60 
The above arrangement of the British butterflies is based upon the 
specialisation of the neuration within the groups. I have taken the 
families as I found them, and I have used the specialisations of the 
wing as a guide by which to arrange the genera within the families. 
I have merely tried to avoid the course of Mr. Meyrick, of deriving 
a specialised from a generalised genus, and then reversing the process 
and making the specialised give birth again to a generalised group. 
Where I have fallen short it will be found either that my guide has 
forsaken me (as in the Blues, where the neuration is so uniform that I 
have not studied it in all the species, and thus not so thoroughly as I 
should), or that I have not properly attended to its teachings. I state 
this as plainly as possible, in order to avoid the repetition of the 
reproach that I have taken an “artificial” character to the detriment 
of a “natural” classification, though what there is “artificial ” about 
the neuration I do not at all comprehend. What these critics mean, 
probably, is that I have leaned too strongly upon a single character or 
class of characters; and I have tried to show above that this 
criticism arises from the critics not understanding what I have chiefly 
done. But I have certainly not rejected the main evidence of the 
neuration where it has been overwhelming, as in the case of Vemeobzus. 
A critic in “ Psyche” cites my founding a distinct family upon this 
type as a fatal instance of my reliance on a single character. But I 
suspect that this critic himself may have so conspicuously neglected 
the neuration, and the lessons to be derived therefrom, that I should 
be able to find, perhaps, North American Nemeobiidz among his 
Erycinidz or Riodinidz. So far from entertaining any undue par- 
tiality for the neuration otherwise than as a means to an end, viz. 
the comprehension of the phylogeny of the butterflies, I have at once 
allowed the Riodinidz to stand, although they are quite on a level 
with the Zephyrini, and their separation from the Zheclime must be 
urged mainly, I think, on other grounds than the condition of the 
wings. The case of /Vemeodius is that unless we throw the structure 
of the wings entirely overboard, and neglect its teachings altogether, 
we must agree that here we have so distinct a type, as compared with 
the Lyceenid, that the character is sufficiently strong to support a family 
separation. For if the wing of /Vemeodius were studied by itself, it 
would be thought to be that of a five-branched Pierid. A parallel 
suggests itself with the Megathymidz. This North American group 
is a highly specialised one, an offshoot from the Hesperid type, in 
some respects retaining perhaps ancestral features, perhaps specialised 
through what we are apt to call degeneration,—as, for instance, are the 
Citheroniine also. In J/egathymus the middle branch of the media 
has abandoned its indifferent original position, and has succumbed to 
the attractions of the cubitus, as I see from Prof. Comstock’s beautiful 
and reliable figures. As to the position of Papz/io, we can at once 
understand how flagrant a mistake it is to place it between Zycena and 
Flesperia, when we see that the Lyczenid wing is deducible from the 
Hesperid and the Papilionid wing is not nearly related to either. 
