96 NATURAL SCIENCE [February 



latter serve as their point d'ajypui. In rare instances, such as Augiades 

 sylvanus, and perhaps the Megathymidae, IV2, while retaining its central 

 position, has maintained on forewings its connection at base with cubitus, 

 while disappearing, from want of this, on hindwings.] 

 Fam. X. liiodinidae { = Lcmoniad%e, Erycinidac, preocc). Vein I of hind- 

 wings present, not absorbed. Radius specialised, four-branched. Specialisa- 

 tion of the type, Riodiim lijsippus, is comparable with tliat of the 

 Zephyrini and Chrysophanini, tribes of the Theclinae, from which it may 

 prove ultimately difficult to separate the Riodinidae by the neuration 

 alone. 

 Fam. XI. Lycacnidae. Radius specialised, three to four branched ; vein 

 III5 reaches apex. Vein I of hindwings absorbed. 



Sub-fam. 1. Theclinae. Vein IVi of forewings fuses with radius. 



To this belong the Theclini, Zephyrini, Chrysophanini. 

 Sub-fam. 2. Lycaeninac. Vein IVi of forewings not fused with radius. 

 Comprises the typical ' Blues.' 

 Fam. XII. Megathymidae. [Neuration not examined.] 



Fam. XIII. Hcspcriadae. Radius generalised, five-branched, all the veins 

 separate. Vein III5 reaches external margin below apex. On hind- 

 wings vein I absorbed. 

 Sub-fam. 1. Pamjjhilinae. Vein I of forewings absent. 

 Sub-fam. 2. Hesperianac. Vein I of forewings present (a ' thickening 

 of the costal edge '). 



The neuration, as thus explained, gives the following results : — 

 The Parnassi-Papilionidae are separated from all other day butter- 

 flies by the presence of vein IX (the short vein on internal margin) 

 on primary, and by the absence of vein VIII on secondary wings. 

 Pariiassius may be aptly compared with Attacid genera, the most 

 specialised among the moths. The brush-footed butterflies resemble 

 the Pieridae, and, except by the radial specialisation in the latter 

 family are difficult to separate by neurational characters. In the 

 position of vein IV3 of the hindwings, the Nymphalidae, Parar- 

 ginae, and Libytheidae agree. A change in the same character 

 brings the Pieridae, Agapetidae, and Limnadidae together. The 

 Heliconidae also share this same change of character, and are 

 generalised butterflies as compared with the Nymphalidae proper. 

 Unless proof can be offered that the pattern of neuration has been 

 here secondarily acquired, we must class the Nemeobiidae with this 

 group of the Hesperiades, the Pieri-Nymphalidae. The Lycaeni- 

 Hesperiadae meet upon a common distinctive wing-pattern ; the 

 Lycaenidae differ in the main by the reduction of the radial 

 branches. The character separating the Theclinae and the 

 Lycaeninae lies in the degree of absorption of vein IV^ by the 

 Radius, the typical ' Blues' being the more generalised group. The 

 value I would give the Papilionides in classification is that of a 

 super-family. Whether Dr Dyar can show us any character in the 

 position of the larval tubercles to aid us here is not yet ascertained. 

 The Hesperiades divide into two groups, having a higher value than 

 that possessed by the modern ' family ' idea. I call them simply 

 the Pieri-Nymphalidae and Lycaeni-Hesperidae. Notwithstanding 



