MEMOIRS OF THE NATIOXAL ACADEMY OF SCIENCES. 29 



tlie larval forms, esi)ecially of tlie. earliest stages, it is not difficult to coustruct a geuealogicai 

 tree, of tUc subfamilies Ileterocampiiia- aud Cerurina-. Wbeii taking into account llie larval stages 

 of the entile family, even with our present imperfect knowledge, it is easj^ to see that Datana 

 stands at the base, is the more generalized primitive form, and was perhaps the first to diverge 

 from the stem form of the family. 



The first author to call attention and at the same time to treat in a philosophic way of what 

 he has called "the incongruence of form relationship, between larviBon the one hand and imagines 

 on the other" is TVeismann, in his well-known work entitled Studies in the Theory of Descent. In 

 Chapter II of the second volume, entitled -'Does the form relationship of the larva coincide with 

 that of the imago?" he points out certain incongruences between the larval and adult characters. 

 He claims that "neither the group of Microlepidoptera nor those of the NocUiina, Bovihycina, 

 SphiiKjina, and RhopaJocera can be based .systematically on larval characters," adding the quali- 

 fication, "Several of these groups are indeed but indistinctl,y defined, and even the imagines 

 present no connnon characteristics by which the group can be sharply distinguished." Within the 

 families, howev(!r, he states: "There there can be no doubt that in an overwhelmingly large 

 majority of cases the phyletic development has inoceeded with very close parallelism in both 

 stages; larval and imaginal families agree almost completely. On the other hand, "in the butter- 

 flies a perfect congruence of form relationship does not exist, inasmuch as the imagines constitute 

 one large group of the higher order, whilst the larva' can onlj- be formed into families." But in 

 this case Weismann does not seem to be aware that the imaginal lihoijalocera as such is ijuite an 

 artificial group, and that the imaginal families recognized by Bates, Scudder, and others have 

 perhaps more equivalent, congruent, or nondivergent larval forms than his remarks would seem to 

 imply. 



But without attempting to enter into an exposition or criticism of Weismaun's general 

 statements, his whole discussion being most suggestive and stimulating, we will turn to what he 

 says of the Notodoutidic: 



Au especially striking case ol' iucou^nieuce is ofteretl Ijy tbe family Xofodonlidir, iiuder wliicli Boisduval, 

 (lepemlinj; only on imaginal characters, united genera of which thu larvie differed to a very great extent. » ♦ » 

 In fact, in the whole order Lepidoptera there can scarcely be fonud associated together such diverse larvic as are here 

 placed in one imago family. 



lie then refers to the short cylindrical caterpillars of CnetJiocnmim, which, however is not a 

 Notodontian, but a Lasiocampid. lie then briefly refers to the larva' of llarpyia (Cerura) and the 

 caterpillars of Stauropus, Ilybocampa, and Notodonta. Without g'ving further attention to the 

 family, he returns to the butterflies. This family, then, presenting "an especially striking case of 

 incongruence," we will briefly discuss, referring the reader for fuller details to the figures on the 

 plates. 



In the first place, as a matter of fact, the more one liecomes familiar with the Lepidoptera 

 and their larval forms the easier it is to distinguish the larvie by their "family" characteristics, 

 premising, however, that the term family is of very uncertain meaning, and that ditterent 

 authors differ as to what to call a family as much as they do what to designate a species. But 

 no one, we think, need to err in correctly picking out or identifying any Bombycine larva except, 

 jierhaiis, a few Notodont larva', which are liable to be confounded with certain Thyatirid;e, and 

 the hairy x^octuid:e, but even then a careful examination will show family differences even when 

 adaptation and modification have nearly bridged over the fundamental differential characters. 



In this work I have divided the family into seven groups, which may be for convenience 

 regarded as so many subfamilies. 1 was first led to do so by the larval characters alone, but 

 found that this classification Avould also apply in general to the moths, so that there proved not 

 to be so much incongruity as was expected. There appear to be, then, seven larval subfamilies 

 and seven imaginal subfamilies. Others may not agree with this view, but it is the most rational 

 classification I have been able to make. 



Beginning with the most simple forms of larva, those of the Glt(2>hi>iiiuv, which, both as 

 regards those of the Old and New World, are tolerably constant, the adults certainly difler notably 

 from those of other subfamilies, as also do the larva' and pupte. 



