Allometric scaling in comparative biology: problems of concept and method

Allometric scaling, a widely used comparative approach for studying the relationship between size, shape, and function in organisms, is examined in both concept and application. The general adoption of an "allometric method" with several standardized conventions has inhibited the creative study of size-correlated phenomena and has falsely simplified a complex area of research. Discussed here are some of the consequences of using a power function (and the resulting logarithmic transformation of data) to describe allometric trends, the dependence on correlation coefficients as a measure of strength of association, and the assumptions underlying use of empirical data to determine the relative size of structures. Several alternative methods for data analysis are suggested.