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

Epidendrosaurus Phylogenetic Discussion Pt. 2

Dear all,

  Here's the first fo two analysis, this one unlike the next focusing on
maniraptorans. The next, Holtz' GAIA analysis, has more characters but is
a general theropod discussion. So that should be interesting.

  Mickey and I last evening and this ran about twenty runs of various
versions of the Xu et al. (2002) *Sinovenator* analysis (with
*Microraptor* added from Hwang et al. 2002) including *Epidendrosaurus
ningchengensis* with an accounted 14% of the complete codable matrix (29
of 206), the smallest percentage of included taxa ... even *Alvarezsaurus*
has more codings. On various runs I experimented with removing taxa
(namely some poorly known deinonychosaurs), ordering all states, and
unordering them. These produced various results, and steps varied from 633
to 584. The most taxa with the fewest steps is, obviously, the preferred

  In half of the analyses, Mickey and I got *Epidendrosaurus* in a
trichotomy with pygostylian birds (*Confuciusornis* and modern birds,
etc.) and *Archaeopteryx*; this is close to the original placement in the
Zhang et al. (2002) analysis. In the other half, it was a basal
maniraptoran, and Maniraptora excluding *Ornitholestes* formed an intense
polytomy; steps varied in both between 630-580. Several analyses produced
584 steps, and several 631-633, reflecting a concentration of data centers
on taxonomic placement. Other features of the trees included
Dromaeosauridae being a self-contained polytomy, and often not including
*Microraptor*; Troodontidae including *Sinovenator* in a distinct
comb-like structure in agreement with the other published results;
*Archaeopteryx* and birds in a monophyletic Aves without other taxa;
*Microvenator* and *Caudipteryx* as oviraptorids, and *Avimimus* is a
caenagnathid; *Alvarezsaurus* outside of a monophyletic Mononykinae.

  I ran the matrix with 19 ordered and 187 unordered characters, and
ignored characters relating to the coracoid and furcula based on
disagreement between the both of us. Though erlier I advocated the
presence of an astragalus, this was also not coded. Characters were not
coded from the Zhang et al. (2002) *Epidendrosaurus* analysis (adding
*Epidendrosaurus* to the Forster et al. (1998) *Rahonavis* (there and in
both analysis, *Rahona*), results of which I presented earlier along with
recodings based on lacks of assumption of features).

  Because a tree with ordered multistate characters is prefered and where
one can provide that a character may not be ordered, I ran the analysis
with stated ordering in Xu et al. (2002). This run under PAUP* 4.0 (beta
10) (Swofford, 1998) produced a whopping 101,430 trees. Statistics for the
analysis are:

steps: 584
consistency index: 0.4332
homoplasy index: 0.5668
retention index: 0.7029

and produced the following concensus tree, which is remarkably resolved in
some areas:

        |  `--Tyrannosaurus
           |  |--Pelecanimimus
           |  |--Garudimimus
           |  `--+--Struthiomimus
           |     `--Gallimimus
                 |  `--Confuciusornis
                 |  |--Sinornithosaurus
                 |  `--+--Unenlagia
                 |     |--Velociraptor
                 |     |--Saurornitholestes
                 |     |--Deinonychus
                 |     `--Dromaeosaurus
                 |  |--Byronosaurus
                 |  `--+--Troodon
                 |     `--+--Saurornithoides junior
                 |        `--Saurornithoides mongoliensis
                 |  `--+--Shuvuuia
                 |     `--Mononykus
                    |  |--Segnosaurus
                    |  `--Erlikosaurus
                       |  `--Avimimus
                                |--Oviraptor philoceratops
                                |--Oviraptor mongoliensis
                                `--GI 100/42

  Variations in this analysis are diverse, though most of the variation
occurs in the oviraptorosaur--segnosaur part of the tree, and increases
the tree count horribly. *Epidendrosaurus* moves about only a little,
however, and two main positions reflect it's concensus position. Position
1) a basal alvarezsaur relationship, 2) a basal avialian relationship;
also, several positions reflect the variable conditions which are less
frequently found: 3) a basal Eumaniraptoran position; 4) a sister-group
relationship with *Rahonavis* and other variations on the second position;
5) and a sister-group to the oviraptorosaur-segnosaur group.



  In each of these trees, Alvarezsauria is always outside the Oviraptor +
Bird clade, and are basal Maniraptora. This tree occurs in about 35% of
the total. Some trees also provide *Alvarezsaurus* outside a
*Epidendrosaurus* + *Mononykus* clade.



  In some of these trees, Avialae is either sister-group to
Deinonychosauria, or Troodontidae. A concensus position would be a
trichotomy of Troodontidae, Dromaeosauridae, and Avialae. *Unenlagia* is
always a dromaeosaurid. This tree occurs less frequently than above, but
in over 30% of the total.



  This tree occurs very rarely and in less than 5% of the total.




  |  `--Rahonavis

  These trees occur very rarely, in less than 5% of the total.

and 5)

  This is one of the rarer trees, in less than 2% of the total.

  With this, and the recent proposition where Alvarezsaurs are
maniraptoran but not avialian, the trees provide a general concensus where
*Epidendrosaurus* may be a eumaniraptoran, or an alvarezsaur sister-group.
Characters that diagnose these groups are very limited:

  An alvarezsaur relationship is supported by only 2 characters in all
trees except some where *Epidendrosaurus* was the sistergroup to
mononykines, where it also included all the alvarezsaur synapomorphies:
63(2) and 125(1).

