58 MEMOIRS OF THE NATIONAL ACADEMY OF SCIENCES. 



on the eighth iibdominal segment, due to the fusion in embrvonic life of tubercles /,• this horn 

 being long, acute, not a knob or rounded tubercle as in Saturniida>. 



Pupa. — The head-end of the bodj' unusually full and rounded, overhanging the mouth 

 region; the sutures between the abdominal segments deeper than usual; no primitive characters 

 on the under side of the head, i. e. , no traces of maxillary palpi and of second maxillw, but the 

 first maxillffi either long or short; in the latter case the two maxillse together form a nearly 

 equilateral triangle (Adelocephala and Anisota); cremaster usually very long and stout, forked, 

 though sometimes (Citheronia) nearly atrophied: groups and lines of line sharp spines on 

 the head and thorax, or the body quite smooth and unarmed (Cithoi-onia). The pupa is 

 subterranean, the larva spinning no cocoon. 



Geographical dhtrihat'ion. — The species are mostly antarctogteic, i. e., confined to South and 

 Central America,_ from which a few species have migrated into America north of Mexico, and 

 to the African continent south of the Sahara {Afrofinea). The only European form is Aylla taa, 

 while Salassa of the oriental region belongs to this famil\-. 



Coiaparison with the Satiirniidpe. — The Ceratocampidre as here considered is a much larger 

 group than that to which Harris gave this name. As a large proportion, the Ceratocampinie 

 excepted, have hitherto been associat^ed with the Saturniidw, we may here recapitulate our 

 reasons for I'emoving such a large body of genera from that group, and for establishing a new 

 family or group for them. Oui- attention was first led to this conclusion by the great and 

 apparent differences from true Saturniidre in the shape and structure of what larvae of the 

 African genera were known, and their transformations. 



But we will first call attention to the imaginal characters. In the first place the Ceratocam- 

 pinfe are sufficiently distinct. In the Buneinaj we have the greatest approximation to the 

 Saturniidie. Our chief guide and reliance here has been the venation of the two groups. This 

 is seen in the number and position of the veins, and the form, size, and completeness of the 

 discal cell. The normal numl)er of veins in the fore wings of CeratoicampinK is eleven, though 

 the second branch of the I'adial vein (vein IIIj) is sometimes absent, and very rarelj^ the vein 

 is undivided, onl} vein III3 being present. In the hind wings the usual normal number of veins 

 is eight. 



On the other hand, the normal number of veins of the fore wings in the Saturniid.e is eleven, 

 but in Rothschildia, Samia, Telea, Anthenea, Rhodia, and a few other genera there are but ten, 

 while in Graellsia, Perisomena and Caligula, japoniea and aiinla there are but nine. This differ- 

 ence, however, is not significant, as it is due to the presence or absence of III, or III.,, or both, 

 these being short, unimportant veins. The usual, indeed so far as we know invariable, number 

 of veins in the hind wings is eight. It thus appears that the normal number of veins is the same 

 in the two families, the group Ceratocampinte. which have nine veins in the hind wing, excepted. 



Subfamily 1. CERATOC^MPIISr.^ G-rote. 



Ceratocamjiidu' Harris, Report Insects inj. Veg. Mass., p. 287. 1841. 



Ceratocampcid:r Packard, Proc. Ent. Soc. Phil., iii, p. 381. November, 1864. 



Ceralocampinve Grote, List Lep. N. Amer., p. xiii. September, 1868. 



Ceratocampada:- Grote, New Check List N. Amer. Moths, p. 20. May, 1882. 



Ceralocampinie Smith, Entomologica Amer., ii, p. 20. April, 1886. p. 44. ]May, 1886. 



Ceratocampidw Kirby, Syn. Cat. Lep. Het., i, p. 739, 1892. 



Cfratocampin.-e Packard, Journ. N. Y. Ent. Soc, i, p. 7. March, 1893. 



Ceraloaiinpidn' Neoioegen and Dyar, .lourn. N. Y. Ent. Soc, ii, p. 147. December, 1894. 



CilheronUd.it Dyar, Journ. N. Y. Ent. Soc. ii, p. 174. December, 1894. 



Ceratocampidic Dyar, List N. Amer. Lepidoptera, p. 75, 1902. 



Snhfaiiiily characters. — Head triangular, wide at the insertion of the antenna? and narrowing 

 toward the oral region more rapidly than usual, coming to a point (especially so in Adelocephala, 

 S^'ssphinx, and Eacles; considerably wider toward the oral region in Anisota and Citheronia). 

 When denuded the front is seen to be more regularly triangular than in any other group of the 

 family, or in the Hemileucida; or Saturniida>. 



Antennae of S bipectinate on the basal two-thirds or three-fourths, beyond filiform, the distal 



