[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Subject Index][Author Index]

Mickey's Coelurosaur Analysis Update

I've been working on my coelurosaur analysis for some time now.  I'm by no
means done, but I think I'm at a point where I should give you all an
update.  My preliminary goal is to include every valid character from the
recent major analyses.  These include the AMNH team's (2001-2002), Holtz's
(1994, 2000), Chiappe's (1996-2002), Novas' (1996-2002), Clark et al.'s
(2001-2002), Sereno's (1999), Maryanska et al.'s (2002), Rauhut's (2000),
Makovicky and Sues' (1998), Forster et al.'s (1998), Xu et al.'s
(1999-2000), Makovicky's (1995), Russell and Dong's (1994) and probably more
I forgot.  This sounds like a lot, but most of the characters are
overlapping in many of them, and some include characters inapplicable to the
taxa I'm using (eg. only useful within Ornithurae, or outside of
Coelurosauria).  Of course, I've modified many characters to ensure I think
their states are distinct, and many characters are my own as well.  Right
now, I have 165 cranial and axial characters coded for 113 taxa, including
three outgroups.  This is the same number of cranial and axial characters
used in my old analysis, and many are the same.  The taxa are all
genus-level OTU's, except Tyrannosauridae (Alectrosaurus+), Ornithomimidae
(more derived than Archaeornithomimus), Oviraptoridae, Mononykinae and
Ornithurae.  Technically, I also have a few species analyzed separately
(Caudipteryx zoui and C. sp. nov., Saurornithoides mongoliensis and S.
junior, Microraptor zhaoianus and M. gui).

Although only cranial and axial characters are included so far, there are
165 of them, which is more total characters than all but the AMNH team's
(79%), Chiappe 2001-2002 (98%), Sereno's (81%), Maryanska et al.'s (85%),
Clarke et al.'s (82%) and the more inclusive analyses of Holtz 2000 (43%)
and Rauhut (74%).  So it's not really that bad, especially considering the
fact I think mine are coded/structured better than the others (which sounds
arrogant, but who wouldn't think their own analysis was coded better) and
more taxa than any of them.  Of course, some taxa are coded for only a few
or no characters so far due to the lack of cranial and/or axial material or
description.  Also, it's not like 165 cranial/axial characters are
inherently inferior to 165 characters from all across the body, assuming
they are all independent.  Sure, there's a lot of data missing, but there is
for the "full representation" analysis with the same amount of characters as
well.  So I've decided to run a few preliminary analyses in PAUP just to see
what I get.  It's rather similar to Livezey and Zusi's recent neornithine
phylogeny- lots of characters from only a few anatomical areas completed so
far.  And like their paper, I can give a distribution of coded characters
and estimated uncoded characters left-
cranial and mandibular- 94 coded, ~100 left.
cervical- 20 coded, ~10 left.
dorsal- 14 coded, ~10 left.
sacral- 10 coded, ~3 left.
caudal- 27 coded, ~10 left.
appendicular- 0 coded, ~200 left.
integumentary- 0 coded, ~3 left.
So I expect to have ~500 characters when I'm done.

First, I ran it with only taxa from the AMNH analysis included (up to the
Incisivosaurus paper, making for 40 taxa).  In the AMNH runs, I included all
three outgroup taxa too (Monolophosaurus, Sinraptor, Allosaurus).  I was
able to include every taxon except Adasaurus (which is coded for 2
characters).  Also, they have multiple OTU's for my first four family-level
OTU's.  The result was 1211 most parsimonious trees (MPT's) of 500 steps
each.  The consensus was-
   |  |--Utahraptor
   |  |--Dromaeosaurus
   |  |--Deinonychus
   |  |--Saurornitholestes
   |  |--Velociraptor
   |  |--IGM 100/1015
   |  `--Sinornithosaurus
         |  |  `--+--Erlikosaurus
         |  |     `--Segnosaurus
         |  `--+--Confuciusornis
         |     |--Caudipteryx zoui
         |     `--+--Avimimus
         |        `--+--Incisivosaurus
         |           |--Microvenator
         |           |--Chirostenotes
         |           `--Oviraptoridae
            |  |  `--Pelecanimimus
            |  `--+--Garudimimus
            |     `--Ornithomimidae
               |  `--+--Alvarezsaurus
               |     `--Mononykinae
                  |  `--+--Microraptor
                  |     `--Archaeopteryx
                           `--+--Saurornithoides mongoliensis
                              `--+--Saurornithoides junior
Interesting, no?  Many characters uniting dromaeosaurs to troodontids and
birds must be in the appendicular skeleton.  All of their shared caudal
characters seemed to not matter.  But Microraptor managed to stay with
birds.  Confuciusornis didn't, and wound up as a basal oviraptorosaur due to
its convergent cranial similarities.  Avimimus is an oviraptorosaur too, and
not close to alvarezsaurids.  Alvarezsaurids are next to both birds AND
ornithomimosaurs.  How's that for a consensus of Chiappe and Sereno?

