Working for birds in Africa

Studies of Socotran birds II. One, two or three species: towards a rational taxonomy for the Golden-winged Grosbeak


Summary. Rhynchostruthus, generally treated as monospecific, is one of a number of taxonomically enigmatic Old-World finch genera whose close relatives have long intrigued systematists. Using morphology and morphometrics, we investigated species limits in the Golden-winged Grosbeak Rhynchostruthus socotranus, which has traditionally been viewed as a polytypic species comprising three taxa: nominate socotranus on the island of Socotra; louisae in northern Somalia; and percivali in southern Arabia. Recently, however, Fry & Keith (2004) suggested that two species should be recognised within this genus: louisae in mainland Africa and socotranus (including percivali) in Arabia and Socotra. Our analysis suggests that as many as six plumage features can be used to separate males of the three taxa (three being diagnostic and the others virtually so), and five features to distinguish females (all of them diagnostic). Morphometric data subjected to a Principal Components Analysis suggest that the three taxa are rather better separated in size and shape than was previously thought, especially amongst males. In some respects Socotran birds more closely resemble Arabian populations (principally in the presence of a white cheek patch), rather than louisae of mainland Africa, but are nonetheless readily distinguished from both. This is unsurprising when one considers that most of Socotra's endemic avian taxa are either truly unique (i.e. species) or are probably best considered as synonyms of African forms (Kirwan in press a,b). Our results demand molecular testing, but provide strong indication that three allospecies, perhaps even full species, are involved, based on the guidelines for assigning species rank of Helbig etal. (2002), but marginally weaker evidence for the recognition of more than one species if the quantitative system used by Collar (2006, to be elaborated in full by Collar et al. in prep.) is employed. We recommend that R. socotranus be henceforth regarded as either three species or one, but suggest that to recognise two species within the genus is either over-estimating or under-estimating biodiversity.

Copyright © African Bird Club. All rights reserved.
UK registered charity 1053920


Web site designed and built by