CLASSIFICATION OF LEPIDOPTERA. Ill 



The Micropterygids and Nepticulids, therefore, will be considered 

 in connection with the Sphingo-Micropterygid group, although, as we 

 have just shown, they might almost as well have been treated with all 

 the lower superfamilies together. It is, however, necessary to break 

 down the idea that there is a great hiatus between the lower (Incom- 

 plete) forms, and the higher (Obtect) forms on the same stirps, or a 

 similar hiatus between the smaller species and the larger. For the same 

 reason, although we place the Eriocraniids at the base of the Geometro- 

 Eriocraniid stirps, they might equally well be placed almost at the 

 bottom of all Lepidoptera. The Hepialids, owing to their relationship 

 with the Zeuzerids and Cossids, are placed at the bottom of the Noctuo- 

 Hepialid stirps. 



There are one or two points in which we think the sum total of 

 characters shows that Dyar and Hampson have erred in their group- 

 ings. Thus the PSYCHIDES belong rather to the Sphingo-Micropterygid 

 than to the Geometro-Eriocraniid stirps. The LASIOCAMPIDES most dis- 

 tinctly belong to Dyar's BOMBYCINA, i.e., to our Sphingo-Micropterygid 

 stirps, whilst the NOTODONTIDES as certainly belong to the Noctuo- 

 Hepialid stirps. We have already shown that Dyar's NOCTUINA is 

 heterogeneous on the egg characters, and the two main branches 

 included must be differentiated on other larval characters than those 

 already used, so as to separate the true Noctuids from the Geometrids. 

 We observe that Hampson finds neurational characters to make this 

 separation. 



We are totally unable to accept the conclusions reached in Dyar's 

 " Synopsis of the Families of Bombycides " (Proc. Bos. Soc. of Nat. 

 History, vol. xxvii., pp. 129-130) as being of any real classificatory 

 value. In this we find the Notodontidae, Pseudoipsidae, Noctuidae, 

 AfHttclidae, Arctiidae, Pericopidae, Nolidae, Litlto.nidae, Euchromiidae, 

 Ltjinantriidae. (of the Noctuo-Hepialid stirps), Brephidae, Geowetridae, 

 T/tyatiridac&nd Drepanidae (of the Geometro-Eriocraniid stirps), united 

 with the Bombyddae, Eupterotidae and Lasiocampidae (of the Sphingo- 

 Micropterygid stirps), to make up the " Higher Bombyces." To 

 explain away the position of the Eupterotids, in which Dyar recognises 

 that tubercles iv and v do not satisfy his definition of the group, we 

 learn that " warts iv andv are degenerate, and have come to assume a 

 generalised position, probably secondarily " (p. 128). Concerning 

 the Bombycidac, Dyar states, what is a fact, that " the warts are 

 small and degenerate, but they are true warts of the typical Lasio- 

 campid pattern" (p. 140). Kegarding the warts of the Lasiocampids, 

 we read that "on the abdominal segments (of Tolype velleda), v is 

 smaller than iv, and all except i and vi are greatly reduced. These, 

 two warts alone persist in the adult " (p. 144). The reduction of iv 

 is quite characteristic of the higher branches of the Lasiocampids, 

 Bombycids, Eupterotids, etc., and entirely different from the well- 

 defined post-spiracular iv of the Noctuids, Arctiids, etc. 



In its broadest lines, then, our scheme of classification assumes 

 three main general evolutionary branches, along which the various 

 superfamilies of Lepidoptera have developed, two of these being flat- 

 'egged and one an upright-egged stirps. These, with the main super- 

 families included in each, have already been given (ante, p. 109), so 



that there is no need to repeat them. 



* Annals and Magazine of Natural History, Oct., 1894, p. 259, 



