RE: Wukongopterus and Darwinopterus

["Unwin, David M. (Dr.)" <dmu1@leicester.ac.uk>]

Darwinopterus – a true test of pterosaur phylogeny?

In response to comments by A. Kellner and D. Peters:


AK (Nov 17th 2009) asked:
“By the way, what are the differences between Changchengopterus and 
Darwinopterus? I hoped to learn more from it, but Lü et al. did not compare 
both taxa in their paper.”

The diagnosis given by Lü et al. distinguished Darwinopterus from all other 
pterosaurs described up to that point which (obviously) included 
Changchengopterus (variously spelled “Changchengopterus” and “Changhengopterus” 
by Wang et al. 2009). Metric data suggests that the holotype and only known 
specimen of Changchengopterus is distinct from all described specimens of 
Darwinopterus. For example, wing-phalanx 1 is longer than the tibia in 
Changchengopterus, but shorter in Darwinopterus – this can easily be 
established by comparing the metric data published by Lü 2009 and Lü et al. 
2009. 


DP (Nov 16th 2009) asked:
“Shall we keep Wukungopterus and Darwinopterus separate for awhile? Or, which 
came first according to the rule book?”

I have now scored “Wukongopterus” for both the Lü et al. 2009 and Wang et al. 
2009 cladistic data sets and, in both cases, where comparable, the character 
state codings are identical to those of Darwinopterus. Morphometrically 
speaking these taxa also appear to be identical. Evidently, therefore, there is 
a good case for synonymy (with Darwinopterus taking precedence as it was 
published well over a month before “Wukongopterus”). It remains to be seen 
whether the small differences in the dentition and jaw shape that I noted 
previously (see DML, Nov 15th 2009) are sufficient to justify the retention of 
“Wukongopterus”. They might, for example, be an artefact of compaction damage. 


A. Kellner (Nov 17th 2009) commented:
“While describing Wukongopterus we run [sic] into two problems: first, we did 
not have a complete skull, showing the confluent external naris and antorbital 
fenestra.”

Interesting point. I presume from this that Wang et al. now assume that 
“Wukongopterus”, like Darwinopterus, had a confluent nasoantorbital opening 
(this supposition is supported by an illustration accompanying various news 
reports on “Wukongopterus” which shows it with a confluent nasoantorbital 
opening). So, having recently managed to replicate Wang et al’s 2009 
phylogenetic analysis, I reran it scoring this character as present for 
“Wukongopterus”. It had no impact on the analysis, which still recovered 
“Wukongopterus” in a trichotomy with Dorygnathus and a bunch of other basal 
pterosaurs, including Dimorphodon and a series of Triassic taxa that, according 
to this analysis (and as in Wang et al. 2009) are more derived than 
“Wukongopterus”. 

As mentioned above, I also scored Darwinopterus for the Wang et al. 2009 data 
set. It is isometric with “Wukongopterus” which can thus be safely deleted from 
the analysis. According to Wang et al. 2009 “Wukongopterus” has missing data 
for 34 out of the 89 characters used in their analysis (and is thus 62% 
complete). Darwinopterus adds data for 28 of these characters and is thus 93% 
complete. So, I reran the Wang et al. 2009 data set using this near complete 
sequence for Darwinopterus/"Wukongopterus". To be fair, the taxa Scleromochlus, 
Angustinaripterus, Cacibupteryx and Harpactognathus were excluded following 
Wang et al. 2009 (actually, I couldn’t follow the logic of this [the 
exclusions, not the ‘following’] as these exclusions appear completely 
arbitrary: for example, Cacibupteryx was scored for 36 characters while 
“Ornithocheirus” compressirostris* was scored for only 22, but the latter was 
included, while the former was excluded). Anyway, one might predict, as 
Kellner’s comment implies, that the better known Darwinopterus (compared to 
“Wukongopterus”) should now fall out as a sister group to pterodactyloids. Not 
so. This more complete data set generated a lot more trees (109980), and the 
strict concensus was very poorly resolved, failing, for example, to recover a 
monophyletic Pterodactyloidea. Darwinopterus formed a polytomy with many other 
basal forms and several pterodactyloid clades. The 50% Majority Rule tree did 
recover a monophyletic Pterodactyoidea, but “Ornithocheirus” compressirostris 
was located basal to Dimorphodon. 

Preliminary conclusion. The unusual character combination exhibited by 
Darwinopterus can be used to test the behaviour of phylogenetic analyses of 
pterosaurs. The Wang et al. 2009 phylogenetic analysis (an iteration of 
previous Kellner data sets including: Kellner, 1996, 2003, 2004, Wang et al 
2005, 2008, generates poorly resolved trees in which 
Darwinopterus/“Wukongopterus” occupy a highly anomalous position and thus fail 
the test. Anyone who inspected the poorly resolved cladogram presented by Zhou, 
Zhou and Schoch at SVP in Bristol this year, which was also based on Kellner’s 
older data set and incorporated yet another example of Darwinopterus (which 
also appears in an odd location), will have already seen the same result. 

* Yes, I know this is a lonchodectid and that it does not exhibit any true 
apomorphies of ornithocheirids, or even ornithocheiriods – don’t worry, this is 
being dealt with elsewhere