On the arbitrary identification of real species

Introduction The detection of a new species is the result of a decision-making process, one that has traditionally and primarily been based upon the discovery of distinguishing characters (Cronquist 1978;Mayr 1982;Winston 1999; Sites &Marshall 2004). This process, called diagnosis, is acutely important as it necessarily lies at the crux of the discovery of biodiversity, including the identification of conservation units. With the rise of quantitative phylogenetic methods, and the increasing accessibility of molecular data, numerous methods for diagnosis have been proposed in recent years (Sites & Marshall 2004). For many situations the identification of a new species does not particularly require quantitative methodology. These are the cases where the organisms of the putative new species are conspicuously divergent from all known species and, thus, where the new species is identified as a sister taxon to previously identified groups. But increasingly, as more new species are described and as more species are the subject of additional investigation, the questions of diagnosis arise within previously described species, wherein patterns of differentiation among populations of the same species must be interpreted in taxonomic terms. In short: how do we decide when a closer look at one taxonomic species actually reveals the presence of more than one species? A similar question arises in conservation contexts: do the data from one species (or one conservation unit) actually reveal the presence of multiple units, each of which merit recognition and possibly protection? This question may not be cast in terms of the taxonomic rank of species, and so the criteria used for diagnosis of a new species may differ from those used for diagnosing a population in terms of meriting conservation status. For example, in the United States the language of the amended Endangered Species Act (16 USC §§1531– 1544) refers to a ‘distinct population segment of any species of vertebrate fish or wildlife which interbreeds when mature’. In effect this conservation policy has identified a taxonomic category of ‘distinct population segment’ or ‘DPS’ as it is called in discussions on biodiversity conservation, one for which criteria and protocols for diagnosis have been much discussed (Waples 1991a; Moritz 1994;