Re: Ankylosauromorpha page

[Mike Taylor <mike@tecc.co.uk>]

> Date: Sat, 19 Jan 2002 19:43:28 -0800
> From: "Mickey_Mortimer11" <Mickey_Mortimer11@email.msn.com>
> 
> Isn't the _obvious_ phylogenetic taxonomy for Ankylosauria-
> Ankylosauria- everything closer to Ankylosaurus than Stegosaurus
> Nodosauridae- everything closer to Nodosaurus than Ankylosaurus and
> Polacanthus
> Polacanthidae- everything closer to Polacanthus than Nodosaurus and
> Ankylosaurus
> Ankylosauridae- everything closer to Ankylosaurus than Nodosaurus and
> Polacanthus
> What would have been so hard about that?

... feeling suitably chastened by Thomas Holtz's gracious rebuke, I
venture to dip my poor little toe in the freezing cold waters of
classification once more ...

I have a question about Mickey's suggestion here.  It's nothing to do
with ankylosaurs _per se_, but just with how different taxa can be
related.

I guess we're all familiar(*) with node-stem triplets in which a
node-based taxon is neatly partitioned into two parts, which are
stems.  For example, the node {Neornithes + Allosaurus} = Avetheropa
is partitioned into the stems {Allosaurus <-- Neornithes} =
Carnosauria and {Neornithes <-- Allosaurus} = Coelurosauria.

(*) If that didn't make any sense to anyone, there's a primer at
http://www.miketaylor.org.uk/dino/faq/s-class/phyletic/index.html

But what Mickey's talking about here implies a sort of node-stem
quadruplet:

Nodosauridae = stem {Nodosaurus <-- Ankylosaurus, Polacanthus}
Polacanthidae = stem {Polacanthus <-- Nodosaurus, Ankylosaurus}
Ankylosauridae = stem {Ankylosaurus <-- Nodosaurus, Polacanthus}
        ... and therefore we can't help but think about ...
SomeNode = node {Nodosaurus + Ankylosaurs + Polacanthus}

(PLEASE, let's not get sidetracked into discussion stuff like whether
SomeNode is Ankylosauroidea, or equivalent to Ankylosauria, or
whatever.  And no, I am NOT trying to have people consider the name
"SomeNode" as published :-)

My question is this: if that kind of three-fold splitting of stems is
used, don't people feel uncomfortable about the left-over stuff on the
stem closest to the root?  Suppose it happens that ankylosaurs and
polacanths are more closely related to each other than to nodosaurs,
like this:

        Nodo.  Pola.  Anky.
          \     \    /
           1     2  3
            \     \/
             \    /
              \  4
               \/
                \

Then in this topology, it's clear that everything on the branch I've
labelled 1 is in Nodosauridae, everything on 2 is in Polacanthidae and
everything in 3 is in Ankylosauridae.

But what about the poor animals on branch 4 (and on further branches
that branch off from there, of course)?

* They're certainly not nodosaurids, because they're more closely
  related to both Polacanthus and Ankylosaurus than to Polacanthus;
* They're not polacanthids, becuase they're just as closely related to
  Ankylosaurus as to Polacanthus; and
  [you can work this out for yourselves, right?]
* They're not ankylosaurids, becuase they're just as closely related
  to Ankylosaurus as to Polacanthus.

(But they are, of course SomeNode-ids :-)

In effect, what we have here is an aggravated form of the node-stem
triplet discontinuity in which the MRCA itself is a member of the
node, but not of either supposedly-partitioning stem.  We just shrug
at that problem, because it's only one animal (as in literally on
specimen, which we'll never find).  But in this case, whole subtrees
could be affected.

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?

Thanks for listening.  Hope this wasn't too stupid a question.

 _/|_    _______________________________________________________________
/o ) \/  Mike Taylor   <mike@miketaylor.org.uk>   www.miketaylor.org.uk
)_v__/\  "As with most things in life, punctuation doesn't have to
         be perfect, just good enough to get by" -- Terry Cox