COLTJBEID^. 



171 



rhachis by bending the body dorsally. The application of this 

 character, first employed for taxonomic purposes by Prof. Cope, 

 but which had not hitherto been tested in a thorough manner, 

 leads to some unexpected results as to the aiHnities of many genera 

 and species the position of which appeared somewhat problematical. 

 Thus all the Madagascar Coluhrinm have the h)'papophyses deve- 

 , loped throughout the vertebral column, and thus differ from the 



Fig. 12. 



Posterior dorsal vertebrae of : — 



A. lAoheterodon madagascariensis. B. Heierodon nasicus. 



a. Back view. &. Lower view, u. Side view. 



American genera Liojphis, Heterodon, Bromicus, &c., with which 

 they have long been associated, although, indeed, a careful com- 

 parison of their external structure alone shows these views to 

 have been based merely upon very superficial resemblances. As 

 regards those dwarfed, degraded forms which have hitherto been 

 associated as Calamariidce, I have endeavoured, as far as possible, 

 to bring them into nearer neighbourhood to such more powerful 

 Snakes, from which, so far as the value of their characters can 

 be correctly estimated, there is reason to believe they are derived. 

 Thus Haldea and StreptopJiorus are placed near Tropidonotus and 

 allies, as proposed by Cope, Simotes and Oligodon near Coronella, 

 and so on ; in the same way as in the family Boidce the genera 

 Eryx, Lichanura, Charina, Bolieria, and Erebophis have been 

 incorporated among the Boas, and Loxocemus and Calaharia 

 among the Pythons instead of being grouped together as '■'■Eryddce.'" 

 In the case of the Boas the course followed is so obviously in 

 accordance with the spirit of a natural classification, that I do 

 not anticipate any objection being raised against it. It has been 

 my aim to carry out the same principle in dealing with the large 

 and far more difficult group of the Oolubridce. 



