Characters, Computers, and Cladograms: A Review of the Berkeley Cladistics Workshop

Abstract

A workshop on cladistic methods was held in Berkeley, 23-28 Mar 1981. It was unusual in that it emphasized analysis of data sets submitted by the participants. This focused attention on the construction of data sets and the differ- ences among computer algorithms rather than on issues of classification or historical biogeography. Like any approach to phylogeny reconstruction, cladistic methods depend heavily dn choice of characters and character states, as well as determination of evolutionary polarity and transformation series, and implicit or explicit character weighting. If the group being analyzed is not monophyletic or if it includes hybrids the cladistic relationships of the taxa may be misinterpreted. Minor changes in the choice of characters or the choice of taxa often produced major differences in clado- grams. Such problems may be magnified by computer analyses. Different algorithms produce different results because they are based on different assumptions about the likelihood of character state changes. Despite these problems, the great strength of cladistic methods is that they force careful consideration of the nature and distri- bution of taxonomic characters and in doing so provide us with our best estimates of phylogenetic history. Since the time of Darwin, systematists have attempted to reconstruct phylogenies by looking at the distributions of character states in organ- isms. In general they have assumed that those organisms with more char- acter states in common are more closely related and that overall similarity should reflect genealogy. It was not until 100 years after Darwin that Hennig (1950, 1966) emphasized that only synapomorphies (shared de- rived character states) provide information about phylogenetic relation- ships; at any one level, similarity based on symplesiomorphies (shared primitive character states) cannot be used to determine relationship. In our view, this is the core of what has come to be known as cladistics. For some workers, however, cladistics is merely the study of the number and sequence of character state changes in the evolution of a group, without regard for synapomorphy (for discussion of this approach see Eldredge