574 Correspondence — Prof. E. D. Cope. 



the result would be more complex than that which I have preferred, 

 for not only would the significance of the name Hyracoidea be 

 changed, but I would have to make a new name for the Hyracoidea 

 proper of older authors. The same reasoning applies to the case 

 of the Insectivora. I must either change the significance of the now 

 generally adopted term, or give a new name to the major division 

 of which it forms only a part. 



I now come to a second branch of the subject, and that is, the 

 question of the definition of divisions. I lay down the principle, 

 which is, I believe, a generally accepted one, that classification con- 

 sists of precise definitions ; and that the aim of the systematist 

 should be to attain as great precision as the nature of the case will 

 permit. The fact that all definitions which separate adjacent groups 

 will be ultimately found to be fallible, does not permit us to fall 

 into inexact and inconsistent methods of definition. Any absolute 

 difference in the number of parts must be noticed in the system as 

 indices of the steps on the lines of descent. It is only proportions, 

 dimensions, and texture which define species, as represented by their 

 fossil remains. Any system which places animals with four digits 

 in the same genus with others possessing three digits, is inexact. 

 Any system which places animals with four premolar teeth in the 

 same genus with those with three premolars is inexact, unless it can 

 be shown that some species has indifferently three or four premolars. 

 A system which includes animals with a scapholunar bone in the 

 same division with animals with distinct scaphoid and lunar bones, 

 is inexact. I think it will be ultimately agreed that animals with 

 tritubercular superior molars must be more widely separated from 

 those with quadri tubercular s, than has hitherto been accepted. 

 Nevertheless I admit that every character loses its value somewhere 

 because of variability. But cases where such are not variable must 

 be carefully distinguished from those where they are. 



To again illustrate this point from those brought forward by Mr. 

 Lydekker. My critic " prefers " to arrange the Creodonta with the 

 Carnivora, although they have no scapholunar bone. Now, I ask, 

 what will be the definition of the order Carnivora, if we reject this 

 character ? There will be none, as the latter constitutes the only 

 bond of union between its diverse forms. As if an unconscious 

 cerebration opposed this view, Mr. Lydekker maintains the division 

 Creodonta, but gives it a new name, " Carnivora primigenia," a 

 name to which the law of priority opposes an objection. Secondly, 

 on any character at present known the division called Insectivora is 

 heterogeneous, and so soon as exact definitions are demanded, it 

 requires subdivision.^ In order to distinguish these subdivisions as 

 a whole, with the allied ones, which are clearly neither Edentata, 

 Chiroptera nor Eodentia, I applied to them all, as an order, the 

 name Bunotheria. This group is as definable as any of the others 

 named, and has equal rank. It is the primitive form of placental 

 unguiculate mammal, just as the Taxeopora is the primitive type of 

 primitive ungulate mammal. 



1 Some of this work remains to be done. 



