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

Re: they keep on growing despite early fusion



">" signs standardized.

David Marjanovic wrote:

> That sounds simple, but you're talking about a PhD thesis in terms of > amount of work. The fact that you expended far less effort than a thesis > does not bode well for the quality of your work.

Wait a minute. You guys love cladograms! If you don't have one already prepared, then you can start with just Iguana, Eudimorphodon, Preondactylus, Marasuchus and Euparkeria. That's not a big deal.

That's rubbish! What good can possibly come out of a cladogram with _five_ taxa? (And you forgot to mention an outgroup -- in the orthodoxy, *Iguana* would suggest itself, but that's precisely what we're trying to test.)


Or use an established data matrix and plug in the excluded taxa.

That's what I'll do. Except that I first need to unite something like five (...or more, if I'm going to do serious work on pterosaur origins...) partially overlapping matrices, and then check every single character for miscodings, indefensible or outright unpublished state separations, ordering and so on.


I did part of the work in uniting two pretty small matrices recently (the result is one of the matrices that I allude to above). That took me alone several weeks full-time, and I'm just the third author. Even though I added taxa (among other things to replace the compound Squamata OTU with real species) and a dozen characters, the finished matrix has only 35 taxa and 116 characters, IIRC.

I also recently checked a published matrix with 19 taxa and 41 characters for mistakes because that matrix had been published in a prestigious journal and resulted in a very strange tree. There were so many mistakes that a 100-page manuscript has come out of it. Together with adding three taxa (two to replace the "all-zero ancestor" and one because it was the most glaring omission), it took something like... I think two months full-time. Two months for an excruciatingly tiny matrix (for comparison, a more recent matrix on the same problem had 102 taxa and 341 characters... it, too, is full of mistakes and very strangely defined character states, and it, too, could do with the addition of yet more taxa, but I digress).

And why insult a work you are unfamiliar with? Ten times more taxa than any prior study, and growing every day, is not "far less effort".

Well, in sum, how many full-time days did you spend on it? When you sent me your character descriptions 2 1/2 years ago, they looked careless.


And who says it takes alot of effort?

I, from experience. Coding a single taxon for 116 characters takes half a workday, assuming that the characters are all in an acceptable state -- finding out whether a character should be ordered can be tricky, and dividing two continuous characters into states in a reproducible way took a day each for the 22 taxa mentioned above (some of them supraspecific OTUs), even though I used a primitive method that just involved measuring all taxa (that is, more, because of the supraspecific ones), plotting the results in Excel, and looking for the biggest gaps in the distribution (so it's dependent on the taxon sampling).


Either you don't like your boat rocked, or you are happy with keeping your various enigma taxa as enigmas. That's fine by me,

It isn't fine by you, and it isn't fine by me either. Also, I'm already in the business of rocking boats, what with having published molecular date estimates for the origin of Lissamphibia that are much younger than all previously published ones, and boasting about how my M. Sc. thesis finds little evidence of Cope's rule in dinosaurs even though it's nowhere near publication. What I don't like is when people act as if a cladistic analysis were child's play. Cladistic analysis consists mostly of potential sources of error, and hunting all of those down takes ages.


but quit beefing and come up with a genus that's demonstrably closer to pterosaurs!

Patience, young padawan. I still haven't read Atanassov's thesis (and of course the two new taxa in it aren't properly published), and the new Triassic pterosaur with the parasagittal hindlimbs hasn't been published yet even informally, so I haven't seen it. Without knowing these three, I am not able to have an informed opinion. Please accept "I don't know yet, and neither do you" as an answer.


> Because this one has Pterosauria as the sister-group to the rest of
> Archosauromorpha (fig. 17 and p. 154).

But also next to Lepidosaurs and far from Euparkeria + all other archosaurs.

What do you mean by "next to"? The pterosaurs are next to the rest of Archosauromorpha in that figure, and Archosauromorpha as a whole, including the pterosaurs, is (by definition) next to Lepidosauromorpha (which still included the younginiforms).


Check the entirety of the literature. For all its faults, this is the one and only time Lepidosaurs have NOT been excluded from consideration. Benton 1985 was a breakthrough and a one-time-shot. Also, perhaps, an unfounded embarrassment, because Benton dropped Lepidosaurs from all subsequent lists. <<<<

Sure, the lepidosaurs are fairly close to the pterosaurs in that matrix, but the archosaurs are _even closer_. While far enough from the orthodoxy (i. e. Pterosauria _within_ Archosauria, or, as Benton called the crown-group, Neoarchosauria), it is still closer to the orthodoxy than to your hypothesis. That is my point.


> Also note that it does not present a cladistic analysis. I don't know if > the
> computing power for a matrix with that many taxa was even available in > 1985.
> Instead, Benton presented subtrees (called "cladograms") and supplied a > list of apomorphies for each node, without trying to find out if his > subtrees together were in fact the most parsimonious tree that explains > the data. There is no data matrix, just those lists!


However done, this is one of the first papers of the cladistic era.

It is one of the papers of the "we know how cladistics works, but we won't do it because we don't have hardware and software for it, and doing it by hand would take years" era. Other examples are Milner (1998, 1993) on lissamphibian origins.


> At the time, it was the most thorough treatment of the question that had > ever
> been made. For today's standards, it's _not even publishable_ in the > form it
> was published. (Even though the data -- the mentions of which character > states are present in which taxa -- must not be ignored in modern > studies.)


