Peters comments on the E and P of pterosaurs

[h0662eka@rz.hu-berlin.de]

Here are some brief responses to Peters recent comments on my phylogenetic 
analysis of pterosaur relationships published in the E and P of pterosaurs 
volume 
by Buffetaut and Mazin. 


Peters wrote:
>I think Unwin's choice to lump genera into terminal taxa for his cladogram 
>unfortunately will yield chimaera after chimaera and rob him (and us) of a 
>chance to see the finer changes happening at the entrances of exits of grades 
>and clades. But then again, I'm a splitter. With this system Dave was able to 
>generate a matrix grid that was 94% filled in, which is an admirable goal. 
>Still, I think it's the subtle character by character blending between taxa 
>that makes a cladogram seamless, rather than blocky.'

I would urge those interested in this subject to read the entire paper in which 
you will find that I commented at some length on my reasons for using 
suprageneric OTU's. Yes, ultimately we should attempt to analyse pterosaur 
relationships at the species, or even specimen level. I have done this at the 
generic level (which in many cases is the same as the species level) for 
'rhamphorhynchoids'. The resulting cladogram was the same as that just 
published, 
except with a lot more terminal taxa. This should, I hope, appear shortly in 
the 
Eudimorphodon symposium volume. 


Peters wrote
>I would have liked to see Dave's appendix 2 list follow standard description 
>order (head to tail, forelimb, then hindlimb) rather than jump around the 
>body, 
>apparently ordered as phylogenetic changes (starting with Preondactylus).

Originally there was an additional table that did this. I was made to take it 
out 
for space reasons. 


Peters wrote:
>Some characters like no. 5: 'Humerus, shorter (0), or longer than femur 
>(1) would probably have been better stated as 'shorter, or not shorter' 
>to take in the possibility of a subequal humerus. 

Makes no difference to the results. 


Peters wrote:
>Character coding had a few problems. 

>No. 44. Mandibular rami elevated about symphysis scored for Nyctosaurus and 
>Pteranodontidae, but not for Anurognathidae, which is considerably more 
>elevated.

>No. 60 AOF vs. orbit height scored for Tapejara, Tupuxuara and Azhdarchidae, 
>as 
>expected, but missed Anurognathus (not that it would matter much).

First of all, I believe that what Peters meant to write was 'I would have coded 
the characters slightly differently'. Because someone interprets anatomy 
differently does not mean that there is a 'problem' with the character or with 
someone else's coding of it. Perhaps more importantly, I reject Peters' two 
proposed changes because neither of them are visible in any of the 
anurognathids 
I have been fortunate to examine. Even supposing real evidence can be found for 
these reinterpretations, which I strongly doubt, rescoring the data set has no 
significant impact on my published analysis. 


>I was disappointed to see no palate characters and only one extremity 
>comparison. I would think the various proportions of the pedal and manual 
>phalanges would be a great place to find hidden phylogenetic patterns. 

My analysis uses only a small proportion of possible characters/character 
states. 
As noted in the paper, the characters I used were selected because I had been 
able to check them directly against fossil material for the vast majority 
(practically all, in many cases) of the genera belonging to the OTU's. And 
please 
note, anyone who is carrying out phylogenetic analysis of fossil verts - I 
found 
many many examples where having checked a character against the literature, and 
in particular the figures - it did not (let me repeat that) IT DID NOT 
correspond 
with what was actually present in the fossil material. If you want to have any 
degree of confidence in your data (and consequently any hope of convincing 
anyone 
else of the validity of your results) turn off the computer and GO LOOK AT THE 
FOSSILS - its the only way. 


Peters wrote:
>Dave was quite right in taking to task my earlier placement of Sordes in the 
>Dimorphodontidae. Further work has spun my head around. Yes, Sordes is close 
>>to
Scaphognathus. But it is also equally close to Campylognathoides and 
>Dorygnathus and only a step removed from Eudimorphodon and basal 
>Dimorphodontidae. It really is kind of a common brown sparrow or shrew-type 
>from which so many grander taxa emerged. Funny though, it's closest cousin is 
>the new Pterorhynchus found by S. Czerkas, which is not very common looking.

To quote my brother 'I'm always right, it just takes everyone else a little 
while 
to realise this'. 

Peters wrote:
>I'd like to see someone character code a number of the putative juvenile 
>specimens and see where they pop up on the cladogram, rather than 
>prejudicially 
>ignoring their potential contribution due to their apparent immaturity. 
>Remember, there are bee-sized bats and birds in this world. It is an apriori 
>assumption not to consider the wee ones. At worst we could find that they lump 
>in with some putative adult. At best we might find a clade of micro-pteros. 
>Size shrinks and a sprinkling of neotony could easily explain the dimunition 
>of 
>'pterodactyloid' caudals, among other characters.

This is a contentious area, and I'm by far not the best qualified to comment on 
it. However, following the excellent work by Wellnhofer and more recently by 
Bennett on pterosaur ontogeny I have looked very carefully at 'small' 
individuals 
and considered them when coding at the species level. In fact, it does not make 
a 
lot of difference to the phylogenetic analysis, because juveniles still tend to 
fall out in close relationships with their elders. Consequently, although I 
believe this is an issue with regard to pterosaurs I really doubt it has any 
serious implications for the general structure of my preferred cladogram. 

So, to finish with something controversial. I bet that in 21 years time, when I 
retire, that (excluding new taxa found between now and then) the preferred 
pterosaur cladograms of 2025 will have practically the same general structure 
as 
my preferred cladogram of 2003. 

Go on, prove me wrong!

Dave