Packard ] ^^^ [March 17, 



impcrialis in its third stage, a lliiiig Mr. Poulton could not do for want of 

 specimens. The resemblance between the two genera at this stage is 

 most striking, although the fully fed larvae are so different, Aglia passing 

 at a single molt (the third and last, this larva only hav'mg four stages), 

 from one family to another ! We know of no parallel case, or at least of 

 one so very striking and conclusive.* Thus the ontogenetic development 

 of this caterpillar epitomizes that of two families, whereas that of most 

 Bombycesis simply usually only an epitome of that of a subdivision of a 

 family, or of a small group of genera. 



AgF,a tau in its third stage differs from Eades imperialis in its third 

 stage in having a pair of dorsal "horns" on the first and third thoracic seg- 

 ments, where E. imperialis has only minute ones on the prothoracic seg- 

 ment, while those on the second thoracic segment are as well developed as 

 those on the third segment ; those on the second segment are minute ; all 

 the " horns " are forked as in Eacles. The dorsal spines on the abdomi- 

 nal are simple and minute, like those on the second thoracic segment. 

 The shape of the head and of the anal legs is much as in Eacles, but the 

 suranal plate differs strikingly in being produced into a rather large, 

 spinulated spine, a feature not known to exist in any Ceratocampids. 



It should be observed in regard to the large size of the prothoracic horns 

 of Agiia, that those of Citheronia regalis are quite well developed, being 

 about two-thirds as long as those on the two succeeding segments. 



Upon examining the adult of Aglia, I find that its head and antenn;x? 

 are closely similar to those of Ilyperchiria io, and the Hemileucidas in 

 general ; the antennoe form a close approach to those of H. io, as on care- 

 ful examination wiih a good lens a second branch of the pectinations of 

 the male antennae can be perceived ; it forms a long, separate branch, but 

 is in the dead and dry specimens very closely appressed to the anterior 

 main pectination. In the venation of both wings Aglia shows a most un- 

 expected resemblance to that of Eacles imperialis ; like that and other 

 Ceratocampidae and the Hemileucid;e, having five subcostal branches, 

 while in the Saturniidae there are only four, the first one wanting in the 

 latter family. 



Thus the moth belongs with the Ceratocampidae, while the larva after 

 the last molt loses all its spines and becomes very much like a Saturnian, 

 perhaps of the type of Telea, though it is without tubercles or spines, and 

 especially like a smootli lorui, the larva of Attacus betes Walker, figured by 

 Burmeister in his Atlas of the Lepidoptera of the Argentine Republic. 

 We therefore suggest that Agli'i tau should be regarded as the type of 

 a distinct subfamily of Ceratocampidae, and thus the latter group may be 

 divided into the two subfamilies, Ceratocampina? and Aglilnae. 



* Over twenty years ago, in 1803, when first beginning ray studies on tlie Bombyces, 

 ray attention was attracted to tlie singular changes of Ar/ha tau and I compared the 

 younglarva to the full-grown larva of Citheronia regalis and pointed out that the latter 

 genus was an " embryonic form and therefore inferior in rank to the Tau motli " {Ainer. 

 ^'aiuralist, June, 1870, and Our Common Insects, 52). 



