An analysis of two methods of fitting the von Bertalanffy curve to capture-recapture data

Abstract

The asymptotic inconsistency of the Fabens method of fitting the von Bertalanffy curve to capture- recapture data is proved under reasonably realistic assumptions on the growth of an animal, in the case of constant recapture intervals. The bias and inconsistency can be large, as is demonstrated by some simulations, and the unreliability of the estimates is also demonstrated by application to two data sets on the western rock lobster. A modified method proposed by Kirkwood and Somers suffers from the same deficiencies. It is concluded that, in general, neither method is trustworthy and a new method is required.