VARIATION OF THE PIT-VIPER. 177 



The fourth type — L. jararacussu (text-fig. 5) — constitutes the 

 third branch in the evolution of the markings. A 1, 5, 2, 6, and 3 

 become confluent, forming a subcrescentic marking, in the middle 

 of which A 4 may be present or not. All the primitive spots of 

 group B unite together to form a dark triangular area, which in 

 the course of evolution becomes confluent with group A — i. e., 

 the apex of A with the base of B — along the dorsal line. At this 

 point the markings darken, and the usual brownish, greyish, or 

 olive ground-colour becomes a bright clear yellow. Finally, more 

 and more black pigment becomes deposited at A and B, until the 

 markings have usurped the place of the ground-colovir, even the 

 crescentic marking (A) being partially filled in with black. The 

 general effect of Type IV. is therefore a black snake, with yellow 

 forked markings issuing transversely from the yellow ventrals, 

 which are checkered with black. The dorsal surface of the head 

 is completely black, owing to the development and confluence 

 between the streaks of the dorsal and dorso-lateral series. The 

 post-ocular streak is separated from this area by a narrow stripe 

 of the clear yellow ground-colour, which continues slightly below 

 the commissure of the jaws. 



The most striking feature about this type of marking is the 

 clear yellow of the interspaces in the doi:sal pattern ; in other 

 types this is only found on the ventral surface of the snake, the 

 dorsal ground colour always exhibiting a superficial powdering of 

 dark pigment. It therefore seems probable that in the course 

 of evolution this pigment has all aggregated in the spots of 

 Groups A and B, and, as more and more becomes deposited, these 

 areas further encroached upon the yellow ground-colour. 



V. Phylogenetic Relations between the Types. 



In tracing the evolution of markings from one type to another, 

 I have assumed that the anterior parts represent the primitive 

 pattern, as in lizards and many mammals (ex. Lacerta 'niuralis, 

 Equus quagga). It might appear, however, that in X. cdrox this 

 is not the case, the more regularly spotted type, ciiaracteristic of 

 the West Indies representing the original pattern, from which the 

 clustered, annulated, and forked markings of C. and S. American 

 forms were derived. On careful consideration I cannot adopt 

 this view for the following reasons: — 1. The manner in which 

 the pattern of the Jararacussu develops from Type I. into the 

 complicated design of Type TV., and develops on each individual 

 posteriorly until the tail is a uniform black, cannot be looked upon 

 as primitive, and if this elaborate pattern, become simplified, 

 is a further development of the clustered type of marking, the 

 other forms must also be more advanced. 2, I hold the greater 

 number of spots to be the more primitive (and for this assumption 

 there is much correlative evidence among other snakes and 

 lizards). The clustered pattern is therefore less advanced than 

 the more evenly spotted, for in the former groups A and B 



Phoc. Zool. Soc— 1918, No. XII. , 12 



