I20 THE PERMO-CARBONIFEROUS RED BEDS OF 



As he says, the classification does not take into consideration the ten- 

 dencies of evolution in the group. 



"I am aware that this classification is what might be called 'horizontal,' rather 

 than 'vertical;' but I also insist that horizontal classifications are absolutely im- 

 perative until such time as we have more than vague surmises and guesses as to 

 the true lines of phylogeny." 



The opposite point of view is adopted by Osborn in his paper on the 

 "Reptilian Subclasses Diapsida and Synapsida, and the early history of the 

 Diaptosauria." He says:^ 



' ' In my opinion classification should directly follow, formulate and express our 

 progressing knowledge of phylogeny ; for example, if the Proterosauria prove to be 

 ancestral to the Dinosauria, as I suspect, then the order Proterosauria should be 

 transferred to the superorder Dinosauria, as constituting a common stirp, a common 

 blood relationship, a common tendency to evolution. As soon as a certain evolution- 

 ary direction is assumed by a certain group of animals toward a higher group, 

 although the evolution may not have proceeded very far, the lower and the higher 

 groups should be united in classification, and thus segregated from groups diverging 

 in other directions." 



After all, these two writers are not far apart, for, were Dr. Williston con- 

 vinced of the correctness of the phylogenies proposed, he would assuredly 

 follow the plan suggested by Osborn. 



The author has discussed the early reptiles in great detail with Dr. 

 Williston, and is in accord with him in most of his statements and conclu- 

 sions, but can not entirely agree with him in his proposed classification. He 

 is inclined to think that Dr. Williston has assumed too conservative a posi- 

 tion in the matter, and to think that enough of the main phylogenetic lines 

 are determinable to warrant a more definite statement. 



WiUiston, in his analysis of the Permo-Carboniferous reptiles, gives the 

 following list of constant characters in the Theromorpha : 



One or two temporal vacuities on each side. 

 Neck of greater length. 

 Neural arches never stout. 



The last character can not be strictly maintained, as Pcecilospondylus 

 Case has the neural arches of a portion of the vertebral column quite stout. 

 This is the only exception to the rule, so far as I know. 



The contention of Dr. WiUiston is that all the other characters, inconstant 

 ones, are not sufficient to separate the American and the South African forms 

 ordinally, and they must be combined. But the constant characters are 

 extremely primitive, exactly such as would be found in a common ancestral 

 group, and are such as might be retained in many variations in widely 

 divergent lines. 



Although it is frankly admitted by the author that divergence has not 

 gone far enough in the Permo-Carboniferous reptiles to eliminate all the 



" Osborn, Memoir Amer. Mus. Nat. Hist., vol. i, part viii, p. 504, 1903. 



