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

Re: Defining clades like a REAL man (Was: New PaleoBios pa

> Date: Tue, 04 Oct 2005 12:29:54 -0500
> From: Tim Williams <twilliams_alpha@hotmail.com>
>>>> Ye-es.  I think it's a matter of philosophy whether you like such
>>>> magic disappearing clades or not.  I am not personally wild about
>>>> them.
>>> Why on earth not?
>> Warning: high subjective opinions follow.  To me, such a
>> self-destructing clade definition seems like a coward's way out.
>> It encourages people to go ahead and name the clade irrespective of
>> the support for the topology in which is was conceived, because the
>> magic will make it vanish if it goes wrong.  But I think it's
>> better to wait until all the phylogenetic ducks are in a row before
>> naming the clade in the first place.
> In an ideal world, all clades would be well-supported and stable in
> content.  In the real world, a change in the polarity of a
> relatively small number of characters can drastically alter the
> topology of a phylogeny.

That is why defining clades well requires taste and forsight.

> If you're waiting for every clade to be rock-solid in content and
> position, then I think you'll be waiting till the cows come home.
> The monophyly of Diplodocoidea looks set in stone, but I wouldn't
> hold my breath when it comes to resolving the relationships of many
> Chinese taxa (_Euhelopus_, _Mamenchisaurus_, _Omeisaurus_,
> _Shunosaurus_, etc).

... which is precisely why it would be premature to erect a clade for
those guys.

In any case, there are other issues here.  You say "The monophyly of
Diplodocoidea looks set in stone".  Well, sure it is -- since
Diplodocoidea is defined as the stem clade (_Diplodocus_ not
_Saltasaurus_), it can't _not_ be monophyletic.  Given sufficiently
weird topologies, it might consist of _Diplodocus_ alone, or of all
sauropods except _Saltasaurus_, but it would still be monophyletic.

You know this, of course; so what do you mean when you talk about the
monophyly of Diplodocoidea looking secure?  I think (stop me if I'm
wrong) that what you really mean is that the _content_ of that clade
is stable -- the Diplodocoidea as "traditionally" conceived (i.e. in
the last ten years) has not had too much happening in the way of taxa
dropping out or others coming in.

Now that is true deep within the clade, for sure: no-one has
cladograms in which the diplodocine _Barosaurus_, the non-diplodocine
diplodocid _Apatosaurus_, the non-diplodocid flagellicaudatan
_Dicraeosaurus_ or the non-flagellicaudated diplodocimorph
_Limaysaurus_ come out closer to _Saltasaurus_ than to _Diplodocus_;
so those genera all stay diplodocoid and we're all happy.  But of
course the nemegtosaurs that Upchurch (1995) considered diplodocoid
are not considered to be titanosaurs closely related to
_Rapetosaurus_.  And _Haplocanthosaurus_ is a diplodocoid only in
months with an "R" in them.

The moral of this story is that you shouldn't defined Diplodocoidea on
_Haplocanthosaurus_ or _Nemegtosaurus_ (and, thankfully, Wilson and
Sereno (1998) didn't).  The typical situation is that you have a big
clade with some deeply nested taxa that are well represented and well
resolved  (_Diplodocus_ in this case), some that are nearer the base
of the clade and not at all strongly supported (_Haplocanthosaurus_)
and others in between that look more or less secure at the moment, but
that might yet drop out elsewhere in another analysis (e.g. the
rebbachisaurs).  Definitions should be anchored on the deeply nested
and well resolved taxa.

Right, so where does that leave us with respect to the Chinese
sauropods?  The answer is, _nowhere_, because there is not one single
relationship between the Big Four (_Mamenchisaurus_, _Omeisaurus_,
_Euhelopus_ and _Shunosaurus_) that is consistently recovered by
multiple analyses.  It's true: revisit Russell and Zheng 1994, Calvo
and Salgado 1995, Upchurch 1995 and 1998, Salgado et al. 1997, Wilson
and Sereno 1998, Curry Rogers and Forster 2001, Wilson 2002 and
Upchurch et al. 2004 (pant, pant) and you'll see all sorts of weird
relationships between these four taxa, and all sorts of weird
positions for them.  In conclusion, you can't pick a well-resolved
representative of the group you want because (A) there isn't one, and
(B) "the group" might not exist anyway.

Finally, then (sorry for the long digression), I think the honest
thing to do is just hang back and not propose clades for the Chinese
sauropods.  It's more realistic just to say "the Chinese sauropods",
since that's what's you usually mean anyway.

And that's what I think.  :-)

> For example, it doesn't take much to convert two sister taxa into
> two successive paraphyletic taxa.  This can have huge ramifications
> for a clade that has these two taxa as specifiers.  The
> _Mamenchisaurus_+_Omeisaurus_ clade is a good example.  If these two
> taxa switch from being sister taxa to paraphyletic, then a
> Mamenchisauridae that is defined with _M_ and _O_ as specifiers can
> go from a clade that includes two genera to a clade that contains
> most of the Eusauropoda.

Er ... didn't I say exactly that a couple of days ago?

Anyway ... the solution to this problem is the same one that the
Doctor said to the patient who compained "It hurts when I do this":
Don't Do That, Then.


Finally (having already said "finally" once, a couple of paragraphs
ago), I just want to mention this:

> There are many traditional family-level groups that contain genera
> that we can be fairly sure are closely related, but which are now
> regarded as paraphyletic: Megalosauridae, Andesauridae,
> Brachiosauridae, Hypsilophodontidae, Iguanodontidae,
> Protoceratopsidae.

Pet peeve: the "evidence" that Brachiosauridae is paraphyletic is a
sequence of unsubstantiated assertions in Salgado and Calvo 1997,
backed up by the phylogenetic analysis in the companion paper Salgado
et al. 1997, which claims to have recovered a paraphyletic
Brachiosauridae based on the inclusion of two (2) potentially
brachiosaurid OTUs -- _Brachiosaurus brancai_ and _Chubutisaurus_.  Of
these, the latter had already been reassigned (Bonaparte 1986b) to
Titanosauridae, so the fact that Salgado et al. recovered it closer to
titanosaurs than to _B. brancai_ is hardly the death-blow to
Brachiosauridae that it has often been portrayed as.  (Also, there is
only one rather unconvincing synapomorphy linking _Chubutisaurus_ with
titanosaurs in this analysis.)  Now Salgado et al. _might_ have been
right about the paraphyly/polyphyly of Brachiosauridae, and indeed of
the genus _Brachiosaurus_.  But there is certainly no evidence for it
in their papers (nor published elsewhere to my knowledge).

Sorry, just had to get that off my chest.

> Some of these families can be utilized and defined for less
> inclusive clades.  Yet, the internal relationships of these
> paraphyletic arrays are prone to rampant flip-flopping, and so it
> would be hard to pin down two genera that could be used as
> specifiers.  How do you define a monophyletic Protoceratopsidae, for
> example, when the closest relative(s) of _Protoceratops_ change from
> one analysis to the next?

Wilson and Sereno's solution in the case of Brachiosauridae was to
define a stem-based clade (_Brachiosaurus_ not _Saltasaurus_) rather
than a node; which seems eminently sensible.  (I don't know enough
about ceratopsians to know whether that approach could be made to work
well for Protoceratopsidae, though.)

 _/|_    ___________________________________________________________________
/o ) \/  Mike Taylor  <mike@miketaylor.org.uk>  http://www.miketaylor.org.uk
)_v__/\  Remember that "ministry" means "service": if what you do in the
         church doesn't serve the church, it's not a ministry.