648 S. W. WILLISTON 



going back to the first position held by Cope, as well as that 

 held by Baur and Osborn. It may truthfully be said that it 

 would not require a very great extension of the list of inconstant 

 characters given on the preceding pages to include most of Os- 

 born's Diaptosauria. I draw the greater dividing line elsewhere, 

 that is, between the Proterosauria and Rhynchocephalia rather 

 than between the Pelycosauria and Cotylosauria or Therapsida. 

 I am proposing no new scheme of classification in the present 

 paper; I am only endeavoring to show the futility of some of 

 the present ones. Perhaps, when we know a great deal more 

 about the older reptiles, we may be able to recognize a true phy- 

 letic classification, but I do insist that we have not that requisite 

 knowledge at the present time, that the most we can do is to 

 venture guesses, shrewd guesses perhaps, but still guesses, and 

 that it is better to endure our present ills than to fly to others 

 we know not of. 



A minute comparison of the characters of Paleohatteria with 

 those given under the precedinglistsof the American Theromorpha 

 would require too much space here, and is unnecessary. Suffice 

 it to say that the diagnosis given by Credner (op. cit., p. 545) 

 applies almost word for word, to the American forms with the 

 exception of sclerotic plates in the orbits, which can be interpreted 

 only as an aquatic character. Not only do the known characters 

 agree throughout with those of the Theromorpha as a whole, 

 but almost all of them will apply to the family Poliosauridae, 

 and Ophiacodon and Varanosaurus especially; the first two 

 paragraphs of Credner's diagnosis will, indeed, apply verbatim 

 to Ophiacodon; and, so far as the characters of Proterosaurus 

 from the upper Permian are known, there is little or nothing 

 discrepant; but I omit their discussion here. 



While it is very possible that Paleohatteria has two temporal 

 arches on each side, the presence of the upper fenestra has never 

 been positively proven. Credner in his first description merely 

 says that the presence of the supratemporal fenestra is 'wahr- 

 scheinlich.' But its presence or absence is really unimportant 

 in consideration of the fact that the two vacuities, both upper 

 and lateral, are known to occur in at least one genus of American 



