Allometry as a tool for extrapolation of biological variables

Abstract 1. 1. An allometric power function, statistically calculated from a set of measured data, is an empirical equation. Theoretically, a power function can be the expression of a similarity constant, but whether this is true for an actual problem requires proof by a causal description. 2. 2. Investigations of a large number of model data sets show that from the three linear methods often used for indirect statistical computation of the power function, linear regression and major axis determination give more reliable results than the calculation of the reduced major axis. Non-linear regression seems not to be generally useful for biological data sets. 3. 3. As measures for the goodness-of-fit of a power function the mean percent deviation and the standard error of estimate are more useful than the correlation coefficient.