106 BRITISH LEPIDOPTERA. 



spiracular, and iv becomes a post-spiracular tubercle. We do not think 

 the pre-spiracular tubercle (which is more or less adventitious) of 

 much value in classification, but the two above characters appear to 

 be so. 



Now it is evident from the above brief summary, that the structure 

 of the larval prolegs, the characters offered by the movable pupal seg- 

 ments, the broad characters of neuration, and of the jugum, only help 

 us to separate, as it were, the generalised from the more specialised 

 super-families. These characters still leave them unsorted, and give 

 us no clue to their relationship to each other. 



We are not alone in our objection to the division of the Lepidop- 

 tera into the two sub-orders, JUGATE and FRENAT/E, as proposed by 

 Comstock. Packard considers that the characters used are too slight, 

 and do not agree with the more fundamental pupal characters, or with 

 important imaginal features. He says : " The jugum is of slight, if of 

 any, functional value, and, in Micropteryx (i.e., Eriocrania),&s in Trich- 

 optera, occurs both in the hind- and front-wings, a point apparently over- 

 looked by Comstock. The Hepialids are much less generalised forms 

 than the Eriocephalids (i.e., Micropterygids), or even the Mieropterygids 

 (i.e., Eriocraniids) ; the pupae of both these groups have free limbs and 

 abdominal segments, belonging to what Speyer calls a group of 

 ' Pupa libera.' The Hepialidae, also, possess neither maxillary 

 palpi nor vestigial mandibles ; they are borers in the larval state, 

 and the pupa has not free limbs, but is a ' Pupa incompleta.' 

 They are scarcely ancestral, though very primitive, forms, but have 

 already become modified, having no traces of mandibles and no maxillae, 

 and, in the American species, the labial palpi have already begun to 

 degenerate. We, therefore, scarcely see good reason for placing the 

 family at the very foot of the order, below Microptery.r (i.e., Eriocrania), 

 but should regard the family as a side branch of the Palaso-lepidoptera, 

 which, very soon after the appearance of the order, became somewhat 

 specialised. Comstock's FRENAT^E comprise a heterogeneous collec- 

 tion of families, some of which have no frenulum at all, and, when the 

 frenulum is present, they offer secondary sexual characters. The ab- 

 sence or presence of a frenulum is hardly, then, a sufficiently funda- 

 mental character to be used in establishing a great primary division. 

 Besides this, there is a rather close alliance between the Hepialidae 

 and Cossidae, the latter having a rudimentary frenulum. Chapman 

 remarks, that while Cossus and Hepialus are quite distinct in pupal 

 characters, there appear to exist in Australia many forms uniting 

 them with Zeiizera into one family. The neuration is also quite similar, 

 and while the two families of Cossidae and Hepialidae are, in some most 

 important respects, quite far apart, one being, so to speak, Tineid, and 

 the other Tortricid, in structure, yet it would, we think, be a forced and 

 unsound taxonomy to assign them to different sub-orders " (Bombycine 

 Moths of America, p. 57). 



We have the same objection to Packard's own primary sub-division 

 of the Lepidoptera into two sub-orders : I. The LEPIDOPTERA-LACINIATA 

 or PROTO-LEPIDOPTERA [comprising only the Micropterygids (i.e., Erio- 

 cephalids)] . II. The LEPIDOPTERA-HAUSTELLATA sub-divided into : 

 1. Palaeo-lepidoptera [comprising only the Eriocraniids (i.e., Microptery- 

 gids)] . 2. Neolepidoptera (comprising the whole of the Incompletae 

 and Obtectae). Such a sub-division, in spite of the elaborate 