63.     All maxillary and dentary teeth without serrations.
125.    Digit I bears large ungual and unguals of other digits distinctly
smaller . (Mickey and I are of some disagreement on the coding on this,
though it is clear *Epidendrosaurus* has a larger first claw than second.
In this manner, the character defines some clades, but not others, and
clearly supports *Epidendrosaurus* and *Shuvuuia*, whereas other
alvarezsaurs lack second digits for comparison. However, other taxa have
such grossly large first claws, including *Ingenia* and *Allosaurus*,
which the matrix does not code for. This will be corrected.)

  An avialan relationship is supported by three characters [63(2), 169(1),
190(1)], with five additional characters when *Epidendrosaurus* is next to
*Rahonavis* [98(2), 148(2), 190(1), 191(1), 204(1)].

63.     All maxillary and dentary teeth without serrations.
98.     No more than 25 caudal vertebrae. (The exact caudal count is unknown,
so this is supportive only through optimization, and doesn't count.
*Epidendrosaurus* was coded "?".)
148.    Lesser trochanter and greater trochanter completely fused (or absent)
to form crista trochanteris. (Same as above, where the trochanters of the
femur are not apparent. *Epidendrosaurus* was coded "?".)
169.    Metatarsal I attaches to posterior surface of distal quarter of
metatarsal II.
190.    Humerus longer than scapula.
191.  Preacetabular portion of ilium markedly longer (more than 2/3 of
total ilium length) than postacetabular part. (Same as above, where the
ilium is not preserved. *Epidendrosaurus* was coded "?".)
204.  Pubic apron less than 1/3 of shaft length. (Same as above, where
pubis is not preserved. *Epidendrosaurus* was coded "?".)

   This shows that optimization is the main source of relationship, and
the avialian position is actually supported only by 3 characters, and the
*Rahonavis* position by 2. However, the support for this group is as
parsimonious as for the alvarezsaur clade. Support for the eumaniraptoran
basal position is with character 190(1). Relationship with the
oviraptorosaur--segnosaur clade is supported by 50(2):

50.     Symphyseal region of dentary strongly recurved medially.

  This is also known in troodontids (*Troodon* and *Saurornithoides*), and
I strongly disfavor this scenario.

  Another run of the same matrix, but with the codings corrected for 5
characters after discussion with Mickey, reduced the percentage of coded
characters to the matrix to 12%, and I get an identical topology with 587
steps (four more than above), and stats of 0.4327 for the CI, 0.5673 for
the HI, and 0.7035 for the RI, and there are only 2,835 trees. The
following is the concensus tree produced:

        |  `--Tyrannosaurus
           |  |--Pelecanimimus
           |  |--Garudimimus
           |  `--+--Struthiomimus
           |     `--Gallimimus
                 |  `--+--Patagonykus
                 |     `--+--Shuvuuia
                 |        `--Mononykus
                 |--+--Epidendrosaurus, position 1
                 |  `--+--+--+--Microraptor
                 |     |  |  |--Sinornithosaurus
                 |     |  |  `--+--Unenlagia
                 |     |  |     |--Velociraptor
                 |     |  |     |--Saurornitholestes
                 |     |  |     |--Deinonychus
                 |     |  |     `--Dromaeosaurus
                 |     |  `--+--Sinovenator
                 |     |     `--+--Sinornithoides
                 |     |        |--Byronosaurus
                 |     |        `--+--Troodon
                 |     |           `--+--Saurornithoides junior
                 |     |              `--Saurornithoides mongoliensis
                 |     `--+--Rahonavis
                 |        `--+--Archaeopteryx
                 |           `--Confuciusornis
                 `--+--Epidendrosaurus, position 2
                       |  |--Segnosaurus
                       |  `--Erlikosaurus
                          |  `--Avimimus
                                `--+--Oviraptor mongoliensis
                                      `--+--Oviraptor philoceratops
                                            `--GI 100/42

  Position 1 is advocated by character 190(1) [see above] and position 2
by 50(2) [same].

  Mickey will seperately present what he found using different
interpretations of the character codings, so this will be a study in how
two people can get different results with the same analysis. Much fun


  Hwang, S.H.; Norell, M.A.; Ji Q.; & Gao K.-q. 2002. New specimens of
Microvenator zhaoianus (Theropoda: Dromaeosauridae) from northeastern
China. _American Museum Novitates_ 3381: 44pp.
  Xu X.; Norell, M.A.; Wang X.-l.; Makovicky, P.J.; & Wu X.-c. 2002. A
basal troodontid from the Early Cretaceous of China. _Nature_ 415:
782-784. (w/ supplementary info)
  Zhang F.c.; Zhou Z.-g.; Xu X.; & Wang X.-l. 2002. A juvenile
coelurosaurian theropod from China indicates arboreal habits.
_Naturwissenschaften_. published online: DOI 10.10007/s00114-002-0353-8.


Jaime A. Headden

  Little steps are often the hardest to take.  We are too used to making leaps 
in the face of adversity, that a simple skip is so hard to do.  We should all 
learn to walk soft, walk small, see the world around us rather than zoom by it.

Do You Yahoo!?
Yahoo! Finance - Get real-time stock quotes