Chap. I. SPECIAL CLASSIFICATION OF TESTUDINATA. 245- 



founded, though these characteristics are confined to certain parts, instead of extend- 

 in<f to the whole organization. 



The next question which we have to consider here is, whether these sulj-urders 

 exhanst the natural subdivisions existing between the order and the genera; or, 

 in other words, whether in this class the orders coincide with the families or not, 

 for we have not yet examined the question whether every order has necessarily 

 more than one family or not. My remarks in the third chapter of the first 

 part of this work can leave no doubt that each of the four branches of the animal 

 kingdom contains several classes, for we have seen that every one of them dis- 

 plays the plan of structure on which it is founded, as carried out in different 

 ways and with difierent means. But we have seen from a supposed case, that 

 if such a class included only a few species, or even several genera, or perhaps 

 one or more families, there might be no foundation for a distinction of orders, 

 if all these species, genera, and families presented only such a diversity of vdtimate 

 structure and such modifications of form as would not distinctly indicate among 

 them a difference of rank, an appreciable gradation.' But where a class contains 

 groups in which such differences as mark gradation and rank are clearly percep- 

 tible, then we have distinct orders, even should these orders coincide with the 

 limits of the families, that is to say, be combined with such modifications of 

 form that, though expressing a gradation, these groups would correspond with the 

 characters upon which families are to be founded. Now it remains for us to 

 examine whether this is the case among Testudinata ; and since the Chelonii 

 constitute so natural a sub-oi'der, when contrasted with the Trionychidae, the Emy- 

 doidae, and the Testudinina, we may select it as a test of the existence of sub- 

 orders in nature, and we shall afterwards extend our remarks to the other minor 

 groups with the view of ascertaining how many divisions of this kind there truly 

 are in the order of Testudinata. 



Ever since naturalists have attempted to subdivide the Testudinata, those with 

 pinnate limbs have been considered as a natural group, raised by most to the dig- 

 nity of a family, and embracing, in all modern classifications at least, two genera, 

 Chelonia and Sphargis, though some authors subdivide farther Chelonia into several 

 genera, and even go so far as to consider Sphargis and Chelonia proper as the 

 types of distinct families. Now, whether that group contains one or two families, 

 it iniquestionably exhibits very great uniformity of structure as a group, wlu-a 

 compared to the other Testudinata. In the first place, the dermal ossification 

 remains imperfect ; next, the limits preserve through life a character Avhich is uni- 

 form in Testudinat^a, as long as their development is not complete, that is to say, 



» Sec Part I., C'lmi). 1, Sect. 1, p. b-1. 



