792 The American Naturalist, [September 
The two pieces (l p) at the base of the maxilla may possibly 
prove to be the labial palpi, if so, is the piece marked / the la- 
bium? The two paraclypeal pieces or tubercles (p. ) appear 
to be the homologue of those in the Psychide. 
The pupe of this family are very extraordinary, but it will 
be seen that they are Pups incomplete, and prove that the 
Fre. 9. 
family should stand much above the Micropterygide, rather 
than below them, so far as regards pupal characters. 
Fig. 9 shows the front of the head and maxille of the Cos- 
sid, Prionoxystus robiniæ, which is more Tortricid than Hepi- 
alid ; pe, paraclypeal piece; mx. p, maxillary palpi; l, labial 
palpi; mz, maxille. 
The very primitive, generalized shape of the thorax of the 
Hepialide is noteworthy. In Hepialus mustelinus the collar or 
prothorax is very much reduced; while in H. tacomz it is very 
long and generalized, as in Sthenopis and the Australian 
Abantiades argenteus. The mesoscutum is considerably shorter 
than in H. tacome. In the latter species the metascutum is 
entirely divided by the large scutellum, while in H. mustelinus 
it is only partly divided, the apex of the scutellum passing a 
little beyond the middle of the scutum. 
It is thus quite evident that Sthenopis is an earlier form 
than H. tacomex, and that the latter is more generalized, hav- 
ing undergone less modification than H. mustelinus. 
