The use of primitive character state distributions in the assessment of holophyly

Abstract

Cladistic analyses are based on the distinction between primitive and derived character states (hypotheses of the polarity of evolutionary transformations) and a complete reliance on only derived character state distributions as bona fide evidence of holophyletic assemblages of taxa. The cladistic premise that only derived character state distributions provide evidence of holophyly is reconsidered and shown to be both unjustified and inconsistent with the desire or methodological prescription of using all the available evidence. Cladistic techniques are here viewed primarily as methods for the ordering of character states so that they may be differentially weighted. The problem of assigning realistic differential weight to primitive and derived character state distributions is briefly discussed. One possible use of primitive character state distributions as evidence of holophyly that is largely free of the problem of weight is proposed. This application is illustrated with an example from the caecilian amphibians (Amphibia: Gymnophiona).