True. I also have beefs with those who put today's standards on yesterday's papers -- without demonstrating more parsimonious solutions.

We don't know if Benton's results were the most parsimonious ones that follow from his own data. Finding that out would be a lot of work, _because most taxa are not "coded" for most characters_, and useless, because we can simply do the work anew, with more taxa, more characters, and updated information. For most of his tree, this has been done several times; that's why I'm so confident that, contra Benton 1985, younginiforms are not lepidosauromorphs, but far away from the diapsid crown-group.


> Pterosaurs possess an antorbital fenestra, but Wild (1978: 247) > considered that this may be a convergence.

True. Also in Varanodon,

No, *Varanodon* lacks it. It's described as having one, but that's just damage. This explains why the poor beast is illustrated (Olson 1965) with two lacrimalia (one in front of the fenestra, one behind): the damage removed the entire middle portion of the thin lacrimal, so that nothing is left maxilla and nasal.


So it does happen. Often enough.

I never said it should be coded as Dollo. Of course it shouldn't be. (There are rodents with a very similar opening.)


> "Further, Wild (1978: 246--253) reviewed numerous similarities between > the early pterosaurs and various 'eosuchians' and differences from early > thecodontians."
> Sounds phenetic to me.


Not to me. Completely different lineage--NOT both derived from Youngina, but Wild did not know that back then.

"Phenetic" means "counting similarities vs differences", as opposed to "counting synapomorphies" which is what "cladistic" means.


> "The characters shared with *Youngina*, *Prolacerta* and others are all > primitive to diapsids as a whole, except for the reduced quadratojugal, > the ossified sternum, the 'hooked' 5th metatarsal [which is now > considered an autapomorphy of the diapsid crown-group, now that the > younginiforms have been thrown out of that clade], and the 3-pointed > teeth seen in *Eudimorphodon*."

This is a case of looking for your wallet under the streetlight when you know you lost it in the alley, because the light is better. A priori [always a sin] and pre-cladistically, Wild was looking for something lizardy with a long finger four. You don't find that among small archosaurs. If you do, let me know. This is the BIG sticking point.

Aren't you just weighting this character, like Wild did? Isn't this an argument from personal incredulity? "No, I can't imagine that finger IV grew secondarily, so it cannot have done that"? I'm not saying you're wrong; I'm saying that if you're right, you're right for the wrong reasons.


Understood. And I understand Benton's world view that both archosauromorphs (rhynchosaurs, Trilophosaurus, Prolacerta and Archosauriformes (Proterosuchus to Dinosaurs) and lepidosauriformes are both derived from Youngina and kin.

*_WRONG_*. Benton had Younginiformes (as a whole!!!) and "Lepidosauriformes" derived from a common ancestor, and he had that common ancestor and Archosauromorpha derived from another common ancestor. This common ancestor is *_NOT_* "Youngina and kin".


Why do you keep making such confusions? An ancestor on a cladogram is a node, not a terminal taxon!

If you test that you'll find it no longer true. I know, it's a struggle to find sister taxa to the Pterosauria without Longisquama, Sharovipteryx, Cosesaurus and Langobardisaurus in the mix. I can see why earlier workers, desperate for a solution, plugged pterosaurs into archosauria or archosauromorpha

No. If you plug them into a cladistics program, they will come out _somewhere_, as the sister-group to _something_. This is unavoidable. No matter if it requires the pterosaurs to have 3000 autapomorphies. The program will not give you an error message "sorry, the sister-group of Pterosauria is missing".


[but notice no single genus is EVER closely associated! Even by Hone & Benton who had the last word.].

You are making strange assumptions both on the fossil record and on the way evolution works. Why can't it happen that, at least with our imperfect knowledge of the imperfect fossil record, two large groups can be sister-groups of each other?


But you and I both know this was by default. Nothing in the archosauromorpha looks like or even smells like a pterosaur, and that [paraphrased] according to Bennett 1996 quoted in Hone & Benton 2007.]

That's what I used to think (skeletal pneumaticity excepted), but the unpublished pterosaur with the parasagittal hindlimbs certainly does smell like a dinosauromorph, so we better both shut up till it's published.


> so I conclude the paper I quote above is what you mean and misremember. > It is, after all, the closest the pterosaurs have ever come to the > lepidosaurs -- and to the younginiforms! -- in a published paper.

Good, then I've made my point: Exclusion is bad.

I agree with that point, but I don't see how you've made it here. Please explain.


Try to plug in a lepidosaur.

Into what? They are in Benton's non-analysis, and they will be in my thesis.

Take another look at Huehuecuetzpalli (free pdf if you Google 'basal lizard Reynoso') Happy hunting, David. I am always at your service. DP

I have read that paper and used it to code *H.* for the unification of the two small matrices I mentioned above (as one of the replacements for the Squamata OTU). I didn't compare it to pterosaurs, but I must say it looks like an unspectacular squamate to me, complete with the highly derived squamate skull and various lepidosaur-only features, all of which pterosaurs AFAIK lack. Also keep in mind it's immature (all epiphyses, even those of the suproccipital, are lying around freely, not even sutured to the diaphyses). What is so pterosaurian about it?


Incidentally, *H.*, the other four lepidosaurs, and *Prolacerta* all came out in their orthodox positions. But, again, the matrix wasn't made for this part of the tree, so its taxon and character sampling are inadequate for that.