36 MEMOIRS OF THE NATIONAL ACADEMY OF SCIENCES. 



whatever muj' be said of tlie .socondury armature; at least we have not observed any, certainly 

 none of taxonomic value, though this subject has not j-et been carefully or extensively examined. 

 The habits of nearly all lepidopterous larva- being the same, we should not expect any decided 

 ditierences except in the leaf miners and the more primitive forms, such as Eriocephala (Monog. 

 Bombycine Moths, I, p. «)0). 



The head in the Smerinthina' genei'a is conical, the vertex tending to be narrow and some- 

 what elevated, but in the higher Sjihingidaj the head becomes round. Now, the conical shape of 

 the head of the Smerinthinje seems to have become directly inherited from the conical shape 

 characteristic of the Ceratocampida?, especially Adelocephala, the most primitive genus f)f the 

 group. 



In tlie armature the position of the setifei'ous tubercle v directly in front of the spiracle is a 

 trenchant or ditierential character, and, as Dyar says, it is characteristic of the sphingid larvw. 

 We are able to contirm all he sa^ys as to this feature. As we have already stated (p. 81), 

 the freshly hatched larva of Oeratomia ajnyntar absolutely differs in this respect from anj' larviv 

 of the Saturniides. It should lie observed, however, that these primary tubercles disappear after 

 the tirst molt and that only the freshly hatched young of Ceratouiia (onyutor, and C. undnlom have 

 been examined. Provisionalh', however — namely, so far as our'present knowledge extends — I 

 quite agree with Doctor Dyar as to the significance and value of this grouji character. The posi- 

 tion of the tubercle in the Sphingida? is one apparently which was suddenly acquired, as if by a 

 leap, or "mutation, ".but as to the cause of the change of position, since in all the great groups 

 of bombyciue moths (Saturniides) it is universally situated below the spiracle, while tubercle 

 iv is moved up behind the spiracle — we are quite in the dark. 



The glandular setaj of Ceratomia are like those of certain Notodontid;^. 



This is a primitive character, not occurring in Ceratocampidie. in which the seta? are all 

 acute, nonglandular, or the seta? at time of birth are, with the tubercles, converted into large 

 spines. Another feature in smerinthine, but not the ' higher"' sphingid larvis, is the occur- 

 rence of crowded miiuite secoudary tubercles on the integument of the head and trunk, render 

 ing the skin rugose or shagreened. An approach to this is, however, seen in larval Adelocephala 

 and Anisota. Also sphingid larva; have no specialized lateral or submedian spines on the sur- 

 anal plate, or rough, coarse granulations on the edge of the anal legs. It will be remembered 

 that the head in the smerinthine larva? is subconical.. narrowing above, while in the more special- 

 ized groups the head is more rounded. This conical shape of the head seems to have been 

 directly inherited from tjie Ceratocampida?. 



A salient feature of the Sphinx larva is the caudal horn; as has been shown by a number 

 of entomologists, and, as we have repeatedly observed, it differs in no respect in its general 

 shape and originally double origin fi-om that of the Ceratocampidie, and seems to be, like the 

 other characters here mentioned, an heirloom from the Ceratocampidie. It appears, then, that, 

 with the exception of the position of the tubercles iv and v, the larval characters of Sphingida^ are 

 such as indicate the direct descent by divergent and saltatory evolution of the group from some 

 primitive ceratocampine form, like Adelocephala. 



In the ver3' interesting larva of the South African sphingid, Lophostefhus dm/ui/hu'I, we have 

 the unique occurrence of a larva beset with a complete armature of long, rathei' stout, chitinous 

 spines, those of the dorsal and subdorsal series about or uearl3- as long as the median caudal 

 horn. Doctor Dyar " has thoroughly discussed the armature of this larva and shows that while 

 it has a remarkaV)ly ceratocampid-like appearance the tubercle r is situated directly in front of 

 the spiracle, a characteristic of its position in those larvic of the sphingids y(>t known. There 

 ai-e no spines on the tirst thoracic segment; tubercle // is not represented. The larva, he says, 

 "is a true Sphinx, not moi'e nearly related to the Ceratocampids than any other Sjihinx, since it 

 possesses true sphingid tuliercles, ii: above r and before the spiracle, not united with /■ as in all 

 the Saturnian phylum. Functionally, indeed, it is a Saturnian like the African Saturnians, 

 with thorn-like tubercles; but the character is evidently adaptational, an irregular hypertrophy 



«Proc. Eut. Soc, Washingtou, iv, May 13, 1901, p. 440. 



