Serological Data and Phylogenetic Inference: The Problem of Rates of Change

Abstract

Kirsch, J. A. W. (Department of Zoology, Univ. Kansas, Lawrence, 66044) 1969. Serological data and phylogenetic inference: the problem of rates of change. Syst. Zool., 18:296–311 [Immunology; cladistics, phylogeny, evolutionary rate; Marsupialia].—Inferences about the evolutionary history of a group based upon studies of its living members can only be made utilizing assumptions about evolutionary rates. One such assumption current among biochemists and serologists is that biochemical evolution has been steady and the same in all lineages. If this were so, cladistic relationships could easily be inferred from phenetics. The assumption of strict time-dependency may be interpreted either as a deterministic or as a stochastic statement. In the former case, the form of phenetic groups is logically determined in ways incompatible with the results of serological studies; this is illustrated by examples from the comparative serology of marsupials. A simple computer model for the evolution of a group of hypothetical taxa of known phylogeny, operating according to the rules of a stochastic interpretation of time-dependency, generates data about those taxa very like empirical data on marsupials. However, the inferred cladistics of the hypothetical group are incorrect in many details, and this is indicated by the incompatibility of phenograms representing the affinities of the taxa with their known cladistics. It is not claimed that the model reproduces the conditions of antigen evolution, but that it does indicate that such evolution, if time-dependent, must be so in a stochastic rather than a deterministic sense; and further, that having introduced chance as a factor in the evolution of biochemical characters, one has only a probability of inferring the correct phylogeny of a group from the resulting data about the similarities of contemporaneous taxa.