Some statistical aspects of the maximum parsimony method.

Abstract

The last three decades have seen considerable debate concerning the relative merits and problems associated with two competing approaches to phylogeny--approaches based on the parsimony principle versus maximum likelihood methodology. Although the two approaches may seem quite opposed, there are in fact some close relationships between them. For example, we describe a recent result that shows how maximum parsimony can be regarded as a type of maximum likelihood estimator when there is no common mechanism between sites (such as might occur with morphological data and certain forms of molecular data). Distinguishing between this and other implementations of maximum likelihood helps clarify some of the dispute that has surrounded the two methodologies. We also provide a brief overview of some mathematical and statistical properties of the maximum parsimony criterion.