Tawa hallae: everything you know about basal saurischians ISN'T wrong...

[Michael Mortimer <mickey_mortimer111@msn.com>]

I just wrote a reply on Parker's Chinleana blog about a pet peeve of mine, and 
your post title sums it up perfectly, Tom.  Before I begin, I want to clarify 
this isn't meant to specifically pick on Tom or the Tawa authors.  It's a 
problem I see as common in dinosaur paleontology.  Namely, whenever a new 
analysis comes up with a different result than an old analysis, it's seen as 
proving the old topology was wrong.  But it's rarely so simple.
 
First you have to ask if the new analysis includes all the relevent taxa the 
old analysis did.  Nesbitt et al.'s doesn't include Guaibasaurus, Panphagia, 
"Dilophosaurus" sinensis, Dracovenator, Lophostropheus, Elaphrosaurus, 
Berberosaurus or Spinostropheus for instance.
 
Then you have to make sure the new analysis includes all the relevent 
characters the old analysis did.  Nesbitt et al.'s excludes the following 
coelophysoid characters found by Tykoski (2005) for instance- skull elongate 
(>3 times longer than tall); antorbital fenestra elongate (>25% skull length); 
premaxillary tooth row ends anterior to external naris; premaxilla and maxilla 
loosely articulated; anterior tip of maxillary alveolar margin curves 
mediodorsally; anterior tip of maxillary alveolar margin curves mediodorsally; 
anterior dentary teeth enlarged; narrow cervical epipophyses which project 
posterolaterally; no axial diapophyses; axial parapophyses reduced or absent; 
pubic peduncle of ilium with two facets- facing anteriorly and ventrally; 
distal tips of pubes lack contact; anterior trochanter conical.  Maybe some of 
these are invalid or correlated with other included characters, or found in 
Tawa.  But it only takes six of these thirteen characters to move Dilophosaurus 
back to Coelophysoidea in Nesbitt et al.'s paper.  Am I saying Dilophosaurus is 
a coelophysoid?  No, because Tykoski probably didn't include every supposed 
neotheropod character of Dilophosaurus.  Most of the big recent analyses have 
found basically ambiguous support for either option- Rauhut (2003) found 
Dilophosaurus to be equally well supported as a coelophysoid when Shuvosaurus 
was (correctly) excluded, Yates (2005) found dilophosaurids were coelophysoids 
with only one added step, Smith et al. (2007) found dilophosaurids were 
coelophysoids with only six added steps, and Ezcurra and Novas (2006) found 
equal bootstrap support for either option. 
 
Even after those are taken care of, coding differences probably exist and must 
be evaluated.  But you don't even need to go through those tedious steps to see 
the basic picture.  Let's say Nesbitt et al. coded everything correctly and 
didn't leave out any relevent taxa or previously used characters.  It still 
only takes one more step to place Liliensternus in Coelophysoidea, and two more 
steps to place Zupaysaurus there.  That's such weak support that I'd say it's 
basically ambiguous whether those taxa are coelophysoids or not.  Imagine how 
easy it would be to find one more character that optimizes as a synapomorphy of 
Liliensternus + Coelophysidae.  It doesn't even have to be unique to 
Liliensternus and Coelophysidae, it just has to most parsimoniously map out as 
a synapomorphy of that clade.  
 
The basic message is that each new analysis does not make prior analyses wrong. 
 One point to Nesbitt et al.'s credit is that they ran constraint analyses to 
determine just how unlikely alternate arrangements are.  We need to keep these 
in mind whenever we look at a cladogram.  Cladograms always show us the MOST 
parsimonious arrangement, but perhaps it's more useful to be aware of which 
nodes are SIGNIFICANTLY MORE parsimonious.  How significant is significant?  
There are tests to determine this (used in molecular analyses), though I've 
rarely seen them used in Mesozoic dinosaur analyses.  My previous post 
http://dml.cmnh.org/2008Mar/msg00195.html might suggest anything 10 steps or 
more is implausible.  After all, in that tree (which combined Tykoski's, Smith 
et al.'s, Rauhut's, Ezcurra and Novas', etc. matrices) it took 8 and 9 steps 
respectively to place Dilophosaurus and Zupaysaurus in the position they have 
in Nesbitt et al.'s paper.  I'm currently working on a section of my website 
that will examine just how well various matrices support each of their clades, 
which should illuminate matters.
 
Mickey Mortimer
The Theropod Database- http://home.comcast.net/~eoraptor/Home.html