422 



AMERICAN TESTUDINATA. 



Pakt II. 



be plain that the characters of the immediate instruments of these acts are 

 essential characters, and that any peculiarities and identities among them must be 

 important in determining their natural relations. In Turtles the jaws and the 

 neighboring pai'ts are the principal organs concerned in these acts; and the claws 

 and limits, which generally perform so large a part in the movements connected 

 with the function of nutrition in some of the higher types, have here little or 

 nothing to do with it. Moreover, in Turtles the structure of the jaws and their 

 muscles determine, to a great extent, the structure and form of the whole head. 

 About the jaws and head, then, are we to look, in this order, for the structural 

 characters which belong to the voluntary acts relating to nutrition ; and here, and 

 here only, do we find the distinguishing characters of the natural groups that may 

 be distinguished within the fimilies and sub-iamilies. Months of research in the 

 family of Cinosternoid£e, and in corresponding groups of other families, have failed 

 to point out any other organs as bearing distinctions and characters for these 

 groups. Indeed, leaving out .specific characters, it is impossible to identify any 

 other pai-t of the body of these animals, when examined isolatedly, as belonging 

 to one or the other of these groups.^ It thus appears that there are, among 

 Turtles, natural groups founded upon the organs w^ith which these animals take 

 their food, and upon them only. These groups, unquestionably, are genera. 



In preceding families I have not hesitated to insist at once upon the generic 

 value of similar characters, trusting that the similarity in the range assigned 

 to the genera which I was led to adopt upon such a foundation, with other gen- 

 era already acknowledged as such, would not foil to convey the same conviction 

 to the minds of other naturalists. But, the Cinosternoidse are to this day so imper- 

 fectly known, the genera proposed by the ablest herpetologists are still so unsatis- 

 factorily characterized, and, above all, the opinion expressed by Schlegel and Tem- 

 minck- upon these Turtles is so diametrically opposed to the results to which I have 

 been led, that- I felt it indispensable to show, on this occasion, in what way, and 

 by what evidence, I have satisfied myself, step by step, that the family of Cinos- 

 ternoida3 is a natural family, embracing two distinct sub-families,^ each of which 



^ I mean to say, that parts of the body of a Turtle 

 found separated, as is mostly the case with fossil re- 

 mains, cannot be referred to their genus with cer- 

 tainty, unless the jaws be among them ; or unless the 

 parts found bear specific characters that occur only in 

 well known genera. This result is of the utmost im- 

 portance to Palceontology, and may explain why 

 Cuvier did not attempt to determine the generic char- 

 acters, and to give specific names to many of the fos- 



sils which he described. It may also serve as a warn- 

 ing to those paliBontologists who never hesitate to 

 distinguish fossil species without sufficient preliminary 

 comparisons with their living representatives, and 

 sometimes upon the most insignificant fragments, 

 which do not exhibit the first specific character. 



- Fauna japonica ; Chelonii, p. 59-62. 



' Already alluded to, (p. 250 and 251,) when 

 contrasting Ozotheca with the old genus Cinosteruum. 



