Species as units of analysis in ecology and biogeography: are the blind leading the blind?

Riddle & Hafner (1999) suggest that genetically differentiated lineages should be formally recognized as evolutionary significant units (ESUs) and that ESUs should replace currently recognized species as units in quantitative ecological and biogeographic analyses. Riddle & Hafner imply that if comparisons of desert rodent communities across many localities in south-western North America by Brown & Kurzius (1987) and Kelt et al. (1996) had used ESUs rather than species, substantially different results and conclusions would have been obtained. Here we defend our use of species in these studies, and question the wisdom of formalizing the concept of ESU and applying it in most biogeographic, ecological and behavioural studies. Studies of genetic variation and description of ESUs are uneven across taxa of mammals and geographical regions, and those ESUs that have been described are often difficult or impossible to identify on the basis of the morphological traits used to distinguish currently recognized species. Consequently, most of the specimens in museum drawers, nearly all of the fossil remains, and most of the animals studied in the field or laboratory by ecologists, behaviourists and other biologists cannot be assigned to ESUs. The currently recognized species, usually described originally on the basis of morphological and biological species concepts, provide the only relatively consistent, operational taxonomic units. Results of our studies on the biogeography of desert small mammals are not seriously altered by recent subdivisions of some species into multiple ESUs.