l_raty
laurent raty
Kennedy M, Spencer HG. 2014. Classification of the Cormorants of the World. Mol. Phylogenet. Evol. 79: 249-257.Martyn Kennedy, Hamish G. Spencer. Classification of the Cormorants of the World. Molecular Phylogenetics and Evolution. In Press.
The full paper (with--unusual, and definitely to be applauded--the supplementary material added at the end of the pdf) is obtainable [here].
I played a bit with the data, they seem clean, and the relationships generally robust.
I have some reservations, however, about the position of Phalacrocorax gaimardi. The dataset the study was based on included sequences of four mitochondrial (12s, atp8-6, cox1, nd2) and 5 nuclear (bfib7, cryaa, irf2, park7, rapgef1) markers. Unfortunately, for gaimardi, besides the mtDNA, they obtained only irf2, which is the shortest of the five nuclear markers. This would not be a problem if the signal in irf2 was congruent with the signal present in the mtDNA--however, it is not: instead of making it highly distinct, irf2 places gaimardi within Phalacrocorax sensu stricto with a support that, although not huge, is nevertheless disturbingly high (see the attached nDNA tree). Of course, when the complete dataset is analyzed, 645bp of slow-evolving nDNA are unable to offset the signal brought by >3500bp of fast-evolving mtDNA, and the position of the taxon is in effect entirely determined by the mtDNA. But the conflict remains, and is unexplained. I'd really love to see more data from this species (particularly more nDNA, but mtDNA from other individuals, that might confirm what we currently have, would be a good thing as well).
(Incidentally: note also that this much more limited amount of nDNA for gaimardi makes the distances involving Poikilocarbo in Table 5 much higher than the rest, for a reason that has nothing to do with evolutionary distinctness. In all the other groups, the values are taken down by the inclusion of more slow-evolving nDNA.)
All that being said, I'm afraid that I remain to be convinced that so much splitting is really the best way to go. A rapid look at the attached mtDNA tree suggests that, if the distance separating Leucocarbo from Nannopterum marks a generic limit, then we should also split Gavia in two, Sula s.s. (Morus and Papasula already excluded) in three, Anhinga in three, and Pelecanus in four... Not to mention that, as indeed noted by the authors, the distance between the two sampled Microcarbo is also larger than between these two "genera". And the distances withing Phalacrocorax s.s. are dangerously close as well.
Personally, I'm happy with the two genera recognized in H&M4.
Attachments
Last edited: