27 



In some snakes the number of rows of scales settles down 

 very soon to the normal number which continues till very- 

 near the tail, in other snakes barely four-tenths of the trunk 

 is occupied by the normal number. Thus, in Bungarus 

 fasciatus and the genus Callophis, the number of rows 

 settles down, at about the tenth ventral, to the normal 

 number 15, and remains at that quite down to the vent ; 

 but' the more common arrangement is shown in the following 

 diagram, dividing into tenths the body of a snake with 19 

 rows of scales : — 



Tenths. Bead. 123456789 10 Tail. 

 No. of rows. 21 19 19 19 19 19 17 17 17 15 10 8 6 4 



We have seen that the scales may be more or less over- 

 lapping or i/mbricate ;. another important character is the 

 presence or absence of a raised keel along the centre of each 

 scale. This keel is very capricious in its presence or 

 absence, and does not correspond to the habits of the snake. 

 It is found in ground-snakes, in water-snakes, and in tree- 

 snakes, indifferently, and in each of these groups, the 

 snakes with smooth scales appear to progress just as well as 

 those with keels. The viperine snakes have generally 

 strongly keeled scales. These keels are generally most 

 marked on the dorsal rows, they become faint towards the 

 abdomen, and rarely appear on the outer row of scales. 

 This row is generally of larger size than the others. In the 

 ErycidoB, the keels on the caudal rows of scales are so 

 strongly marked as to form sharp longitudinal ridges ; and 

 as they rake backwards like the teeth of a ratchet-wheel, 

 they would apparently be of great use as aids to progression 

 in burrowing. 



The presence of an opposite feature — grooves — is fre- 

 quently noted ; they are far from obvious, and are best seen 

 by letting the light glance off the scales; they are then 

 visible as one or two minute gougings at the apex. 



