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

Re: Ankylosauromorpha page

Mike Taylor wrote-

> But what about the poor animals on branch 4 (and on further branches
> that branch off from there, of course)?
> So finally, my actual question is just this: doesn't anyone else but
> me feel uncomfortable about using sets of definitions which leave some
> specimens in this taxonomic no-man's-land?

The problem stems from a couple things.
One is the classic view that all ankylosaurs are either nodosaurids or
ankylosaurids.  Taxa like Minmi and Liaoningosaurus show us that some are
more basal than Nodosaurus and Ankylosaurus, making this view false.
Another problem is that Nodosaurus has never been used in a good recent
phylogenetic analysis.  Only Carpenter (2001) included it, but he assumed it
was part of a monophyletic Nodosauridae with many other taxa a priori, so
never tested which taxa might be closer to Nodosaurus than other taxa.  This
makes it currently impossible to define Nodosauridae eponymously and still
know what is included.  A clade of taxa usually classified as nodosaurids is
found by both Kirkland (1998) and Hill et al. (2001), but it had <50%
bootstrap support in the latter at least.  So nodosaurids as commonly seen
have a good chance of being paraphyletic.
Also, the recent consensus Polacanthidae is a clade containing Polacanthus,
Hoplitosaurus, Gastonia, Mymoorapelta, Gargoyleosaurus and even Hylaeosaurus
hasn't been well tested.  At least Gastonia and Mymoorapelta were found to
be close to Polacanthus by Kirkland (1998), but Hill et al. (2001) didn't
even include Polacanthus, but found Gargoleosaurus and Gastonia didn't form
a clade.  Thus, Polacanthidae could easily be para- or polyphyletic, though
its members seem closer to ankylosaurids than taxa usually allied with
So what are we to do?  Perform a large cladistic analysis with genus-level
OTU's including both Polacanthus and Nodosaurus.  Until that happens though?
We obviously must define Nodosauridae and Polacanthidae based on Nodosaurus
and Polacanthus because there is a good chance the families won't include
their eponyms if we do otherwise.  We could always choose another possible
nodosaurid to name the family after and define it based on that genus, like
Sauropeltidae (Ford, 2000), Panoplosauridae (Nopsca, 1929) or Edmontoniidae
(Russell, 1940).  That would be an odd transition from the usage in the past
few decades though.  We could solve the possible paraphyletic polacanthid
problem (ah, alliteration) by expanding the Ankylosauridae to be everything
closer to Ankylosaurus than Nodosaurus (or whatever nodosaur you named the
basal family after).  This actually sounds like a good solution to me and
would nullify problems of just how polacanths, shamosaurs and other
ankylosaurids are related to each other (we could worry about subfamilies
later).  If we want our precious traditional Nodosauridae, Ankylosauridae
and now Polacanthidae though, the obvious choice is to define each by their
eponyms and exclude the others' eponyms.  This ensures each family will
exist (even if monotypic) and leaves room for other families to be named for
taxa on Mike's branch 4 or the basal branch, once stable phylogenies are
found.  But we don't know what may be in that "taxonomic no-man's-land"
until a good phylogeny comes out.  According to Kirkland, it's Shamosaurus
(so we would ressurect Shamosauridae if there were more taxa to include,
like Cedarpelta or Gobisaurus).  It's the only way to make safe definitions
at this point while keeping traditional family names.
Or (here's a novel idea) we could always..... hold off on definitions until
someone performs that big analysis.  Until then, definitions are rather
useless anyway, since we don't know what taxa are nodosaurids and

Tracy Ford wrote-

> First, Stegosaurs and Ankylosaurs are not related, not even closely.
> to the missbelief of this and cladistics it keeps getting perpetuated (I'm
> just say'n in this case :) ).

How about some evidence for this?

> Scelidosaurus is an ankylosauroid (or ankylosaur) and NOT at the junction
> thyrophean, which doesn't exist. Norman is working on this and should have
> the paper out soon. Jim Kirkland also believes this (pers. comm..).

I'm still curious as to why this is just becoming accepted recently.  I
think it sounds pretty cool, but it's never been tested!  Galton (1990)
listed at least twenty-four characters present in stegosaurs and
ankylosaurs, but not Scelidosaurus (supporting the Stegosauria +
Ankylosauria clade Eurypoda).  Carpenter (2001) counters with five(!)
characters uniting Scelidosaurus and ankylosaurs, ignoring the eurypod
characters and omitting them from his analysis.  Anyone else see a problem?

> And where is Stegopeltinae? Oh yea, I didn't use a cladgram so we all can
> ignore this, right? The armor is different and I will be giving a talk at
> the Tate Symposium on Ankylosaurus and ankylosauridae and showing how
> different the scutes are and their similarities.

As Mike Keesey was thinking of Kirkland's (1998) analysis (or something
close to that time, since polacanthids have been recognized as a distinct
family recently) when he presented the
shamosaurine-polacanthine-ankylosaurine hypothesis, Stegopeltinae was
nowhere to be seen at the time, being created by you just last year.  They
may deserve their own subfamily, but if you are right and they are most
closely related to the Ankylosaurus-Euoplocephalus clade, I would think
keeping them in the Ankylosaurinae would be best (lest we have a
preponderance of ankylosaurid subfamilies).

Mickey Mortimer