MEMOIRS OF THE NATIONAL ACADEMY OF SCIENCES. 33 



It appears from Lt. Col. .1. A\'. Fawrctt's (l('scri|iti(iii ami liji'iirc of the larva of I'rotoparce 

 maiiritil Butler, of South Africa, that there are '•paired humps on first and second somites." 

 Judoiiio- hy his lio-ure. these humps ar(> h)wer, more rounded, hut ai'e situated on the front edge 

 of each seuuicut. as in Ceratomia. We thus have in two ([uite ditl'erent genera of Sphinginaj this 

 singuhir mimicry of the thoracic tubercles of Ceratocampina". 



Tin- ^lirxt lari-al stage of ('■'ri(t(>ini(i undulom. — (PI. XXXIV, tig. i>.) Stiige I. The larva is 

 long, slender, cylindrical, but widely diflers from that of C. innyntor in the same period of life 

 in being destitute of the four boss-like rudimentary horns, the segments lieing in this respect 

 normal. The hinder or fourth pair of mid-abdominal legs are, judging by Mr. .Toutel's drawing, 

 even smaller than the third pair, which are much larger than those in front. The body is of an 

 uniform pale whitish flesh color, with no markings. 



Incongrut-nce in, the (jeniiK Ceratomid. — Ceruhniiui ai«lid(i»i seems to f)e geneiically distinct 

 from O. amyntor in stage I, both as regards the absence of the rudiments of dorsal false horns 

 and in the mid-abdominal legs being smaller than the others, or at least no lai'gor than those of 

 the second pair. 



The larva of 6'. catalpx in all its stiiges, as descril)ed and figured by Kiley, is also entirely 

 unlike C. cunyutnr, being, after the first molt, smooth l)odied, while its markings are very 

 different, the body being without any oblique whitish lines, and in the two last stages the skin is 

 "smooth and velvety." In the second stage the head is smooth and polished, with no traces 

 whatever of thoracic false horns (Kep. U. S. Entomologist, 1S82, p. IS'.t). 



The larva of C. hae/eni resembles, according to Rilej', that of C. umhilotta; it is marked with 

 oblique vellowish-green lines. The incongruence between the larva? of the species now referred 

 to Ceratomia, is striking. It is to be observed that Riley refused to associate either 

 C. catedpm or C. liagenl with C. amyntor, l:)ut referred them to Sphinx. Certainlj' C. amyntor 

 stands alone, and the pupa3 and imagoes of the genus Ceratomia as now accepted should be 

 revised. 



VIII. PHYLOGENY OF THE CERATOCAMPIN^E. 



The most primitive, generalized genus of the group is Adclocephala. unless it should be 

 found that Astylis hellatrix is still more so. Unfortunately the larva and iiuago of this form are 

 not obtainable. 



That Adelocephala is the stem-form from which the other genera have originated is 

 suggested by the larval armature, the presence in the undetermined Brazilian larva (PI. XLIX, 

 fig. -f) of quite well developed prothjracic horns, and by the equality in length and shape of the 

 dorsal horns both of the thoracic and abdominal segments. 



.Already in Syssphinx the armature has undergone a very considerable reduction, showing 

 that it has diverged from the main line of descent. 



A decidedly remote side branch, with no annectant form, is Anisota, which notably diflers 

 from all the rest of the group in the reversion of the caudal horn to a pair of separate setiferous 

 tubercles. This would seem to be the result of a j>er .m/fiu/t retrograde mutation — a case of 

 reversional evolution: also the decided reduction in length of all the thoracic horns, except those 

 of the second thoracic segment, is a case of discontiiuious evolution l)y partial atrophy. Corre- 

 lated with these modifications are certain differences in habits in the case of the species of 

 Anisota. which are gregarious and seem to l)e avoided by fiirds, judging by their feeding in 

 exposed situations and by the great luunber of individuals. 



Returning to the main evolutional path of the group, we observe that Eacles is not remotely 

 disconnected from Adelocephala, although in this genus toward the end of larval life, in contrast 

 with the exuberant growth of horns in the earliest stage, there is a reduction in the length of 

 the spines. The pupa is of the type of that of Adelocephala, and in the imago there is not a 

 great difference. 



