282 BRITISH LEPIDOPTERA. 



This seems not at all improbable when we consider that specialisation 

 is so frequently accompanied by reduction in number of bones, 

 organs, etc., in vertebrates, and nervures, tubercles, free segments of 

 pupae, etc., among insects. Chelepteryx fcollesi), which is said to 

 have a tufted larvae, is probably an early branch from the Dimorphid 

 stock that has diverged widely in regard to its larva. So far as 

 I can determine from a superficial examination it is more specialised 

 as an imago than Dimorpha verskolora, judging by the fact that 

 the pattern of the upperside is not so exactly repeated on the under- 

 side as in D. verskolora, and also that the forewing pattern is not 

 so nearly repeated on the hindwings (Bacot, in lift.). 



With regard to the antennas of Dimorpha compared with those of 

 Lachneids, Sphingids and Chelepte?yx, Chapman writes : " There is 

 a close general resemblance between the antennae of I), verskolora and 

 those of the Lachneids, but nol more than between the antennae of 

 widely separated forms with pectinate antennae. The chief difference 

 is that, in the antennae of D. verskolora, the long hairs clothing the 

 ventral aspects of the plumules are irregularly distributed, whilst 

 in Lachneids they are in tolerably regular rows. Though a little 

 bent at the ends, the plumules have the final seta fairly in line with 

 the shaft of the plumule, whilst in most Lachneids the combined 

 bending and clubbing at the end of the plumule make the seta 

 apparently on the forward margin. Sphingids are too different to 

 make any comparison useful. Chelepteryx has antennae very similar 

 to those of D. verskolora as to scaling, plumules, and hairs ; and, 

 except that they are longer and diminish terminally more gradually, 

 and that the pectinations are comparatively much smaller in the 

 female, there is practically nothing definite to seize as a difference. In 

 connection with the antennae it may be noted that in both genera, 

 Chelepteryx (collesi) and Dimorpha {verskolora), the males have very 

 long we'i-developed spurs to the anterior tibiae, whilst in the female of 

 Chelepteryx the spur is short and slender, and in that of Dimorpha it barely 

 exists, being a mere minute knob at the point of origin. This is a 

 greater sexual difference in this structure than is at all usual." 



Packard finds the greatest affinities between Dimorpha and 

 the Saturniids, but notes (Proc. Amer. Phil. Soe., xxxi., p. 141) 

 that he is " disposed to think the family Endromidac ( Dimor- 

 phidae) is a natural one, and that it would be a violation of 

 the principles of classification to include Aglia with it. The two 

 genera, both as regards their larval and their adult characters, are 

 quite distinct." He finds that Endromis {Dimorpha) verskolora has 

 the head, palpi, antennae, and the hairy abdomen very closely like 

 those of Hemileuca maia, but the median nervure of both wings 

 divides into four branches, and the subcostal nervure of the four 

 wings divides into five branches, as in IP maia and the other 

 Hemileucidae. He concludes that " Endromidae is a branch of the 

 Bombycine tree, parallel to, but distinct from, the Hemileucidae, 

 and stands above the latter, connecting the group and the Cerato- 

 campidae (Cithero/iiidat) and the Saturniidae with the higher families of 

 the Bombyces, in which there are four branches of the median nervure, 

 all the families mentioned agreeing with the Notodontidae as having but 

 three. In the general shape, the small retractile head, the mode 

 of coloration, and the caudal horn, the larva of Endromis appears 



