A Method of Fitting Growth Curves of the von Bertalanffy Type to Observed Data

Abstract

A method is described for obtaining the best least-squares estimates of the parameters L[image], k, and t0 when von Bertalanify curves of the type L = L = L[image] (1 - ek(t-t0)) are fitted to observed data. This method imposes no restrictions on the number or size of the samples or on the time intervals between them. It also provides estimates of the limits of error of the parameters. The amount of computation is fairly large, but a method of systematizing it is described which makes manual computation practicable for moderate-sized sets of data. The method has been used to develop a computer program which seems to have advantages over some existing methods. A numerical example is worked out in full to illustrate application of the method.