Next, I added several more complete and important taxa- Compsognathus,
Sinosauropteryx, Scipionyx, Caudipteryx sp. nov., Nomingia, Bagaraatan,
Bambiraptor, Scansoriopteryx (including Epidendrosaurus), Jeholornis,
Iberomesornis, Sinornis, Patagopteryx, Yanornis and Ornithurae (based on
hesperornithiformes and Ichthyornis).  This resulted in 9073 MPT's of 622
steps each.  The consensus was-
      |  |--Utahraptor
      |  |--Dromaeosaurus
      |  |--Deinonychus
      |  |--Saurornitholestes
      |  |--Velociraptor
      |  |--IGM 100/1015
      |  |--Bambiraptor
      |  `--Sinornithosaurus
         |  |  `--+--Segnosaurus
         |  |     `--Erlikosaurus
         |  `--+--Caudipteryx zoui
         |     `--+--Caudipteryx sp. nov.
         |        `--+--Avimimus
         |           `--+--Incisivosaurus
         |              |--Nomingia
         |              |--Microvenator
         |              |--Chirostenotes
         |              `--Oviraptoridae
            |  |  `--Pelecanimimus
            |  `--+--Garudimimus
            |     `--Ornithomimidae
               |  `--+--Alvarezsaurus
               |     `--Mononykinae
                  |  `--+--Byronosaurus
                  |     `--+--Sinornithoides
                  |        `--+--Saurornithoides mongoliensis
                  |           `--+--Saurornithoides junior
                  |              `--Troodon
                           |  `--+--Jeholornis
                           |     `--Archaeopteryx
There's my friend the Basal Coelurosaur Polytomy.  At least dromaeosaurs are
closer to birds than basal coelurosaurs in this tree.  Confuciusornis moves
into the proper position in the tree now that more pygostylians were
included.  The basic topology was still retained though- (dromaeo (enigmo
(ornithom (alvarez (troodo, aves))))).  This is unique as far as I know, and
shows how much our standard topology depends on appendicular information.
Bagaraatan ends up as a basal coelurosaur as in Holtz (2000), unlike my
previous analyses, Rauhut's (2000) and Longrich's (2001).  Scansoriopteryx
is right below Aves, as in Zhang et al. (2002) and the non-cladistic studies
of Czerkas and Yuan (2002) and myself.

I also tried it using only the taxa examined by Maryanska et al. (2002).  I
had to exclude Therizinosaurus and Borogovia due to the fact no cranial or
axial remains are known.  The result was 43 MPT's of 402 steps.  The
consensus was-
   |  `--Velociraptor
      |  |  |--Alxasaurus
      |  |  |--Erlikosaurus
      |  |  `--Segnosaurus
      |  `--+--+--Nanshiungosaurus
      |     |  `--Caudipteryx zoui
      |     `--+--+--Caudipteryx sp. nov.
      |        |  `--Avimimus
      |        |--Nomingia
      |        |--Chirostenotes
      |        `--+--Microvenator
      |           `--Oviraptoridae
         |  `--Ornithomimidae
               |  `--+--Sinornithoides
               |     `--+--Saurornithoides mongoliensis
               |        |--Saurornithoides junior
               |        `--Troodon
Which is the same basic topology I got from the AMNH taxon set.
Interesting.  Confuciusornis ended up as an avian, far away from
oviraptorosaurs.  Nanshiungosaurus landed next to Caudipteryx zoui, but only
19 characters can be coded for it so far.

Anyway, I hope you all enjoyed this.  I'll probably do another update when I
get most of the pectoral and forelimb characters coded.

Mickey Mortimer