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

Re: a closer look at Hanson 2006

----- Original Message -----
From: "David Peters" <davidrpeters@earthlink.net>
Sent: Saturday, June 10, 2006 4:00 PM
Subject: a closer look at Hanson 2006

So, Mike, good to see your work, but Triassic forms are sprinkled
around in your analysis, but they are all basal in mine. Also Mid
Jurassic forms are higher than some Late Jurassic forms here, but
they are in chronological order in mine.

I don't think the pterosaur fossil record is so good that we should expect a good stratigraphic fit of their phylogenetic tree.

When a phylogenetic tree has poor stratigraphic fit, there are three possibilities:

- The tree is wrong.
- The fossil record is incomplete -- there actually was something to fill up the ghost lineages, it just isn't discovered yet (if ever).
- Both.

The third is the general case and should therefore be treated as the null hypothesis.

While Dorygnathus and
Rhamphorhynchus appear similar, there are major differences with
better matches among other pterosaurs. For instance the sternal
complex on all  Campylognathoides and all Rhamphorhynchus are large
and similar in morphology. Not so in Dorygnathus, which has a tiny
sternal complex.

Wild shot into the blue: This screams "ontogenetic character". It may not be -- but please explain why it isn't. (Sorry, I don't have time to figure that out on my own.)

The back- and-forth morphology that appears here should be a red flag
that something is wrong.

Maybe there just isn't enough phylogenetic signal in this character. If you mapped fur color on a mammal tree, you'd get extreme back-and-forth jumps...

(However, it should not be ontogenetic. Dave Unwin gave a talk 3 days ago mentioning that in pterosaurs the ratio of wing finger length to wingspan and of hindlimb length to wingspan stay the same throughout growth. Isometric growth like in sauropod limbs. Sounds like a great criterion for figuring out which juveniles and which adults belong together!)

<< A growth series. JVP 2 or 3 years ago, bone histology: there is such a thing as immature *Rhamphorhynchus* in the fossil record.

True, but Bennett is also a fanatic lumper. >>

The number of *R.* species is emphatically not what I was talking about. I was talking about a bone histology paper. How many species the various *R.* specimens belong to doesn't matter for my argument -- what matters is that some of them are immature (juvenile growth rates at death, lack of external fundamental system).

Why is an X-ray very bad in your experience. Thousands of hospitals
use them every day!

Yeah, but you aren't encased in stone in a hospital. Figure 6 of the description of *Pterorhynchus* has a VERY bad resolution. You can see the red, green and blue pixels (on a B&W pic!) which are at least as big as the pixels on this flatscreen.

<<I've recently been shown photos of the skull of Pterorhynchus and
the characters it shows are strongly pterodactyloid. >>

How so? And which 'pterodactyloid'? There are dozens to choose from.
But then, there's that gigantic tail and short metacarpal.

Two plesiomorphies -- which preclude it from actually being a pterodactyloid, and indeed, it doesn't come out within that clade.

> `--+--+--Pterodactylus
>    |  `--+--Lonchodectes
>    |     `--+--Cycnorhamphidae
>    |        `--Ctenochasmatidae

Unwin has a nearly complete lonchodectid specimen up his sleeve. He showed photos at a talk of his. Lonchodectidae now comes out close to *Tapejara*, *Tupuxuara* and Azhdarchidae.

> "Daitingopterus" 

Oho!

Because pterosaurs inside of eggs have proportions and tooth counts
very close to those of adults, one can employ embryos, juveniles and
adults in the same cladogram.


For these characters, yes. Some of them anyway.

If you want to determine juvenile vs. adult you can look to bone fusion,
texture, etc. for clues. Shape is no longer on the list.

This is not true for things like relative head size and relative eye size -- there we have skull shape, BTW.

<< [...] it has over 120 characters and
over 80 taxa, far more than Unwin's or Kellner's analyses.>>

While this is clearly good, I can't resist submitting the cheap comment that it can't be enough. Not even twice as many characters as taxa... :-/

Above I have listed many 'bumps' in your 'blend'. If your results
produce a hundredth of 34,000+, then your results produced 340 trees.
Right? That's still too many, and more than one, which is the number
you are competing against.

This is a misunderstanding of cladistics. There are, for starters, lots of ways to get a single MPT that is false. There are even lots of ways to get a single _and well supported_ MPT that is still way off.

<<but P. micronyx may be a juvenile Diopecephalus or Gnathosaurus and
P. elegans is a juvenile Ctenochasma. >>

Wow. You're all over the place here. Pick one. Also, which P.
micronyx. We've gotten to the point now that specimen numbers are
required because one P. micronyx may not be the same as another P.
micronyx.

I think I can safely agree with this (I mean, without knowing any "*P. micronyx*" specimen "in person"), but all of these specimens are so clearly babies that including any of them in an analysis that may contain ontogeny- and/or size-related characters is not a good idea in the first place.

<< Sure, the results may make sense, in fact, they are very likely to
make sense one way or another, but they may not reflect the true
evolutionary relationships.  >>

Once again, the weasel-word 'may' demonstrates that this is an
opinion based on a priori assumption.

Untrue. Scenarios -- results that "make sense" -- can happen to "reflect the true evolutionary relationships", but they demonstratedly don't need to.

I have argued here that your tree has some unwarranted reversals
and other unusual morphological alignments.

Why "unwarranted" and "unusual"? Evolution happens. Unless you code a character as irreversible, PAUP* will not treat it as such. A reversal is one step.

It also demonstrates that we already know, essentially, the
entire tree of the Pterosauria. Future
discoveries can simply be plugged in.

It does not even begin to demonstrate this. The addition of new taxa with unexpected character combinations can always change the results. Your tree merely predicts that such taxa will never be found -- but every cladogram does that. "Predictions are very difficult, especially about the future."