180 MISS JOAN B. PROCTER ON THE 



consist of 6 spots each, whilst in the latter group A consists of 5 

 and B of 4. Besides these two points, I attach importance to 

 the fact that the eight primary series of spots never extend 

 beyond the head. 



As we have already seen. Types II., III., and IV. are all 

 directly evolved from Type I. ; but their relative positions have 

 yet to be discussed. Of the stages of development evinced 

 between each of these three and the primitive type, many show a 

 tendency for the markings to break up into the spotted type 

 (Type III.) just before the tail. Nearly all specimens with mark- 

 ings ranging from Types I. to II. show this peculiarit}', also all 

 forms intermediate between Types I. and III. In the series 

 from Types I. to IV. this is the case in intermediate forms, but 

 Type IV. itself maintains its direct line of development, the 

 black markings encroaching more and more upon the yellow 

 ground-colour towards the posterior part of the snake, the end of 

 the tail being black. 



It will therefore be seen that phylogenetically the markings 

 of Type III. are not to be regarded as so primitive as those of 

 Type II., and that Type IV. is the most modified. 



In the next chapter, on classification, varieties will be based 

 on these types of markings in conjunction with their geographical 

 distribution. 



Having traced the markings of L. atrox fi-om the most 

 primitive onwards, it is of some interest to try to ti'ace these 

 markings back to the hypothetical four-paired streaks of primitive 

 snakes. 



The first stage towards this would be the equal development of 

 group B with group A, and the commencement of the fourth or 

 lower labial series on the body. The series would then become 

 more equalised, until a pattern were reached of 4 rows of equal 

 and alternating spots — independent from those on the oppo- 

 site side— which would be but the result of the breaking up 

 of the primitive streaks. But L. atrox is very far ahead of this 

 prototype, and such cases as might appear at first to be an 

 approximation to it, prove on " careful consideration to be 

 secondary, an instance of deceptive apparent reversion. 



VI. Classification. 



It will be seen from the foregoing that no very definite 

 boundaries can be traced between the varieties of L. atrox, 

 whether based on structure or on coloration, so complete are the 

 links connecting them. I can, however, recognize four principal 

 foi'ras based chiefly on markings, more or less in conjunction 

 with geographical distribution, which I think afford on the whole 

 a more satisfactory basis for classification than the characters of 

 scale-structure, etc. 



The first of these forms is L. atrox, var. affinis, which ex- 

 hibits the more primitive type of marking (Type I.), viz., the 



