The problem with multivariate analysis.

[david peters <davidrpeters@earthlink.net>]

Workers have been trying to decipher the phylogeny, nomenclature and affinities 
of Pterodactylus and Rhamphorhynchus for over 150 years. They see differences 
and similarities. They try to quantify their observations in an attempt to 
answer the question: Are there multiple species improperly lumped into 
nomenclature wastebins? Or are many specimens properly lumped into single 
species?

Mateer 1976, and Bennett 1995, 1996 both performed statistical studies on 
pterosaurs using skull and long bone measurements in an attempt see how chart 
dots grouped and lined up. Their work has been almost universally accepted. 

The problem is this, according to a larger cladistic analysis examining dozens 
to hundreds more characters: the variation within these taxa do not occur so 
much at the long bones, but rather at other relatively minor parts of the 
skeleton: teeth, toes, fingers, face, pelvis, etc. After all, all properly 
named Pterodactylus and Rhamphorhynchus specimens are related at the genus 
level (which may still be too broad when compared to extant taxa, but thatâ??s 
for another day). Perhaps that is why both Mateer and Bennett tended to lump 
multiple and various specimens into single species, applying observed 
differences to ontogenetic stages of development.

One might have similar results performing a similar sort of statistical 
analysis on a selection of passerine birds or microbats using only long bones 
to create a data table. (Now thereâ??s a project!) The overall proportions 
[just guessing here] appear to be grossly similar in a number of taxa despite 
some variation in size and genus. That may be because both taxa have 
aerodynamics operating as a selective force. As in pterosaurs, the ignored 
variation, it would appear, would only show up in the smaller overlooked 
details of beak and ear shape, etc. 

So while the statistical analyses of Mateer and Bennett are interesting 
exercises, if one wished to determine affinities at the species and genus level 
in flying taxa then it would appear that one must perform more detailed 
cladistic analyses and probably create reconstructions of the roadkill to 
improve understanding, especially when trying to argue a case. You all know how 
educational Greg Paul reconstructions can be. The differences and similarities 
are illuminated almost at first glance. 

Have a great time in Denver!
David Peters
St. Louis

Bennett, S. C. Â 1995 statistical study of Rhamphorhynchus from the Solnhofen 
Limestone of Germany: Year-classes of a single large species. Journal of 
Paleontology, 69:569-580. 

Bennett, S. C. Â 1996 Year-classes of pterosaurs from the Solnhofen Limestone 
of Germany: taxonomic and systematic implications. Journal of Vertebrate 
Paleontology, 16:432- 444. 

Mateer, N. J. 1976 A statistical study of the genus Pterodactylus. Bull. geol. 
Inst. Univ. Uppsala 6, 97â??105.