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

Re: David Unwin's The Pterosaurs From Deep Time - review

> Enter David Unwin?s The Pterosaurs From Deep Time (2005), self described
> as ?the first complete portrait of the legendary dragons of deep
> time...? [...] Claiming to be the latest word in
> pterosaurs, Unwin states, ?the true nature of these Mesozoic dragons has
> proven elusive ? until now? and ?this book puts that [puzzle] picture on
> display for the first time... ? and ?the pages that follow contain the
> first comprehensive account of our new understanding of how pterosaurs
> were constructed and how they lived their lives..?

Since when do authors write their own blurbs? I'm sure the publisher did 

> Given that the pen is mightier than the sword, Unwin had to have been
> eager to lance, slash and dismember all opposing and weaker viewpoints
> with unprecedented and insightful cladistic analyzes [...]

You're expecting too much of the _secondary_ literature.

> written in a vernacular that might appeal to paleo beginners with juicy
> fiction opening each chapter

This shows that it _is_ secondary rather than primary literature. DA is, in 
part, primary literature. Wellnhofer's encyclopedia was not.

> The family tree of the Pterosauria, as depicted by Unwin (fig. 4.5), is
> a real tree with 17 select skulls figured as metaphorical fruits at the
> ends of branches. The pterodacs are at the top and the rhamporhynchs are
> at the bottom. Of course, this sort of graphic hearkens back to the
> pre-cladistic trees of the 20s through 70s ? quaint ? but ashamedly not
> keeping up with the freshman basics of paleontology.

As long as the topology is right, there is nothing to say against this.

> Here Pteranodon inexplicably
> joins the toothy broad-snouted pterosaurs rather than the other
> toothless sword snouts.

"Inexplicably"? Hasn't Unwin published at least one cladistic analysis in 
which you can read why it "joins the toothy broad-snouted pterosaurs"?

> Later (but on the same subject), rather than employing a real pterosaur
> sister/ancestor taxon

Sister or ancestor???

> as determined by PAUP (and PAUP can nest everything!),

Well, it's a computer program. It cannot _not nest_ anything.

> I suppose it is unusual to be singularly erased from a bibliography and
> yet thanked in the acknowledgements (I actually had nothing to do with
> this book). Not one of my abstracts or peer reviewed published studies
> are to be found here. I can understand Unwin?s reason for doing so. Most
> of my papers cast doubts on his hypotheses and observations or they
> provide alternative views. If my theories and observations were invalid
> or weak, they should have been easily dispatched and this was the forum
> to do it in.

Why drag an academic debate into what seems to be intended as a 
coffee-table book? (I haven't seen the book, I hasten to add.)

> (remember, you get paid to write a book)

In theory at least...

> Ironically. on the
> same page as the ?trimmed? Zittel wing, Unwin shows us the Vienna
> specimen of Pterodactylus in which there is no doubt that the wing
> membrane extends only between the elbow and wing finger  with a small
> fuselage fillet (nor a part of the wing), extending to mid thigh, on
> both wings.

Where is the difference between a "fuselage fillet" and a (moderately) 
broad wing? Especially considering the fact that the "actinofibrils" have 
now turned out to be muscle fibers, which means that the wings (like bat 
wings) did _not_ have a fixed shape.

> I was hoping to see what happens, step-by-step, to a broad chord wing
> membrane when it is folded after landing. Does it droop? Or what? And
> why not? Unwin left it a mystery.

I guess it contracted to become very narrow.

> ignoring an earlier paper that showed that the pteroid and its more
> distal counterpart are former central[ia] that have migrated to the
> leading edge of the wing and act as passive restraints to overextension
> of the wing finger due to drag while in flight.

Naïve question.. why can't it be the ossified insertion of a tendon? This 
would spare us the whole migration hypothesis (and the assumption that a 
centrale, let alone two, was still present). What is that "more distal 
counterpart", BTW???

> I was hoping to see the famous ?flat-footed? Dimorphodon weintraubi set
> in various phases of the walk cycle and at no time extending the toes.

Well, a few degrees of hyperextension are enough for plantigrade walking.

> The human hand is likewise incapable of extension -- except when
> pressure is applied -- and the same could apply to pterosaur feet
> adding a bit of spring to each liftoff.

Not everything that doesn't work is a metaphor... Of course our metacarpals 
have articular ends that allow hyperextension. Our muscles are just too 
long to do all of that hyperextension.

