INTRODUCTION. 645 



the following classifications of Dumeril and Bibron of Ophidia and 

 Lacertilia, and Giinther of Anoura, and Agassiz, as modified by Cope, of 

 Testudinata, are perhaps the best attainable. 



That many of the marks used in describing species in this treatise are 

 not constant, the writer and every student of Herpetology well knows. 

 Thus the presence and absence of a loral or ante-orbital plate may both 

 be seen in the same animal on opposite sides of the head; the number 

 of rows of dorsal scales varies in different individuals of a species; 

 the coloration and arrangements of the spots and stripes, the number of 

 the upper and lower labials, in fact the cephalic plates are liable to 

 become more or less fused and run together ; the number and arrange- 

 ment of the shields in the carapax, and even the shape of the head is 

 more or less variable. In using such marks, for the purposes of 

 description, the writer but acknowledges the imperfections of this 

 branch of Zoology, and hopes that after this suggestion, no one will be 

 misled by any of these variable characters at times used in the synopses, 

 but that they may be found of service in the identification and study of 

 the species. 



One of the most difficult things about the study of these animals is 

 that their colors change so much when placed in alcohol. Thus yellow 

 becomes white ; green, blue ; and red, brownish-black ; while brown and 

 metallic tints remain for some time unchanged. However, exposure to 

 the sun for a season will often enable us to form an idea of the original 

 shade. 



Sex may be told in most if not all Turtles by the males having convex 

 plastra, and in the Anoura by the males, in the greater portion of the 

 species, being supplied with vocal vesicles. 



