42 MEMOIRS OF THE NATIONAL ACADEMY OF SCIENCES. 



The shape of the three last abdominal segiuents, with the anal legs, is eharactei-istic in the 

 Sj'mbombycina. The eighth segment may or may not be humped or bear a horn-like excrescence, 

 but the ninth and tenth segments are well developed, more or less elongated, especially the ninth. 

 In shape and armature these segments widely differ from those in the Syssphingina. 



The caudal hump or horn differs from that of the Syssphingina in being soft and fleshy, 

 usuall_v low conical: it is sporadic in distiibution, being either present or absent in the Lasiocam- 

 pida;; apparent!)' absent in the EiirypterotidiV. 



In Bomhyx viorl,'^ as already stated (p. 20), it arises from a median multisetose wart: present 

 in stage I it is a large conical hump, which is distinctly divided at the end, plainly showing 

 its origin from two separate warts. 



In Brahmiva japoniea in stage I it has lost all traces of its duplex nature owing to the high 

 degi-ee of specialization of the warts in this genus. 



The suranal plate is broad, thick, fleshy, either smooth or somewhat rugose (Endromis and 

 Bombycida^), or in Brahma^a armed with two horns. In Apatelodes the suranal plate is short, 

 very broad and fleshy, witii no armature, and so it is in the Liparida^ and Lasiocampida'. 



The anal legs differ widely from those of the syssphingine group of families in being soft, 

 fleshy, with no decided chitinous plate or granulations. 



The hairy larvse, especially those of the Liparida" and Lasiocampida?, with their conspicuous 

 pencils of hairs have their prototypes in the notodontian genus Apatelodes. In A. torrefacta 

 the tufts or pencils arise from minute ill -defined warts. The tuft on abdominal segment 7 arises 

 from two separate areas (not wai't-like eminences), one on each side of the median line of the 

 bodv: those on the eighth segment form a larger area or double group, bearing numerous micro- 

 scopic papilla? like those all over the l)ody. giving rise to the long secondary hairs, from which 

 the hair-like seta» arise. There are in the full-grown larva no warts, like those piesent in the 

 young before the first molt. (For the arrangement of the warts in stage I, with figures, see 

 Dyar in Psyche, December, 1895, p. 317.) The pupa varies in the group with the habits of the 

 different generic types. All the larvii? are spinners except the Brahmieida'. in which the pupa 

 is sul)ter]'anean and with a well-marked cremaster, somewhat reminding one of that of Ka-Jts 

 imperialh. The pupa differs, however, from that of Eacles in the head not being bent so 

 far forward, the thorax not l)eing so full and overhanging the head. In this respect the pupa 

 is like the imago, that of the Ceratocampida* having the thorax very full and rounded in front 

 and overhanging the head. The base of the maxilla? are also nuich nearer the head-end of the 

 body, while they are about twice as long as in Eacles. The large cremaster reminds one of that 

 of Eacles, ])ut this is evidently a case of convergence; it differs in being constricted at ))ase, 

 beyond much smaller, and ending in two diverging points. 



Pliijlogcny af the S[/iiihn/iJnjf!na. — The origin of the group may with a fair degree of cer- 

 tainty' be traced back to Ichthyura (Melalopha); at least that genus, especially' in the larval stage, 

 appears to be an ancestral type, prophetic of the incoming of more specialized families. It is 

 already as regards the multisetose warts, the secondarj' set», and two double humps and head- 

 characters quite far removed from the less specialized notodontians (Notodontinje) and approxi- 

 mated to the Synibombycina, these being points which I failed to see in preparing the monograph 

 of the group. 



For example, the head of Ichthyura in its general shape is very different from that of the 

 Notodontinffi and Heterocampina?, being more as in the Lasiocampinw. It is broad: the broad 

 and short epicranium is swollen on each side of the epicranial suture, which is nuu-h shorter than 

 in the other notodontians; the clypcus is large. (PI. XLIV. fig. 2.) 



The question ari.ses whether Ijecause of these features and the ancestral relation of the group, 

 the Apatelodina', Datanina», and Ichthyurina' should not l)e removed from the otIuT notodontians 

 and be regarded as collectively forming iui independent family. This may ultimately ha\e to be 

 done. But at present we may consider that the notodontians, originally derived from some 



«I have this spring (1904) raised numbers of B. moti from eggs kiniUy sent me l»y Dr. L. O. Howard, Ento- 

 mologist U. S. Department of Agriculture, affording me an abundance of living material for examination. 



