Paraphyletic groups as natural units of biological classification

Abstract

Despite the broad acceptance of phylogenetic principles in biological classification, a fundamental question still exists on how to classify paraphyletic groups. Much of the controversy appears due to (1) historical shifts in terminology and definitions, (2) neglect of focusing on evolutionary processes for understanding origins of natural taxa, (3) a narrow perspective on dimensions involved with reconstructing phylogeny, and (4) acceptance of lower levels of information content and practicability as a trade-off for ease of arriving at formal classifications. Monophyly in evolutionary biology originally had a broader definition, that of describing a group with common ancestry. This definition thus includes both paraphyletic and monophyletic groups in the sense of Hennig. We advocate returning to a broader definition, supporting use of Ashlock's term holophyly as replacement for monophyly s.str. By reviewing processes involved in the production of phylogenetic patterns (budding, merging, and splitting), we demonstrate that paraphyly is a natural transitional stage in the evolution of taxa, and that it occurs regularly along with holophyly. When a new holophyletic group arises, it usually coexists for some time with its paraphyletic stem group. Paraphyly and holophyly, therefore, represent relational and temporal evolutionary stages. Paraphyletic groups exist at all levels of diversification in all kingdoms of eukaryotes, and they have traditionally been recognized because of their descent-based similarity. We review different methodological approaches for recognition of monophyletic groups s.l. (i.e., both holophyletic and paraphyletic), which are essential for discriminating from polyphyly that is unacceptable in classification. For arriving at taxonomic decisions, natural processes, information content, and practicability are essential criteria. We stress using shared descent as a primary grouping principle, but also emphasize the importance of degrees of divergence plus similarity (cohesiveness of evolutionary features) as additional criteria for classification.