Phylogenetic relationships amongst swifts and swiftlets: a multi locus approach.

ABSTRACT

We recently reconstructed the troublesome swiftlet phylogeny using cytochrome-b mitochondrial DNA sequences. The relationship of the giant swiftlet (Hydrochous gigas) with swiftlets of the genus Aerodramus was, however, unresolved. In an attempt to clarify this issue, we now incorporated mitochondrial 12S rRNA and nuclear beta-fibrinogen intron 7 nuclear DNA sequences with the cyt-b sequences of six swiftlet, two swift, and one hummingbird outgroup species. A partition homogeneity (PH) test, used to determine the congruence of phylogenetic signal between two sets of sequences, suggested that cyt-b and Fib7 sequences were incongruent and therefore should not be combined. However, further analyses revealed that the apparent incongruence was probably due to the high amount of variation in cyt-b sequences. Separate and combined analyses of the three sequences unambiguously placed H. gigas as the sister-group of Aerodramus and supported monophyly of the swiftlets. These results were supported by analyses of combined NADH dehydrogenase subunit-2 (ND2) and cyt-b sequences of H. gigas in combination with sequences previously published by other workers. Recently, it was shown that the pygmy swiftlet (C. troglodytes)--in our phylogenetic analyses consistently placed with other, non-echolocating, Collocalia species--is in fact able to echolocate. Echolocation thereby lost its value to distinguish between different swiftlet genera. Furthermore, the phylogenetic distribution of echolocation can be explained either by its single evolution at the base of the swiftlets, with subsequent loss, or by independent evolution in Aerodramus and C. troglodytes. Because yet unpublished data suggest that only the auditory nuclei in swiftlet brains show adaptations to echolocation, the latter explanation seems the more likely one.