394 NATURAL SCIENCE [December 
Heterocera). Here also we may refer to the features A, B, C of Anosia 
(evidently overlooked by Dr Grote); as already stated, these are 
not to be observed among specialized Nymphalids. Now, the 
cubitus-anal connection, short anal nervure of the forewings, and 
subcostal-radius connection in the hindwings of Papilionids are 
generalized features, homologous with the features A, B, C of Anosia, 
modified and lost in other groups of Rhopalocera ; it is impossible to 
regard them in any other light. In this respect then Papilionids 
approach the generalized Nymphalids. 
Reaction of modification appears to me to have caused different 
lines of specialization of neuration in the Lepidoptera, consistent 
with adequate support of the wing membrane, and the develop- 
ment of wing area and shape; for example, it seems that modifi- 
cation of neuration has been rapid in the case of Opostega 
crepusculella, which is not highly specialized in wing area and 
shape as compared with primeval Trichoptera-Lepidoptera, but 
the neuration has been reduced to several simple (unbranched) 
nervures. The area and shape of Rhopaloceran wings are, however, 
more highly specialized, and the modification of neuration——particu- 
larly of the forewings—has been less rapid. If we draw the line 
separating primitive Rhopalocera from the mass of primaeval 
Lepidoptera at Hesperids, then we have the fact that no 
Rhopalocera possess the basal portion of the median nervures, 
and this loss we may assume to be the point of separation 
(granting, also, modification of hindwing neuration, which every- 
where in the Lepidoptera has preceded that of the forewings). 
Next we have noted the formation of three-branched cubitus 
in the forewings, by the incorporation of one of the median 
nervules with the cubital system in Hesperids and subsequent 
specialization in the direction of Lycaenids. In another direction, 
apparently at some distance from Hesperids, we note that a three- 
branched cubitus is a fixed feature in the neuration of Nymphalids’ 
forewings, that no movement has taken place in the position of the 
other median nervules, and very little subsequently among Nym- 
phalids and Satyrids. But as the Pierids separated from the 
Nymphalid-Pierid stem, the median nervules moved towards, and 
one became incorporated with, the radial system. The cubitus 
‘blotch’ and rudimentary anal nervure of the forewings and the 
subcostal-radius connection in the hindwings of Anosia point to the 
possession of the Papilionid cubitus-anal connection, two anal 
nervures in the forewings, and a subcostal-radius connection in the 
hindwings by those primaeval Rhopalocera from which the Nym- 
phalid-Pierid groups originated; and another important fact is that 
at this point two anal nervures in the hindwings were retained 
(as in Hesperids). Next in sequence come the Papilionids, in 
which movement of one of the (Nymphalid) median nervules forms 
