174 MISS JOAN B. PROCTER ON THE 



posterior parts. One specimen, however (see Table, Trinidad 1 A), 

 is inclined towards the Jararaca type of marking. Anteriorly 

 A 1 is divided, each spot being confluent with A 5 and 6 ; A 4 is 

 absent, and of group B only an occasional- B 1 is present (text- 

 fig. 3, d). At about tlie middle of the snake the twin spots 

 A 1, 1, have become some distance apart, and are in sufficiently 

 close proximity to A 2 and 3 to give the appeai-anee at a distance 

 of paii'ed transverse bands, or, since the intermediate space is 

 darker than the surrounding ground-colour, of a single dark- 

 edged band ; these bands, which may pair with those of the 

 opposite side or not, are very little more than their own width 

 apart. 



In all specimens the pattern tends to break up into a spotted 

 type just before the tail, all the spots being more or less equal in 

 size and distribution. It consists of A 1, 1, 4, 2, 3, and B 1, 

 3, 4, and 2. 



In the markings of the mother, A 1, 2, 3, 5, and 6 are con- 

 fluent, forming a somewhat truncate chevron, in the middle of 

 which A 4 is sometimes situated. Group B is made up of the 

 six primitive spots, all except B 1, 2, however, being indistinct 

 (text-fig. 3, b). On the whole, the individual variation shown by 

 these 27 specimens is very slight : the most apparent differences 

 between them are due not to variation of the actual markings, 

 but to the inconstant I'elationship of those of the two sides of 

 the snake. 



1 have since examined another mother and young, from Anda- 

 goya, Colombia (received from Dr. Spurrell), and in this case 

 there is no variation at all among the young. 



The markings of both families are of the most primitive type 

 occurring in Z. atrox^ all other forms being evolved from it, as 

 will be shown presently. 



Evolution^ 



In the course of evolution, these primitive markings become 

 modified in the following way, which is the same for all forms of 

 L. atrox. 



A 1 tends to divide transversely to the axis of the body, and 

 A 5 and 6 become confluent with it. B 5 and 6 merge into the 

 wround-colour, A 4 and B 3 and 4 disappear (this stage has 

 already been described in the variations of the young snakes), 

 A 1 then divides completely, but is still distinct fi-om A 2 and 3._ 

 A 4 disappears, and the whole of group B tends to merge into 

 the background. The spots A 1, 1 then become a considerable 

 distance apart, narrow, and in- close proximity with A 2 and 3, 

 which, since the interspace is darker than in group B, thus 

 produce the effect of a dark, darker-bordered, transverse band. 

 Except for an occasional B 1, this group is pale and quite 

 indistinct. At this point the bands, which may or may not pair 

 with those of the opposite side, are slightly more than their own 



