NAMED SUBSPECIES AND THEIR SIGNIFICANCE IN CONTEMPORARY ORNITHOLOGY

Subspecies, or geographic races, are diagnosable populations that, at least during the breeding season, are largely allopatric with other subspecies of the same species. In attempts to give objectivity to the subspecies concept, arbitrary rules have been applied for the recognition of subspecies (e.g., the "75% rule/' whereby 75% of the individuals should be identifiable to subspecies; there are several other rules). As a case study, I examined the usefulness of the subspecies concept in describing geographic variation of a polytypic American songbird, the Savannah Sparrow (Passerculus sandwichensis). About 21 subspecies of this species have been recognized in the taxonomic literature, but much of the geographic variation is clinal. I argue that there is little value in subdividing a clinal continuum into different subspecies. Rather, the use of subspecies is best restricted to distinctive, and usually geographically isolated, populations. I show that this has been done for only a few of the named subspecies of Savannah Sparrows. Received 31 July 2006, accepted 6 March 2007. Resumen.Las subespecies, o razas geograficas, son poblaciones generalmente alopatricas (al menos durante la epoca de reproduction) y que se pueden diferenciar claramente de otras subespecies de la misma especie. Se han propuesto diferentes reglas para asignar individuos a una determinada subespecie, e.g. "la regla del 75%" (75% de los individuos tienen que ser identificables como subespecie), en un intento de dar objetividad al concepto de subespecie. Como caso de estudio, discuto la utilidad del concepto de subespecie para describir la variacion geografica de Passerculus sandwichensis. Se han reconocido cerca de 21 subespecies, pero gran parte de la variation geografica es clinal. Argumento que es de poco valor el subdividir una clina continua en subespecies diferentes. Se debe restringir el uso de subespecies para poblaciones bien diferenciadas y aisladas geograficamente. Muestro que esto se ha hecho solo para unas pocas de las subespecies de Passerculus sandwichensis. In recent years, many biologists have used intraspecific geographic variation to test hypotheses about adaptation and evolution, and named subspecies have reflected this variation. For example, M0ller and Cuervo (1998) compared feather ornamentation in birds to test the hypothesis that sexual selection promotes speciation and found that ornamented species had more subspecies than non-ornamented specieswhich suggests an association between subspeciation and ornamentation. Likewise, Sol et al. (2005) examined brain size relative to body size in Holarctic passerines, to test the hypothesis that behavioral changes are an important ^-mail: rising@zoo.utoronto.ca driver of evolutionary diversification, and found that species with large relative brain size have undergone more extensive subspecific diversification. It is clearly important for these studies that named subspecies more or less accurately reflect units of intraspecific diversification. Wilson and Brown (1953) identified several problems with the subspecies concept as then applied. Among these was the arbitrary lower limit of the "distinctiveness" of subspecies (i.e., how distinctive must a population be to earn a subspecific trinominal name?). There are several arbitrary rules, the most common being the "75% rule" (Amadon 1949), though efforts have been made to apply more rigorous diagnosability rules to the classification of subspecies (e.g., Patten and Unitt 2002, Remsen 2005, Cicero and