CuAP. I. THE SKIN. 259 



the tliin, elastic epidermis. As in all Vertebrata, so also is the epidermis in 

 Turtles, composed of characteristic cells, of an hexagonal or irregular form, which 

 are dry and tlat near the surface, and more or less imbricated, while their contours 

 are better defined the deeper we penetrate and the more we approach their matrix. 

 But in relation to the mode of growth and the duration of these cells, upon a 

 larger scale, up to the time when they are cast in moulting, we find the greatest 

 variety among Turtles, as we find, indeed, among all the different types of Verte- 

 brata. The diffei'ence.s in the epidermal formations, observed in Turtles, naturally 

 lead us to expect such a diversity among them in particulai*. In the Sphargididae 

 and the Trionj'chida^ I have had no opportunity of seeing a regular casting off" of 

 the epidermis, though there can be hardly any doubt that a change of the epidermis 

 takes jilace here, and that it is effected by the dropjjing of single cells or of thin 

 layers, for I have noticed it in Trionyx, as we find it in the epidermis of Frogs 

 and of Man himself, in whom it is quite similar. But in all other Turtles, the 

 nature of the epidermis, and therefore its moulting also, ai'e entirely different. In 

 Eretmochelys irabricata, the plates of the shield (the tortoise-shell of commerce) 

 are very large, and imbricated one above the other. These plates increase only in 

 front, where they are imbedded in a thick matrix, in the Malpighian layer, as in 

 a case. As the plates enlarge in front, the older parts must move backwards, 

 where they are worn off" by external mechanical agencies. This process goes on 

 so rapidly in these Turtles, that in a specimen of two feet in length, no trace 

 of those primary scales, Avhich covered the whole shield diu'ing the first year, 

 could be found. This mode of growing and moulting, if we may call it so, is very 

 similar to that in the human nail. But we find a very different jjrocess in land 

 Turtles, and to some degree also in Cistudo, in which the plates rest entirelv, 

 in front and on the sides and behind, upon their matrix, in the Malpighian layer. 

 They are not at all free and raised behind, as is the case in Eretmochelys, 

 and thus they grow not only in front, but with their whole under surface and 

 on all sides ; hence it follows that we find upon the surface of each scale, around 

 a sm;\ll angular central plate, (the scale of the first year's growth,) a smaller or 

 greater number of concentric stripes or regular annual rings, as they are exhibited 

 on a transverse section of an old tree.^ 



' This is remarkably obvious in some specimens perfectly smooth, so tliat llicir successive growtii 



of the Xerohiites carolinus (Testudo polyphcnuis) and their age can no longer be read upon the plates, 



of our Southern and Soutli-wcstcrn States, and always as it is easy to do in many other species of that 



in Testudo radiata of Madagascar, and in Testudo family. The Gopher, and perhaps also the Gala- 



geometrica of the Cape of Good Hope. In ri'Ia- jiagos Turtle (T. indica) burrow into the ground 



tion to the Gopher, (Test, polyphemns.) I have to and live in earth holes, and this accounts, perhaps, 



remark that the plates of most adult specimens are for their worn and polished plates. Hut why should 



