Since Cope (1885) argued that a persistent theme in evolution is a tendency for body size to increase during phylogeny, comparative zoologists have looked for regular rules which would characterize the change in animal form with body size. Beginning with DuBois (1898, quoted in Stahl and Gummerson 1967), power-law equations of the type y = mxb (allometric equations) were used to compare a parameter y, such as brain weight, to another parameter x, often body weight, in individuals representing species of different size within a phyletic group. The method produced consistent empirical descriptions of the effects of growth and body size differences in adults. Brody (1945, pp. 615-617) made extensive use of the method in his Bioenergetics and Growth. Davis (1962) reported on organ weight relationships in cats, and Stahl (1965) and Stahl and Gummerson (1967) presented data on organ weights and somatic measurements in primates. Stahl (1963) also reviewed the literature on biological similarity. Writers seeking the meaning of allometric rules have mainly followed Huxley's (1932) interpretation for relative growth: a particular variable y (e.g., brain weight, or perhaps metabolic rate) and another variable x (e.g., body weight) are like two sets of capital that grow in the bank at different rates of interest. These rates of interest are a given fixed multiple of each other, so that (dy/ydt) = b(dx/xdt). Although this interpretation has been repeated many authors, its usefulness seems scant, because it does not illuminate the question why some anatomical or physiological parameters would grow at different rates of interest than others. Allometry literally means by a different measure, as opposed to isometry, by the same measure (Gould 1966). In the case of isometry, an organism or piece of machinery is scaled up in such a way as to maintain geometric similarity, so that all the dimensions of the small object are multiplied the same factor in designing the larger one. It is frequently impossible to design a large machine to work the same way a small model works while maintaining perfect isometry between the two. A large pendulum clock, for example, will not keep time synchronously with a perfect model of itself made to a smaller scale, because the period of the pendulum depends on the square root of its length. Only if one makes changes in the gearing can the hands of the small clock keep time with those of the large.