Pterosaur phylogeny

[h0662eka@rz.hu-berlin.de]

On November 23 David Peters sent:

'Dear David,
Alexander Kellner also presented a pterosaur phylogeny in the new
Geological Society Buffetaut/Mazin volume on Pterosaurs. Could you
please discuss the differences and similarities between your two
cladistic views?'


The Kellner and Unwin phylogenetic analyses of pterosaurs

In the Discussion section of my analysis of pterosaur phylogeny (Unwin 2003, p. 
172-174) I explored, at some length, the differences and similarities between 
the 
results of my work and that of others including the study which largely 
underlies 
Kellner's (2003) analysis. Now that I have a copy of Kellner's paper (and a 
copy 
of the pterosaur book) I have also begun analysing his 2003 data set, but have 
only dealt with basal pterosaurs so far. 

In a nut shell, the analyses by Kellner (2003) and Unwin (2003) share a lot of 
common structure. Principal differences include the placement by Kellner of 
anurognathids as a sister group to all other pterosaurs. This is largely 
supported, in Kellner's analysis, by two characters (anterior position of nasal 
opening, relatively short wing-metacarpal) that are inflated to form five 
characters in the data set. Taking this into account and some other errors in 
Kellner's codings (e.g. character 8, the nares and antorbital opening are 
separate in Preondactylus, not confluent) then rerunning a modified Kellner 
data 
set yields a concensus tree that, in terms of basal pterosaurs, is highly 
congruent with my preferred tree (Unwin 2003, fig. 7a). The key exception being 
that in Kellner's tree the anurognathids remain as the basal most taxon. Chris 
Bennett has some further data that addresses this issue (presented at the 
Eudimorphodon meeting in Bergamo) and I will present more evidence that 
supports 
the derived position of Anurognathidae found in my 2003 tree in the proceedings 
of the Eudimorphodon meeting - which I hope will be available shortly. 

With regard to the pterodactyloids, while Kellner and Unwin's analyses show a 
great deal of congruence Kellner's 'Archaeopterodactyloidea' is a hodge-podge 
of 
basal pterodactyloids, including forms that, in my opinion, belong within 
Dsungaripteroidea. Once these are removed the resulting clade is practically 
isomeric with Ctenochasmatoidea except that, curiously, Kellner did not include 
Gnathosaurus, Plataleorhynchus, Huanhepterus or Cearadactylus in his study. 
These 
taxa form the Gnathosaurinae of Unwin 2002, but I am rather baffled as to why 
this entire clade was omitted from Kellner's study, which otherwise has good 
taxonomic coverage. 

One serious piece of weirdness in Kellner's study concerns 'Ornithocheirus 
compressirostris'. This pterosaur from the British Chalk has nothing to do with 
ornithocheirids at all, and belongs in the Lonchodectidae (see Hooley 1913, 
Unwin 
2001) a clade about which I will have much more to say shortly, following my 
month long study of Chalk pterosaurs in the UK in 2003 which turned up a very 
important associated skeleton of a lonchodectid! Amazingly, in Kellner's study 
Ornithocheirus (= Lonchodectes) compressirostris is paired with Anhangueridae, 
and nested deep within his Pteranodontoidea (= Ornithocheiroidea) a group 
supported by oodles of characters (including the classic twisted deltopectoral 
crest of the humerus), none of which are found in lonchodectids. In my 2003 
study 
Lonchodectidae ended up in Ctenochasmatoidea (Unwin 2003, fig. 7), but I would 
not be at all surprised if ultimately this clade wound up somewhere else such 
as 
a sister group to Azhdarchoidea. 

Kellner (2003, p. 125) also argues that Tapejara and Tupuxuara form a distinct 
clade: Tapejaridae, and claimed, incorrectly, that Unwin and Lü (1997) accepted 
the validity of this clade. In fact, quite the opposite, we specifically stated 
that Tupuxuara was probably more closely related to Azhdarchidae than to 
Tapejara 
- a result that was corroborated by my 2003 study. 

Again, however, I would emphasise that many of these differences are relatively 
minor and often have little impact on our general understanding of the history 
of 
the group. The fact is that much of the basic structure of the pterosaur tree 
(Preondactylus , dimorphodontids and anurognathids at the base of the tree, a 
monophyletic Pterodactyloidea, and well supported Ornithocheiroidea, 
Dsungaripteroidea and Azhdarchoidea) is firmly established and will act as a 
strong constraint on the remaining possibilities. Future work is likely to 
focus 
on specific problems such as the exact relationships of anurognathids to other 
basal forms and the content and relationships of clades such as 
Ctenochasmatoidea. 

Importantly, the general set of relationships that have emerged now provide us 
with a framework within which to explore important palaeobiological questions 
such as the distribution and function of cranial crests which, lets face it, 
are 
a heck of a lot more interesting than fiddling about with cladistic analyses. 

Tally ho,

Dave