Qiliania a confuciusornithid?

Another quiet month due to working on publishable projects, but here's something I noticed when getting an update for my site ready.

Ji et al. (2011) described the new bird taxon Qiliania, based on a pelvis and hindlimbs.  They included it in O'Connor et al.'s Shanweinao matrix, along with three unnamed Xiagou enantiornithines, Soroavisaurus and Archaeorhynchus.  It emerged as an enantiornithine, which is what the authors describe it as.  But then there's this statement- "DNHM D2522 (the holotype of Rapaxavis pani; Morschhauser et al., 2009) and PKUP-V1069 (the holotype of the basal ornithuromorph Longicrusavis houi; O’Connor, Gao & Chiappe, 2010) were removed (see Supporting Information)."

Well that's weird.  Why would you remove those two taxa?  Unfortunately, the supplementary information just lists tree descriptions and gives the codings for the added taxa.  They're the only undescribed taxa from O'Connor et al.'s Shanweinao matrix, but O'Connor is an author of Qiliania too.  O'Connor was first author of the Longicrusavis description, which was submitted almost a year before the Qiliania paper was submitted.  The Rapaxavis description came out two months before the Qiliania paper was submitted.  So I can't see any reason the Qiliania authors wouldn't trust the codings and have the resources to check them.

In any case, the interesting point is that when Rapaxavis and Longicrusavis are left in, the cladogram is different.  Confuciusornithids, 'Jeholornis' (= Shenzhouraptor) and Sapeornis now form successively more distant outgroups to Ornithothoraces.  Zhongornis is a confuciusornithid as I proposed.  Relationships in Enantiornithes are almost completely different, with all the CAGS specimens avisaurids, a clade of Las Hoyas taxa with Eoenantiornis sister to it and a Gobipteryx+Vescornis clade.  In Ornithuromorpha, hongshanornithids and songlingornithids are sister taxa.  But most important for a paper on Qiliania, that genus is now a confuciusornithid.

This is based on the short ischium, posteriorly excavated tarsometatarsus (also in avisaurids) and J-shaped metatarsal I (also in some enantiornithines).  Forcing it to be an enantiornithine is only one step longer (which rearranges enant topology again), so I'm not arguing Ji et al. were wrong to place it in that clade, but I do wonder why they excluded taxa which were already coded.  At the least, this shows the importance of including taxa in analyses and suggests Qiliania may be better placed as Pygostylia incertae sedis.


Incidentally, O'Connor et al.'s matrix is also the one Kurochkin et al. (2011) used for their Mystiornis paper.  They added Anchiornis, Mei, Avisaurus, Vorona and Mystiornis and found the latter four formed a clade one node closer to Aves than Archaeopteryx.  With Anchiornis being in a polytomy with this clade and more derived birds, this suggests a systematic coding error by the authors for their added taxa, which can unfortunately not be confirmed since the matrix was not published.  I added Mystiornis, Avisaurus and Vorona myself and found the latter two fall out in their normal positions (derived enantiornithine and basal ornithuromorph), while Mystiornis is an ornithothoracine outside of Longipterygidae and Hongshanornis+Aves.  When these taxa and Ji et al.'s taxa are all ran together, Mystiornis is sister to Avisauridae (similar to Cau's Megamatrix) and Qiliania stays as a confuciusornithid.