ANCESTRY AND CLASSIFICATION. . 247 



cate, on the basis of existing affinities, the rela- 

 tive time at which the different groups diverged 

 from each other or from the main stem ; and the 

 height which each branch attains the relative 

 perfection of the highest members of that group. 

 It is of course impossible to represent this with 

 any accuracy on a flat surface ; for one may prop- 

 erly conceive of a group only as a mass, with 

 branches springing from a common central core. 

 The swallow-tails and lycaenids are thus brought 

 at opposite extremities of the tree, whereas 

 they are closely related to each other, and dis- 

 agree with all other groups in the relation of the 

 head of the caterpillar to the segment behind it ; 

 this relationship, however, is indicated by each 

 occupying the lowest twig of the branch on which 

 it is seated, both branches being closely connected 

 at their base. The striking and unique peculiar- 

 ities of certain groups is shown by their extreme 

 divergence from the main stem ; thus the swal- 

 low-tails stand apart from all others in the posses- 

 sion of dorsal osmateria in the caterpillar, and in 

 certain special characters of the butterfly, shortly 

 to be mentioned ; the lycaenids at the opposite 

 extreme, in the onisciform nature and diminu- 

 tive heads of their caterpillars ; the castnioides 

 among the skippers by their close approach to the 

 moths, and the satyrs or meadow browns by the 

 forked tail of their caterpillars. The superficial 



