RE: Sereno's (2005) new definitions

[Tim Williams <twilliams_alpha@hotmail.com>]

Mickey Mortimer wrote:

> Not so.  You can run the Troodon formosus holotype premaxillary tooth in a 
> matrix as an OTU.  In fact, it's easy to do such because there are so few 
> characters to code.

I see what you mean.  Using that criterion, many nomina dubia could be put 
through a matrix so long as at least one codeable character is discernable.  
But it was my impression that a nomen dubium could not be used in a 
phylogenetic analysis on account of the fact that it has no taxonomic 
standing, and so cannot be a valid operational taxonomic unit.  In other 
words, the T in the OTU is missing - although as Mickey showed, you can 
still carry out the O part.  But I don't think it's worth the effort.

> After running this in PAUP (and deleting their Troodon OTU), the Troodon 
> holotype comes out in a clade with Byronosaurus, Sinornithoides and the two 
> Saurornithoides species.  Thus, an exact enough phylogenetic position has 
> been determined to define Troodontidae with.

When a name is declared to be a nomen dubium (or nomen vanum), the name is 
limited to the type specimen; no further material can be referred to the 
taxon that carries the name.  Designating a nomen dubium is a subjective 
choice, and taxa once declared to be taxonomically invalid are often found 
to be valid - and vice versa.  But the designation is not arbitrary, and a 
taxon remains taxonomically invalid so long as it's a nomen dubium.  Putting 
a nomen dubium in a matrix is entirely artificial - and maybe a bad thing to 
do if the resulting phylogeny is influenced by the inclusion of nomina 
dubia.

In other words, we could put a nomen dubium into a matrix - but why would we 
need to?  Why, when we could just replace the nomen dubium with an 
unquestionably valid taxon?

> It's true that a nomen dubium won't ever have an _exact_ position in an 
> analysis containing at least two taxa it's indistinguishable from, but very 
> few taxa have exact positions from analysis to analysis anyway.

This a big problem when it comes to poorly-represented taxa (i.e., taxa 
known only from fragmentary and/or poorly preserved specimens).  With these 
taxa there is the prospect of one day finding more material that can expand 
the taxon's dataset.  By contrast, a nomen dubium is a dead-end: no further 
material can be referred to it, by definition.

Jaime Headden wrote:

> As Tim said, the holotype tooth of *Troodon formosus* is not directly 
> comparable to the types of any other North American troodontid except for 
> *Pectinodon bakkeri*, which is very distinct in
> form to *T. formosus*. Contra Marjanovic's statement about correspondence 
> of the teeth with those of *Saurornithoides*, only *P. bakkeri* is closely 
> corespondent, that of *T. formosus* is not, arguing we would need to find 
> _in situ_ teeth identical to the type to more directly infer it's nature. 
> This is currently not possible.

Very interesting stuff, Jaime.  You actually go a little further than I 
do...  I was prepared to go along with the consensus that all troodontid 
material from the Campanian-Maastrichtian of North America was consistent 
with the presence of a single troodontid species.  If not, we can kiss 
_Troodon_ goodbye.

> Only two mandibles with _in situ_ teeth are known fromDinosaur Provincial 
> Park, and I would argue that of these, only one is similar enough to 
> suggest similarity to *T. formosus*, while the >other has distinct teeth 
> and distinct mandibular morphology.

How does _Polyodontosaurus_ fit into all of this?

> I agree that *Carnotaurus* should be used in Ceratosauria, Neoceratosauria, 
> Coelophysoidea, Abelisauria, Abelisauroidea, and Abelisauridae definitions, 
> since it's use is less ambiguous than the eponymous *Abelisaurus*,

I have to disagree with you here in terms of using non-eponymus taxa for 
coordinated family-level taxa.  _Carnotaurus_ should not be used as a 
specifier for Abelisauroidea and Abelisauridae at the expense of 
_Abelisaurus_, because we don't want to have a situation in which 
_Abelisaurus_ falls out of the Abelisauridae/oidea.  Hey, it could happen!  
But it could only happen if Abelisauridae/oidea is anchored in _Carnotaurus_ 
rather than _Abelisaurus_.

> While we should try to use eponyms as taxa, this may actually be 
> problematic for some taxa, as the original taxa for which the names are 
> formed are often less well known than is ideal: there is only 1 specimen 
> each of *Ornithomimus velox,* *Oviraptor philoceratops*, *Dromaeosaurus 
> albertensis*, and each is hardly complete or comprehensive in the material 
> known

In this case, just re-name the families.  For Ornithomimidae, Oviraptoridae 
and Dromaeosauridae let's have Struthiomimidae, Citipatiidae and 
Velociraptoridae.  Having said that, I think Ornithomimidae, Oviraptoridae 
and Dromaeosauridae are probably OK - so I'm not actually arguing for a 
change.  But I think Troodontidae is a disaster waiting to happen.

Elsewhere in dinosaur nomenclature there is a need for change.  The name 
Titanosauridae has already been effectively abandoned, which was a good move 
(irrespective of whether _T. indicus_ is valid or not).  Ceratopsidae should 
be junked too, given that _Ceratops_ is probably a nomen dubium (and unlike 
_Troodon_, we would be hard-pressed to find its position in ceratopsian 
phylogeny given the _C. montanus_ type material).  The name Hadrosauridae is 
unhelpful, given the status of the _Hadrosaurus_ type specimen.  If we are 
serious about erecting definitions that are stable, then some time-honored 
traditions have to be compromised, ICZN or no ICZN.  Ceratopsidae could 
become Centrosauridae, and Hadrosauridae could become Saurolophidae - but 
Ceratopsia and Hadrosauria would remain.

Cheers

Tim