RE: Pterosaur origins

[Michael Mortimer <mickey_mortimer111@msn.com>]

Dave Unwin wrote-

> The long awaited paper on pterosaur origins by Hone and Benton is now out 
> in the Journal of >Systematic Palaeontology. Hone and Benton reanalyse and 
> reject the conclusions of Bennett 1996 >and Peters 2000, and go with the 
> current orthodoxy that pterosaurs are 'derived archosaurs' by >which I 
> suppose they mean ornithodirans.

Very disappointing paper.

Basically, they say Bennett's rationale for excluding hindlimb characters 
(which ended up dropping support for Ornithodira) was unwarranted, which is 
true enough.  Yet this only takes two paragraphs to show and no more 
interesting tests (excluding hindlimb characters from other archosauromorph 
analyses, excluding other regions' characters, examining the homology of 
ornithodiran hindlimb synapomorphies) are run.

Peters' analyses get more detailed treatment, but no methodical reanalysis.  
They note some basic problems (unjustified exclusion of characters; 
characters coded differently in each analysis; Cosesaurus and Eudimorphodon 
treated as bipeds; outgroup issues; use of Eudimorphodon and Longisquama), 
but there are few critiques of specific character codings.  The exception is 
their two paragraphs on Sharovipteryx, in which they list several characters 
which Peters coded that cannot be coded in the actual specimen.  They then 
note Peters' identified Sharovipteryx arm is actually a series of ribs and 
that "the true arms of Sharovipteryx have now been found buried in the 
matrix (R. Reisz, pers. comm., 2003) and this confirms that Peters? (2000) 
supposed arm was incorrectly identified."  Hilarious.  Still, there are no 
data given showing how correcting these codings affects Sharivopteryx's or 
pterosaurs' position.  An explicit list of incorrectly coded characters 
would have been more useful, and/or a corrected reanalysis of Peters' three 
trees.

The conclusions are particularly unjustified.  "The results of the 
re-analyses of the supermatrix suggest that the Prolacertiformes should be 
considered the sister group to the Archosauria."  And yet, no supermatrix 
was presented, nor was any previously published matrix reran.  "The 
Pterosauria are not closely related to the Prolacertiformes and should 
instead remain among the Archosauria and probably among the derived 
archosaurs."  Yet this has yet to be tested.  Peters' analyses were flawed, 
but that doesn't mean a better analysis including varied prolacertiforms and 
archosaurs would necessarily result in archosaurian pterosaurs.  "Removal of 
hindlimb characters has no significant effect on their position and so 
cannot be used as an argument for convergence and non-homology between the 
dinosauromorphs and pterosaurs."  But, but, it did.  Hone and Benton never 
found that exclusion of hindlimb characters left pterosaurs as archosaurs.

In conclusion, this is a good outline of a paper.  It presents some valid 
issues but never goes further to explicitly detail most of the problems or 
test the hypotheses.

I should also note Renesto and Binelli's (2006) analysis suffers from the 
opposite problem most do.  While most analyses analyzing pterosaurs' 
position exclude prolacertiforms, drepanosaurs and Longisquama (or just have 
Prolacertiformes as an OTU), Renesto and Binelli's exclude proposed 
archosaur relatives of pterosaurs.  The only included archosaurs are 
Ornithosuchus, Archaeopteryx and Sinosauropteryx.  A useful test would 
include Scleromochlus, Lagerpeton, Marasuchus and such as well.  Renesto and 
Binelli's matrix is also quite small, only including one proximal tarsal 
character, for instance.  There has yet to be a published matrix usefully 
testing pterosaurs' phylogenetic position among archosauromorphs.

Mickey Mortimer