Evolutionary Rates and the Inference of Evolutionary Tree Forms

SEVERAL methods have been developed recently to decide which of a set of alternative trees is the most consistent with genetic, biochemical or morphological information about populations of present day organisms1–5. Most of the methods involve a “minimum evolution” hypothesis to the effect that the observed differences have been established by the smallest possible number of mutations or unit biochemical or morphological changes. The dubious nature of the minimum evolution hypothesis has been emphasized by Inger6 and Rogers et al.7, and the difficulties involved in selecting strictly comparable units of morphological change have been discussed by Lerman8. The fact that the number of possible distinct tree forms increases very rapidly with increase in the number of populations leads to computational difficulties; several authors have used some form of average-linkage cluster analysis on dissimilarity or association measures between populations to obtain a preliminary reduction in the number of tree forms2,3,9.