Bird analysis update 3

[David Marjanovic <david.marjanovic@gmx.at>]

Deleted *Hulsanpes* and *Ambiortus*. The strict consensus (of 1102 trees, 
length 209) is grateful:

+--allzero outgroup
`--+--+--*Caudipteryx*
  |  `--+--Scansoriopterygidae
  |     `--+--*Archaeopteryx*
  |        |--*Microraptor*
  |        |--basal Troodontidae
  |        `--+--*Rahonavis*
  |           |--*Shenzhouraptor*
  |           `--*Jixiangornis*
  `--+--*Sapeornis*
     `--+--*Yandangornis*
        |--*Omnivoropteryx*
        |--*Confuciusornis*
        |--*Jibeinia*
        `--+--Euornithes (see below)
           `--+--*Iberomesornis*
              `--Euenantiornithes (grass)

Euornithes
|--*Vorona*
|--*Patagopteryx*
|--*Yanornis*
|--*Yixianornis*
`--+--*Apsaravis*
  `--+--*Songlingornis*
     |--*Gansus*
     `--+--Hesperornithes
        `--+--*Ichthyornis*
           `--*Anas*

The resolved position of *Iberomesornis*, however, is likely an artifact of 
the restricted number of rearrangements -- in other words probably not all 
MPTs were found. Originally I forgot to remove *Ambiortus*, and in the 
strict consensus of that run Enantiornithes was not resolved at all, even 
though *Ambiortus* stayed in Euornithes.
       But even if Euenantiornithes is really present in the strict 
consensus, the situation still remains that, without reweighting, I haven't 
been able to tell the world anything new about enantiornith phylogeny. :-(

The majority-rule consensus resolves the polytomies as follows:

--100--*Microraptor*
   |--basal Troodontidae
   `--66--*Archaeopteryx*
       `--100--*Shenzhouraptor*
            `--56--*Rahonavis*
                `--*Jixiangornis*

--100--*Sapeornis*
   `--100--*Omnivoropteryx*
        `--100--*Yandangornis*
             `--92--*Confuciusornis*
                 `--100 *Jibeinia* + Ornithothoraces

Euenantiornithes
|--*Protopteryx*
`--60--*Longipteryx*
   `--60--*Noguerornis*
       `--72--*Eocathayornis*
          |--*Eoenantiornis*
          |--84--*Gobipteryx*
          |   `--97--*Longirostravis*
          |       `--*Liaoningornis*
          `--84--*Sinornis*
              `--91--unnamed Lebanon
                  `--91--*Halimornis*
                      `--91--*Concornis*
                          `--91--*Enantiornis*
                              `--57--94--*Neuquenornis*
                                  |   `--*Cuspirostrisornis*
                                  `--54--*Eoalulavis*
                                      `--86--*Soroavisaurus*
                                          `--95--*Avisaurus*
                                              `--*Vescornis*

Euornithes
|--*Vorona*
`--54--*Patagopteryx*
   `--100 *Yanornis* + rest

--100--*Songlingornis*
   `--87--*Gansus*
       `--rest

The Adams consensus has *Confuciusornis* and *Yandangornis* as sistergroups.

In individual trees *Eocathayornis* and *Noguerornis* jump around like 
crazy. For example, tree number 1000 has

+--*Gobipteryx*
`--+--*Eocathayornis*
  `--+--*Longirostravis*
     `--*Liaoningornis*

and

+--+--*Eoalulavis*
|  `--+--*Soroavisaurus*
|     `--+--*Avisaurus*
|        `--*Vescornis*
`--+--*Noguerornis*
  `--+--*Cuspirostrisornis*
     `--*Neuquenornis*

, not surprising considering how fragmentary *N.* is, and quite surprising 
considering how complete and seemingly obviously basal *E.* is. ~:-|

Now for 10 rounds of reweighting...