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

Re: Avgodectes, now: eyeballing



On the webpage: www.pterosaurinfo.com/pterodac_clade.html you
can see that No. 9 and AMNH 1942 have identical torso sizes.

On my screen, No. 9 has a torso length (straight line glenoid to acetabulum) of 7 mm, while AMNH 1942 has one of 10.5 mm (to my surprise there is a half-millimeter ruler in this room), so I wouldn't say "identical"...


IMHO they are juveniles of different species, AMNH 1942 being more closely related to *Pterodaustro* than to No. 9, so I guess it's a ctenochasmatid. As I don't have any pterosaur literature here, this is quite a humble opinion, but perhaps some of the lurking experts can fill us in here... (insert smiley: scroll down to first smiley on http://www.molecularbiology.at/index.php?name=PNphpBB2&file=viewtopic&t=30&sid=419b12dcd58214f4c8aae08baf8d2fcc - - more than halfway down the page)

Are they the same ontogenetic age?

They could be, by coincidence. But of course in the absence of a good photo and/or description I can't hazard a guess on any details.


AMNH 1942 has a much longer neck and rostrum. Is it an adult?

Probably not. But I'm sure there are several published opinions on this.

Similarly on the webpage:
www.pterosaurinfo.com/scapho_clade.html  you'll note that tiny No. 9
(the 'model' baby pterosaur) follows Scaphognathus cladistically

Sorry, but for reasons outlined earlier I can't accept this as long as you don't demonstrate that either both No. 9 and the used *Scaphognathus* specimens are adult or that you don't have ontogeny-related characters in that matrix.


and overall resembles one more than it's presumed parent among
pterodactylids.

Could simply be a case of ontogeny recapitulating phylogeny. But then you seem to be alone so far in thinking that *Scaphognathus* is close to (any part of) Pterodactyloidea, so I'll go with published analyses for now...


Again, note that its cladistic successor, No. 12, has
an identical torso size.

Could it be a ctenochasmatid younger than AMNH 1942?

Cladistic analysis takes into account everything from
palate element ratios to phalangeal proprortions.

Many, if not most, of these ratios are ontogeny-related.

DM:
secondly it's potentially misleading to include species known only from
juveniles because they will likely appear in a too basal position, for
the same reason as above. In
short, do include all tiny pterosaurs in your matrix -- after you've
shown that they are really adult.

Statement falsified by the placement of baby Pterodaustro next to
adult Pterodaustro.

Statement falsified for *Pterodaustro guinazui*, but not for any other pterosaur, see my last post.


DP:
        Others need me to visit every taxon (this is common
         practice in matrices of 7 to 25
        suprageneric taxa, hasn't been done in matrices over that)

DM:
Of course it has been done. I happen to know the example of HP Oliver
Rauhut's really big* theropod matrix in his dissertation.

You need to stick with the subject. The subject is Diapsids.

The subject is most emphatically not "diapsids", it is "cladistic analysis". Whether it's necessary to visit fossils personally doesn't depend on the taxon, it depends on their so far published descriptions and illustrations.** If you want your analysis to be better than all published pterosaur analyses, why don't you take the opportunity* to be _much_ better?


* Connected to a lot of work, and time, and money. I can't help it. As I wrote offlist, a phylogenetic analysis of that scope is about the size of a Master's thesis.
** I'm currently in the process of measuring, among other things, the skull lengths of as many dinosaurs as possible, skull length being defined as "horizontal distance from the tip of the premaxillae to the caudal end of the skull table". This is easiest to measure in reconstructions that show the skull in dorsal view. For most dinosaurs with known skulls, a photo or drawing in dorsal view does not exist. Many values in my database are therefore likely to be wrong by up to perhaps 10 centimeters. -- In your reconstruction of *Huehuecuetzpalli* http://www.pterosaurinfo.com/huehuecuetzpalli_recon.html, there is something that looks like a complete lower temporal bar. Is the fossil preserved in lateral view? If so, you might need a good 3D photo or drawing to tell that that thing is not a temporal bar, it is the rear part of the palate (the pterygoid and the pterygoid process of the quadrate), while the lower temporal bar is completely missing like in AFAIK all squamates. Most reconstructions of fossils are line drawings that give completely false ideas about the 3D shape.


DM:
Adding taxa can be botched up.


This is for the referees or the next generation of grad students to
_show_ that the additional taxa are botched up. To assume or presume
that there is a botch up shows prejudice, not science.

Need I repeat? I wrote "can be botched up", not "is automatically botched up when you do it". I was just trying to send a message of caution: don't think that adding taxa will automatically solve problems.


Er, BTW... you are the author, so it's actually your job to show you haven't included juveniles or miscoded specimens from the literature, not that of the next generation.

DM:
See above for the need to be careful with assessing maturity; see my
misery for the great benefits of personally examining fossils that are
poorly described and illustrated, not to mention the fact (mentioned
several times on this list, years ago) that published data matrices
often contain typos and must therefore not be simply copied; see
theoretical studies and my experience for the need to add characters
when you add taxa (...and even when you don't add taxa).


