Long bones from a taxonomically diverse assemblage of extant terrestrial mammals, spanning more than three orders of magnitude in body mass, have been measured in order to evaluate earlier models proposed for limb allometry as a means of physically coping with increased body size in large species. Linear regression models are unable to explain long bone scaling across a large size range of mammals, as differential scaling is present in large and small species and smaller species tend to approach geometric similarity to a considerably greater extent than larger species. Attempting to explain the morphology of the appendicular skeleton across the large size range of terrestrial Mammalia by means of a standard power function will probably inevitably lead to oversimplification, concealing underlying adaptations for coping with increased size. These include differences in limb posture, decrease in locomotory potential and greater duty factors during fast locomotion in large animals compared to anatomically similar smaller forms, and differential scaling of the appendicular skeleton in small and large mammals.