> He also did not include the one
> undisputed bipedal digitgrade anurognathid track,

Excuse me... how do "undisputed" and "anurognathid track" go in the same 

> the Rhamphorhynchus ?growth? series of skulls lifted from Wellnhofer
> (which has been shown to be a phylogenetic series leading from
> Campylognathoides when tested in PAUP)

You are grossly misjudging the capabilities of cladistics. If you put a 
_real_ ontogenetic series into PAUP*, how on the planet is it supposed to 
find out that these OTUs are in fact one??? It is a computer program. It 
cannot help but assume that all OTUs are equal. Even if the autapomorphies 
of the single taxon to which the growth series belongs are present in the 
data matrix, ontogeny- and size-related characters -- if present -- will 
certainly outweigh them. Isn't there an SVP meeting abstract from this year 
that says actual babies entered in a cladistic analysis come out as more 
basal than their parents?

> Those tiny 'babies' may actually be hummingbird and bat-sized adults.

...which look a lot more similar to baby hummingbirds and baby bats than to 
the respective adults, down to (external) bone histology.

> Speaking of eggs, Unwin reports that the
> pterosaur eggs were thin and soft-shelled, like those of lizards and
> turtles, rather than birds and alligators. Hmm, maybe that clue can help
> us in the search for pterosaur origins.

It is evidence for non-archosaurian pterosaurs... or for a double origin of 
hard eggshell in archosaurs. How old is the oldest known crocodile 
eggshell? I hope that's not *Krokolithus* from the mid-Eocene Geiseltal?

> Perhaps this explains why 3 for 3 pterosaur eggs are full
> term and so few archosaur eggs are as well formed.

Good luck trying to do statistics with a sample of 3.

> With all these mysteries still left unsolved, as often noted by Unwin
> himself, why would he claim that all these mysteries are settled now?

That's what the blurb says, right?

> Unwin describes the primitive uropatagium as binding the hind legs
> together, as if one were wearing a raincoat clamped to one?s pants legs
> ? and that this medial membrane connected to the lateral toes (!) How
> this can happen mechanically is not shown.

This is easily imaginable if the toe bends around, as preserved in *Sordes* 
(shown in Wellnhofer's encyclopedia). Why this should have evolved is 
another question.

> The analog of living bow-legged lizards capable of rapid bipedal
> progression still has not sunk into Unwin?s academic mind. Perhaps it
> will when he learns that pterosaurs are indeed lizards. In fact you can
> throw away all the closest sister taxa and include only archosaurs (as
> many as you like), a pterosaur and either an iguana or monitor lizard
> and the pterosaur will nest with the lizard.

I still haven't answered to your offlist message from July or August about 
your diapsid analysis, and I fear I won't find the time for that in another 
few months. Sorry! However, this is the place to mention that you shouldn't 
include phalangeal proportions in a phylogenetic analysis. Why? Because in 
birds they seem to agree _exactly_ to the amount of time the species spends 
in the trees vs on the ground. No phylogenetic signal seems to be present 
in these characters; they seem to be extremely volatile. It goes without 
saying that using a character without phylogenetic signal in a cladistic 
analysis is a total waste of time and computing capacity. An example would 
be to use color in an analysis of Amniota -- if you're out of luck (such as 
having too few informative characters in the matrix), the Eurasian grass 
frogs, *Rana (Rana)*, will clade with the green anoles and *Lacerta 
viridis* and selected birds, to the exclusion of the Eurasian brown = water 
frogs, *Rana (Pelophylax)*.

But don't take my word for the presence or absence of a phylogenetic signal 
in phalangeal proportions! Instead, test that. Mesquite (at least) has that 
test inbuilt. You enter a tree that you consider correct or at least 
halfway plausible* and tell Mesquite which OTU has which state of the 
character you want to test. The program traces the evolution of the 
character in question on that tree, then it randomizes the tree a large 
number of times and shows on how many trees the character has a more 
respectively less parsimonious distribution than on the input tree. If the 
character needs fewer steps in 95 % of the random trees than in the input 
tree, you know with 95 % probability that there is no phylogenetic signal 
in that character, and you can throw it out of your matrix.

* Must of course be a tree that wasn't made using that character. That 
would result in circular logic.

Lust, ein paar Euro nebenbei zu verdienen? Ohne Kosten, ohne Risiko!
Satte Provisionen für GMX Partner: http://www.gmx.net/de/go/partner