Re: was ... Jaime. Now: M. Baeker has a problem with phylogenetic analysis

[David Marjanovic <david.marjanovic@gmx.at>]

----- Original Message -----
From: "David Peters" <davidrpeters@earthlink.net>
Sent: Monday, June 19, 2006 7:03 PM

> If you find it too difficult to come up with an adult/juvenile pterosaur
> pairing, go to the literature and see what forms Wellnhofer and Bennett
> have promoted as adults and parents of the same genotype. Then expand
> the universe of taxa to include all pterosaurs and you'll find that
> rather than matching parents, the so-called 'juveniles' actually form
> series that fill phylogenetic gaps between taxa. It's not an isolated
> incident. Happens about six times.

I still don't understand why you act as if PAUP* could tell you if an OTU is 
adult or juvenile. It can't.

As you are fond of saying, PAUP* nests everything. If you would put 
"Archaeoraptor" into a theropod matrix, it _would_ come out somewhere. I 
can't predict if it would find itself in a polytomy with *Microraptor* and 
Euornithes, but I bet it would not turn the entire tree into phylogenetic 
grass. Probably it would come out close to *Yanornis* with the tail as a 
reversal -- would be great if someone with a bit of time on their hands 
could test this.

> > Fledglings usually *do not* look like parents.
>
> In pterosaurs they do, as embryos tell us.

The features necessary for flight indeed do: the ratio of wing finger length 
to wingspan does not change in ontogeny, the ratio of hindlimb length to 
wingspan does not change in ontogeny, and the pterosaurs in the eggs have 
wing loadings that fall into the range of adult pterosaurs. (Source for all: 
Unwin's talks last week.)

But that's not all. It is completely inevitable that small pterosaurs have 
relatively larger eyes and brains than big ones; this is a relationship that 
holds throughout at least the vertebrates (where eyes are present) and can 
very easily be explained by the way vertebrate ontogeny works. The 
ossification of tarsals and small phalanges is delayed compared to the 
flight apparatus. Then there's bone histology -- only adults have an 
external fundamental system.

> Chiappe and Co. told us one embryo was Pterodaustro because
> it was found with and looked like adult Pterodaustro.

Not precisely, though.

> More details revealed it to be an anurognathid.

On the third day of the congress we had excursions to the natural history 
museum and the geology museum. There I saw most of the feathered dinosaur 
specimens I knew from the literature (and then some -- there are apparently 
unpublished specimens of *Sinornithosaurus* on display). It turns out that 
despite the good photos in National Geographic I had got some 3D 
relationships wrong. Do you remember how I lamented that the middle portion 
of the tail feathers of *Protarchaeopteryx* had not been prepared and was 
still lying under rock? Nonsense. The white patch is not the overlying but 
the underlying layer. The tip of the bony tail of *P.* is, if anywhere, on 
the counterslab, which is not on display (let's hope it still exists). Do 
you remember how I lamented that the feathers on the snout tip of NGMC 91 
had not been prepared free? Again the feathers are a layer above, not below, 
the rest of the slab; if there were feathers on the tip of the snout, they 
are on the counterslab, to which the same comments apply. The same holds for 
the snout of the holotype of *Sinosauropteryx*.

Lesson: To claim anything based on a photo requires even more care than I 
had thought.

Take-home message: I don't see the slightest reason to believe your tracing 
that "revealed it to be an anurognathid".

> >> Just include all pterosaurs and let PAUP find the relationships.
> >
> > It can't if you plug in units (specimen or species) *that differ in
> > more than just phylogeny*.
>
> Example please.

Well, it _can_, but only by chance, that is, if you happen not to have too 
many size- or ontogeny-dependent characters in your matrix.

Example: *Iberomesornis*. Sure you remember all the cladograms from the 
1990s that had

--+--*Iberomesornis*
 `--+--Enantiornithes
    `--Euornithes (not named yet)

based on several convincing features, like the presence of only 5 sacral 
vertebrae in *I.* (like in Archie) vs 8 in Enantiornithes and 9 upwards in 
Euornithes, or the lack of fusion in its metatarsi. Then came Sereno and 
showed (long paper, badly accessible journal) the specimen was a juvenile 
enantiornithean. Oopsie. Guess why I will never put *Liaoxiornis*, or 
"*Cathayornis caudatus*", into my bird matrix, even after having seen 2 or 3 
specimens in museums.

> Phylogeny is easier than you think.

Phylogenetics is more complicated than you think.