I note that others have copied data matrices, character lists, etc. in
prior work, so your finger pointing should be at them.

I didn't say you copied from other data matrices. I said you should really personally examine the more poorly described/illustrated of your OTUs. Others have made and published mistakes because they failed to inspect the fossils.


DP:

        3. Evidence of what basal plantigrade pteros did with that dang
        big lateral toe while walking (see keyword: 'sauria aberrante' for
        the alt. explanation).

DM:

It looks (er... literally) like it was involved in the "hindwing". But I
wouldn't bet money on that.

Keyword: walking. And no one, even Unwin and Bakhurina 1994, has
shown in detail (not cartoon) that the toe was involved with the hind
wing, or rather the uropatagium

I know. That's why I didn't try to express more certainty. Because of its shape the toe looks as if it was involved in the (cr)uropatagium, and this can be interpreted into one or two fossils; your explanation is equally convincing, plus it does a slightly better job of explaining how the pterodactyloids got away with losing that toe. So I guess I'll just wait for more papers.


DM:
Several published cladistic analyses, plus reportedly a big unpublished
diapsid analysis (lots of taxa, including "prolacertiforms") by Senter.

Pterosaurs were not taxa in Senter's analysis.

AFAIK he has at least a pterosaur OTU. If correctly coded, this artificial taxon should nest in the correct place. But I only know this 2nd- or 3rd-hand. Who knows what Senter has changed since his SVP talk/poster/whatever a year or two ago.


Otherwise, in the
several published cladistic analyses you refer to (all small, by the
way) lizards were not included so pterosaurs nested with dinos by
default.

Sorry, that's true, I can't think of one with included prolacertiforms...

DM:
Sorry for the impolite question, I'm too tired to think in terms of
social intelligence -- have you found out how to order multistate
characters in a NEXUS file?

Of course, but that needs to be done with assumptions. I prefer to
work without assumptions.

It is a _big_ assumption to assume things like "big (0), medium (1), small (2)" or "numerous (0), medium (1), sparse (2)" should _not_ be ordered. Of course there are borderline cases; I think the number of sacral vertebrae within hadrosaurs or within Neornithes should perhaps not be ordered, but by default sacra grow by the addition of already present dorsal and/or caudal vertebrae, so by default the number of sacral vertebrae should definitely be ordered, for example.


DM:
When I saw your pterosaur matrix -- long ago! -- you had several
characters that were morphological series, yet all your 183 or so
characters were unordered
and thus _had to_ produce _bogus_ results.

why bogus?

Imagine this one-character matrix, and imagine that taxon A is the outgroup:

A 0
B 1
C 1
D 1
E 2
F 2
G 2

The strict consensus of all MPTs will look like this when the character is unordered:

--+--A
 `--+--+--B
    |  |--C
    |  `--D
    `--+--E
       |--F
       `--G

because state 1 is a synapomorphy of BCD while state 2 is a synapomorphy of EFG. Now suppose we order the character. The resulting strict consensus will be quite different:

--+--A
  `--+--B
    |--C
    |--D
    `--+--E
       |--F
       `--G

Now suddenly state 1 is a synapomorphy of BCDEFG; state 2 is still a synapomorphy of EFG, but BCD don't share _any_ character state to the exclusion of E, F or G.

Suppose we had a theropod matrix with too few characters, one of which would be the unordered number of sacrals. With enough bad luck (which could be helped by the presence of few enough other characters), we could get a clade that contained all theropods with 5 sacrals but not any with more than 5. I'm sure you see this would be absurd.

I remember Mickey M complaining onlist about a sauropod analysis that did not order the number of sacrals because of a gross misinterpretation of embryology. Fortunately it still had enough characters not to produce as absurd results as the artificial example above.

Name three characters that benefit from ordering.

Any multistate character that forms an obvious morphological series, in the absence of _good_ evidence that you don't need to go through state 1 in order to pass from 0 to 2 or from 2 to 0.


DM:
your interpretation of the fossil doesn't look convincing...


Again: Either one supports cladistic analysis, or one supports
'eyeballing'

http://www.pterosaurinfo.com/avgodectes_fig2.html is not the result of a cladistic analysis. It is the result of your interpretation of a photo -- should I say "the result of eyeballing"?


If you're going to code OTUs in a cladistic analysis after such drawings, you better devote 20 pages of the paper to explaining how your method works and _that_ it works (means, demonstrate that it doesn't only find faces and pyramids on Mars). A good test would be to get your hands on a fossil you have interpreted and look if you can find and prepare, say, a claw on the end of the wing finger or a continuation of the tail of a pterodactyloid. If you can do that (which could be a separate publication in your curriculum vitae), your phylogenetic analysis of Pterosauria would certainly be accepted for publication, and rightly so.