MEMOIRS OF THE NATIONAL ACADEMY OF SCIENCES. 87 



of the tuberclps suporimposed on the phylogenetic chanicter.s of the Sphingidte." The style of 

 coloration, he says, is not sphingid but ceratoeampid. '' The head and cervical shield are con- 

 spicuously striped with black; the anal plates ai'e red. with black borders; the bod^- is green; the 

 spines black, with yellow bases; the foot shields black. A white or yellow bar extends between 

 the second and third spines on the first to seventh abdominal segments." 



We would note the fact that Fawcett places this hawk moth in the subfamily Smerinthinsi?, 

 the most primitive group of the Sphingidse. The pupa has short maxilli*. and a large, short, 

 rounded cremaster. 



It is also interesting to note that the armature is very much like that so prevalent in the 

 South African subfamily Bun<eina\ represented l)y Gynanisa. Nudaurelia, Bunsea, etc., and tliat 

 the caudal horn is not like that of other Sphingidfe, especially the Smerinthina?. in being thick, 

 fleshy, and tuberculated. and more or less flexible, but is solid, stifl', chitinous, like the other 

 spines on the body. In this feature do we not see the efl'ects of the dry, peculiar climate of 

 Africa, where there are so many spiny plants and trees? The spines may have arisen after the 

 ancestors of Lophostethus had established themselves on the African continent. It should be 

 borne in mind that the South American continent (Neogala) is apparently the center of origin of 

 the Sphingidie. The same or similar climatic conditions may have influenced the coloration of 

 this larva. 



The i^upa. — When we compare the pupa of Paonuift e,rc»^c(iti(s S with that of Enchs 

 imperialis S there will be found to be no salient or diagnostic dift'erences, such as we 

 would expect, to separate the pupsv of two great families. The shape of the body is nearly 

 identical; the head of Paonias is slightly more conical in front, not so much rounded; but the 

 surface of the integument is covered with fine spinules. The antenna; are the same in width and 

 in the raised joints and pectinations. The maxilla? are of the same shape and length, but wider 

 at base than in Eacles; the eyes, epicranium. and clypeus are the same, their surfaces similarlv 

 though less rugose, but without any .specialized spines. There are in either form no traces of 

 primitive characters such as occur in the more primitive lepidopterous families. 



The cremaster in Paonias is large and ends in an undivided spine, not forked as in Eacles 

 and other ceratoeampid pupse, with the exception of Citln-ronla rcgulis, in which it is vestigial, 

 and shows signs of an original division. Also the segments of the abdomen are smooth, and 

 segments IX and X ai-e complete in Paonias, the sutures not interrupted and obsolete between the 

 scar of the genital opening and tliat of the vent. This may prove to be a familv or diagnostic 

 character. (PI. LVI, tig. ,s,/.) 



It is to be observed, then, that the pupa? of the Smerinthina? are generalized, and in their 

 head-characters, those which are most fundamental, agree with those of the Ceratocampidse, 

 while in the more variable shape of the terminal alxlominal segments and the cremaster there is 

 a departure from the ceratoeampid shape. 



As we ascend the sphingid series and reach Plilegetbontius with its enormous tongue case, 

 forming a partly free structui-e, we have a feature peculiar to the Sphingida?, but as is well 

 known the maxiUse, even in the more specialized Sphingida^ are exposed to great vaiiation, and 

 they may be in the pupa buried between the fore legs on the breast, or if large, form a salient 

 prolongation of the front of the head, as in Chwrocampa. 



As regards the hal)its of the pupa the Sphingida^ iiave retained tiie subterranean mode of 

 life of their ceratoeampid ancestors, in no case known to us .spinning a cocoon or lining their 

 subterranean quarters with silk, unless in sporadic cases a few silk threads are spun. 



The iiiiago. — There are in Sphingidte eleven veins in the fore wings and nine in the hind 

 wings. The most striking and diagnostic character separating the two groups of Sphingida' and 

 Ceratocampida? is the presence in the former group of radius 2 (111.,), which ari.ses within the 

 middle of the wing before the end of the discal cell. By the addition of this vein the wing is 

 greatly strengthened on the costal border, which receives tlie force of the blow during the move- 

 ments of the wings in flight. This vein is alxsent in all the genera of Ceratocampina?, but it is 

 generally present in the subfamily of Bunainsi?. where, however, it is a very short, weak vein 

 develoi^ed near the apex of the wing. It is absent in Xudaurelia cyiherea and \estigial in Salassa, 



