1898] CLASSIFICATION OF THE DAY BUTTERFLIES 95 



Sub-fam. 2. ' Leptidianae. The first branch of radius springs from a 



point beyond the retreating median cell. Radius five-branched. 



Vein IV2 central on forewings, inclining to cubitus on hindwings. 



Type peculiar, and relatively generalised by the five-branched radius 



and position of I V2, 



Fam. IV. Nymphalidac. Vein III4 of forewings thrown off upon external 



margin below, exceptionally at apices. Radius generalised, five-branched. 



Sub-fam. 1. Nym])halinac, Radius and subcostal vein on hindwings 



fused to a point opposite the issuance of the abbreviated vein I ^. 



To this group belong the Limenitid genera, such as : Nijmphalis, 



Naias, Limcnitis, Basilarchia, etc. 



Sub-fam. 2. Argijnninne. Radius and subcostal. on hindwings fused to 



a varying point, but always before that of issuance of vein I. This 



group contains the mass of forms, the tribes into which it may be 



resolved are not usually definable taxonomieally by distinctive 



characters of venation. The smaller groups vary in resj^ect to the 



degi-ee of specialisation in the one direction of the disintegration of 



the media. The 'Emperor 'butterflies appear to be most specialised; 



the ' Fritillaries ' are most generalised in the holarctic fauna. Here, 



among these latter, vein III4 seems exceptionally to attain apices 



from below in certain exotic genera. 



Fam, V. Agapeticlac {=Sati/)'idae])reocc.). Vein III4 of forewings thrown off' 



to apex of wing ; all traces of Vein VIII appear to have completely 



vanished. Radius generalised, five-branched. 



Sub-fam. 1. Pararginac. Vein IV3 of hindwings joins cubitus, thus 

 agreeing, by convergence, with the position of this vein in the 

 Nymjihalidae. 

 Sub-fam. 2. Agapctinac. Vein IV3 of hindwings joins cross-vein, as in 

 the following family and in the Whites (Pierinae). 

 Fam. VI. Heliconidae. Vein VIII absent on forewings ; II and III on hind- 

 wings hardly absorbed at base. 

 Fam. VII. Limnadidae {Danaidae). Vein III4 of forewings thrown oft' to 

 apex of wing. Vein VIII present. Radius generalised, five-branched. 

 II and III of hindwings absorbed as in Nymphalinae. 

 Fam. VIII. Lihytheidac. Vein III4 of primaries thrown off to costa, just 

 before apex of wing. Vein VIII ]>resent. Vein IVs of hindwings joins 

 cubitas, as in the Nymphalidae and Pararginae, not as in the Limnadidae. 

 Radius, five-branched. 

 Fam. IX, Nemcohiidae. Vein IVi on hindwings absorbed by the radius as in 

 Pieridae. In preceding three families this vein does not leave the upper 

 and outer angle of the median cell. Radius, five-branched. - 

 [Group Lycaeni-Hesperiadae, characterised by the equal spacing between the 

 veins, and by the peculiar progress in disintegration of media, of which 

 the branches remain in situ, leading generally to theii' obliteration 

 (especially of IV2 on secondaries) on disintegi-ation of cross-veins, which 



^ Prof. Comstock also gives this comparative character for the ' Purples ' in the 

 Manual, 406, a work with which I was unacquainted when •writing the "Tagfalter von 

 Hildesheim." Comstock follows (I.e. 410) with Anaea ; but, alone on this character, the 

 Goatweed butterflies are clearly more generalised forms. The latter are also much more 

 generalised than Potamis iris, and probably have no proper claim to the fantastic title 

 of ' Emperors. ' 



2 The immature stages of NcTneohius seem to show the Lycaenid character, while the 

 wings of the butterfly appear to be developed upon the Pieri-Nymphalid pattern 

 involving the movement of the branches of the media. It might be thus a survivor of 

 a separate line connecting with the ancestral stem of both groups, since it shows 

 features of each. In any case, whether the pattern of the wing is here primary or not, 

 Nemeohixcs represents a distinct family type, which is my original contention. It can- 

 not weU be derived from the Lycaenid type, since in this the radius is specialised, while 

 in Nemeobius the five-branched condition is retained. 



