Species Concepts in Ornithology

Abstract

The biological species concept (BSC) has been generally accepted by ornithologists in theory and practice for half a century. The concept has been considered useful both from evolutionary and taxonomic viewpoints. Criticisms of the BSC, largely by nonornithologists, have not led to substantial modification of the concept in ornithology. We discuss the concepts of grouping individuals into taxa and ranking these taxa as biological species, defined as reproductively isolated groups. The dimensionality of species is discussed, and we suggest that the problems in extending the BSC from nondimensional (sympatric) to multidimensional (polytypic species) situations are too serious to ignore. Reproductive isolation has no consistent genotypic or phenotypic correlates that can serve as predictors of reproductive compatibility of allopatric groups. Thus, many biological species are inappropriate groups for comparative biology or analyses of evolutionary history, as they are frequently conglomerates of several independent, not necessarily monophyletic groups. We discuss why hybrid zones do not necessarily contribute to clarification of species limits. The theoretical and practical limitations of the BSC lead us to advocate a phylogenetic species concept (PSC) (sensu Cracraft 1983), in which taxa are monophyletic, diagnosable clusters of individuals and species are the smallest diagnosable clusters. Diagnostic characters are considered to flag independent evolutionary histories and are used to delineate species boundaries. Phylogenetic species are therefore basal evolutionary units that should be the units used in phylogeny reconstruction, speciation analysis, and biogeography. We discuss the potential use of subspecies names, and recommend that they not be used. Problems with the PSC are discussed, such as its apparent neglect of important biological information, its use of seemingly trivial characters to delimit species, its apparent typological nature, and resulting unwieldy classifications. Implementation of the PSC is discussed with reference to practical problems in determining the statistical limits of phylogenetic species.