Laurent, I see flammeolus is basal to the American group in your tree.
Earlier (e.g. by König) it was assumed this was the only American ‘true’ Otus – has anyone else suggested it was a Megascops (which of course is not extremely surprising) after all?
I see: reference 18 in the article. I should stop just looking at trees!
Otus flammeolus actually appeared basal to
Megascops in the trees published by Wink & Heidrich in König's book:
http://www.uni-heidelberg.de/institute/fak14/ipmb/phazb/pubwink/1999/20. 1999.pdf
...but without support (<50% bootstrap support in NJ).
See also Wink & Heidrich 2000:
http://www.uni-heidelberg.de/institute/fak14/ipmb/phazb/pubwink/2000/26. 2000.pdf
and Wink et al. 2004:
http://www.uni-heidelberg.de/institute/fak14/ipmb/phazb/pubwink/2004/28.2004.pdf
What is causing the (or, at least, part of the) trouble in the cytochrome b analyses seems to be
Pulsatrix. The position of this genus is poorly resolved in the published trees; it moves around a lot, being sister to
Strix and/or
Bubo, to (
Megascops +
flammeolus), or even to
flammeolus alone, depending of the analysis, and always with low support. The last position would of course be a problem for the inclusion of
flammeolus in
Megascops. Unfortunately, there are very few published genetic data for
Pulsatrix, so it's difficult to say more about this...
I've attached what I get in ML with publicly available COI sequences (the only other gene available for
flammeolus and a variety of screech- and scops-owls). Here again,
flammeolus groups with
Megascops and not with
Otus - with very high support. (Note also that, amazingly,
Otus sensu lato is monophyletic in this tree [
Ptilopsis not included] - but support is not very high at this node, so I guess this might easily be an artefact. [Or not...? :eek!
Incidentally, the separation of
flammeolus in the monotypic genus
Psiloscops has also been advocated (
http://books.google.be/books?id=0t0...arch_s&sig=ACfU3U2-7Mwzp2bwTOyRYyQpim-gUS0rLw ).
Laurent -