29 



only in diflFerent species, but in individuals of the same 

 species ; yet it is a valuable diagnostic mark. In some species 

 the number of ventrals does not vary more than five per 

 cent., in others it is very variable. Thus Tropidonotus 

 quincunciatus has very constantly either 137 or 145 

 ventrals, whilst T. stolatush&s from 121 to 161. The genus 

 Ablabes is most irregular in this respect, the species having 

 from 122 to 245 ventrals. And the genus Tragops, a 

 well-defined genus of tree-snakes, consists of three species 

 having respectively about 151, 190 and 220 ventrals. The 

 number of subcaudals is very variable, it is generally greater 

 in males than in females owing to the increased length of 

 the tail in the former sex. 



The tail ends in a single shield. In most snakes this 

 shield is not remarkably developed, but in the Uropeltidce 

 it becomes a broad disk either naked or covered with keeled 

 scales, and in Plectrurus it ends in a forfical bispinous 

 shield. In some snakes with rather short tails, it may be 

 found quite pointed, sufficiently so to prick the skin slightly 

 if roughly handled. The Crotalidce have this point well 

 developed, though the tail is not provided with the rattling 

 appendages peculiar to some American genera of the family. 

 It is remarkable however that the Indian Crotalidce have 

 a habit of vibrating the tail when excited, and if it strikes 

 against any hard substance a slight noise is produced. 



In the sea-snakes the tail is compressed laterally forming 

 a vertical fin like that of an eel. 



The extremity of the tail is liable to accidental mutilation, 

 and it frequently happens, in consequence, that the number 

 of subcaudals is found much below the regular number.* 



* The Indians have a fable that the cobra loses an inch of its tail 

 each time that it bites a man. Stumpiness of tail is not however 

 confined to this snake and amongst the less agile species a large 



