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

Re: something's wrong here: Qianosuchus phylogeny

The strict consensus tree in the accompanying pdf file looks 'traditional.' By that I also mean 'bogus' as many of you already know. Let's start the rant right here:

Well, below this line you wrote an indented list that represents the tree in some way. I thought I had the pdf and its 6 supp. inf. files here, but it turns out I had downloaded them in the university and not copied them to here, so I have to guess how your list is supposed to translate into a tree, i.e. which parts are Hennig combs and which parts are polytomies. Can you maybe write the tree again, in the DML standard format as below?

--+--
 `--+--
    `--

Hyperodapedon
Prolacerta
Erythrosuchus
Euparkeria
Proterochampsidae
    Scleromochlus
         Pterosauria
         Lagerpeton
         Marasuchus
         Ornithischia
         Sauropodomorpha
         Herrerasaurus
         Theropoda
    Ornithosuchidae
    Qianosuchus
    Phytosauridae
         Gracilisuchus
    Postosuchus
         Crocodylomorpha
    Stagonolepidae
         Fasolasuchus
              Ticinosuchus
              Prestosuchus
              Saurosuchus


Prolacerta > Proterosuchus > Erythrosuchus > Euparkeria is all just fine.

You should rather use the standard (Newick) format for writing trees as text, (*Prolacerta* (*Proterosuchus* (*Erythrosuchus*, *Euparkeria*))), to make clear that it's a tree (of which the mentioned taxa are endpoints) rather than an ancestor-descendant sequence.

Then the trouble really starts.

Can anyone tell me how you derive Proterochampsidae from Euparkeria?

Not at all. After all, nobody claims to be able to do that. It's a tree, not a ladder.

Then, can anyone tell me how you derive the so-called 'Crurotarsi' from Proterochampidae?

Not at all. After all, nobody claims to be able to do that. It's a tree, not a ladder.

Also arising out of Proterochampsidae: Scleromochlus! What???

Nothing arises out of Proterochampsidae. Like all other OTUs, Proterochampsidae is an endpoint of the tree.

You see, one of the reasons why the adoption of cladistics was such a revolution is that it spelt the end of "ancestor worship". Precladistic paleontologists tried (and in a few cases, like the one of the stratopheneticists, still try) to _see_ ancestor-descendant sequences in the fossil record, and whenever they didn't find an "ancestor", they got rather desperate, wrote things like "probably arose from an as yet unknown thecodont or eosuchian", and drew romerograms* where a stippled lines extend downward from the ichthyosaurs for 100 million years only to curve slightly toward the "labyrinthodonts" and then end in a question mark. Cladistics doesn't do that. Cladistics looks for the closest known relative of everything, _no matter how distant it is_. By explicitly distinguishing plesio- and apomorphies, it led to spelling out the fact that we should not expect an ancestor to possess any autapomorphies; autapomorphies were subsequently identified in lots of famous and less famous "ancestors" (such as *Archaeopteryx*, here on this list, a few years ago). It has also led to a much better idea of how bad the fossil record is. Seeing ancestors everywhere requires the -- unspoken -- assumption that the fossil record is very complete; so complete in fact that we shouldn't expect to find any more side branches, but only to fill in the gaps in the tree -- only to find the intermediates between the intermediates. It should have been obvious, but somehow it wasn't.

If I've overlooked something obvious, please tell me, because I get the distinct but highly improbable impression that you have now, for years, been trying to use PAUP* for "ancestor worship", which does not work.

* Phylogenetic trees with bubbles instead of lines, with the bubble width proportional to the author's vague ideas about the diversity, disparity, and importance of the taxon at any given time.

But wait, it gets worse: Pterosaurs arise from Scleromochlus and its unknown sister taxa!!! 

Nope. The pterosaurs _are_ (part of?) the sister-taxon of *Scleromochlus*.

Suddenly the smallest hands in all creation become the largest!

Suddenly the smallest and the largest hands turn out to share a common ancestor at an unknown distance. That's not terribly surprising.

And that stub of a metatarsal 5 becomes a highly specialized retractable toe??? Bogus.

Indeed. Instead, the common ancestor had a normal 5th toe, *Scleromochlus* on the one hand lost it, and the pterosaurs on the other hand only lost the claw and drastically lengthened the rest.

I know, I know, we're talking about sister taxa here, but you have to go pretty far back in this family tree to get that toe to grow back. 

IIRC, at least one phalanx in it is plesiomorphic for Dinosauria.

Then Lagerpeton arises from the sister to pterosaurs? I don't think so. It's closer to Tropidosuchus and then Chanaresuchus and then Proterochampsa. 

According to your... peculiar matrix.

Out of Lagerpeton arises Marasuchus, 

Nope.

Last but not least, theropods arising out of sauropods and ornithischia? 

Nope.

Shouldn't that be the other way around? 

Nope.

Aren't theropods closer to Marasuchus in every way? 

In every plesiomorphic way, yes.

Sharp teeth, reduction of digits, etc. By the way, toe five makes a return in some of the above. 

Yep, in some sauropods.

No wonder the authors were not able to figure out where Qianosuchus nested. They needed the following taxa: Triassolestes, Turfanosuchus, Pseudhesperosuchus. They need updates to Ticinosuchus. They also need more characters. 

Agreed. Even more taxa would certainly help, too.

aetosaurs, by the way, which are also derived from Ticinosuchus. 

Nope.

So, not sure if that tall slender tail of Qianosuchus is swimming. Especially considering that the transverse processes that would have anchored the tail rectractors are really hard to find.

Retractors? They would have anchored the big femoral retractor (M. caudifemoralis longus) and AFAIK the caudal elevators/depressors.

The world view of traditional workers is that someday some series of taxa will fill in those currently untenable gaps in the evolution of one form into another.

Judging from the above, you seem to think that we already have that whole series, which would mean we can stop digging.

There's a simpler solution. It's called parsimony.

Which won't find us ancestors if they aren't there. It will find us sister-groups, though.

We have enough taxa to make it work so that sister taxa seem to blend into one another morphologically,

Come on. If you really believe the big semiaquatic proterochampsids and the tiny bipedal hopping *Lagerpeton* "blend into one another", or even the clunky plantigrade *Lotosaurus* with its crurotarsan ankle and rauisuchian hips and the gracile dinosauriform *Silesaurus*, look again.

And when all the work is done, you have to step back, take a look at the whole thing and ask yourself: "Does this make sense?" remembering that evolution works in tiny increments. 

And remembering the quality of the fossil record.

Last rant: specimen-based studies: good. Suprageneric-based studies: bad, as shown above.

You're right that using supraspecific OTUs greatly increases the risk of screwing up. It doesn't guarantee screwing up, however. I'm currently working on a paper that (as a side effect) will... probably show this to most people's satisfaction.