as body mass ( m) or, infrequently, some other anthropometrical measurements] have been either proposed or applied when presenting the results of muscle function tests in various medical, ergonomic, and sport related studies. However, the fact that the relationship between S and body size may differ when muscle torque (measured using a standard isokinetic apparatus) and muscle force (measured using a dynamometer) are recorded has not been taken into account. To address this problem, we tested both muscle force and muscle torque under isometric conditions in six different muscle groups. The relationship assumed between S and m was S=k.m(b) and, according to a simple mechanical model based on geometrical similarity we developed, the exponential parameter b would be expected to equal 1.00 and 0.67 for torque and force, respectively. The experimentally obtained values for the parameter b were higher for muscle torque than for muscle force in five out of the six muscle groups tested ( P=0.068; Wilcoxon matched pairs test). Despite a relatively wide scatter, the mean (SD) values were also close to those predicted, being b=0.67 (0.19) (corresponding to the allometric scaling method) and b=1.02 (0.34) (corresponding to the ratio standards method) for muscle force and for muscle torque, respectively. Therefore, we concluded that the ratio standards and allometric scaling should be employed to adjust S for body size when muscle torque and muscle force, respectively, are tested.