26 MEMOIRS OF THE NATIONAL ACADEMY OF SCIENCES. 



The ninth abdominal segment is unusnally well developed in the Attacida' and the Cerato- 

 ■campidii', sometimes, as has been previonsly stated, bearing a true " caudal horn," which takes 

 the idace of that usually growing on the eighth segment. In the Rhopalocera, the suranal plate 

 is in general, especially in Hesperidie and Papilionida>, small and rounded, much as in the 

 Noctuida-, but in the Xymphalida' it is more or less speci;ilized, and remarkably so in the larva of 

 Xeonymphd phoeion and other satyriues, whei'e it is greatly elongated and forked. (See figures 

 in Scuddcr's "Butterflies of New England;" also W. Miiller's figures of larva of Prepona.) 



The par anal lohen. — These are the homologues of the two anal valves (valrulw of Burmeister, 

 "the podical plates" of Huxley) observed in the cofkroach, and occurring in nearly all, if not all, 

 insects. In Geometrid larvie they are full, tlesliy, lobe-like, or papilliform, Ijoundiug the areas on 

 each side, and appear as if projecting backward from the base of the anal legs. 



In the Ceratocanipidie these paranal lobes are not well developed. In the larva of Cerura they 

 are much as in Geometrid caterpillars, where they end each in a seta. 



The paranal forks. — We have already called attention to these two bristles in our description 

 of the larva' of Cerura. (Proceedings Boston Soc. N. H. xsiv, p. .^53.) They are well developed, 

 arising from the end of a papilla in'ojecting directly backward. Their use has been indicated by 

 Mr. John Hellins,' who refers to a pair of sharp points underneath the anal flap, "which are used to 

 throw the pellets of frass to a distance." Occurring in Notodontian and other arboreal caterpillars, 

 notably the tree-inhabiting Geometrids, they are wanting in Noctuidie (including Acronycta and 

 Catocala), Sphingidic, and Rhopalocera, as well as the lower Geometrids and the Microlepidoptera, 

 and are not developed in the Sphingidre. In Ichthyura (Clostera) they are slightly developed. In 

 the European Urupteryx sambncatn (received from M. P. Chri^tien) these lobes are very large, 

 papilliform, and setiferous, and in our Chcerodes, etc., they are similarly developed and the 

 use of the two setfe or the fork is undoubtedly the same as in Cerura. 



The infraanal lobe. — My attention was first called to this lobe or flap while exanuning some 

 Geometrid larvfe. It is a thick, conical, fleshy lobe or flap, ending often in a hard cliitinous point, 

 and situated directly beneath the vent. In appearance it is somewhat like the egg-guide of the 

 Acrydii, though the latter is thin and flat. Its use is evidently to aid in tossing the pellets of 

 excrement away so as not to allow them to come in contact with the body. In a large not iden- 

 tified Geometrid worm, which lives on the ash, this flap is large and conical, ending in a blunt 

 chitinous point. In a large geometer belonging to another genus, the tip is sharper and harder, 

 and in what is probably a larva of Endropia, while the paranal forks are well develo])ed, the 

 infraanal lobe ends in a stift' bristle. Whether this infraanal lobe is the homologue of the ninth 

 nrosternite or ventral plate I will not at present undertake to say. 



Glfindnlar seta: — Among the Notodontida* the freshly hatched larvte of several genera are 

 provided with glandular hairs of various shapes. In Dntana iiifef/ei-rima they are clavate; iu 

 Dasylophia anguina they are clavate, somewhat flattened, and are dark, but clear at the tip,^ while 

 in all the other caterpillars we have observed that the glandular hairs are confined to the body, 

 those on the head tapering to a point, and apparently not fitted for secreting a fluid; those on the 

 head of Dasylophia are glandular, all ending in a slight transparent bulb. 



Other genera of this group will probably on further investigation be found to possess glandular 

 setic in their first larval stages. They occur in the freshly hatched larva of what is probably a 

 species of Heterocampa, also in Xadnta fiibbosa, Iclitliyiit-a incliisa, and Pheos'ui riinosa. 



It is to be observed that the freshly hatched caterpillars of Ceratosia tricolor Smith are 

 provided with glandular hairs. They are flattened at the tip, which is slightly tridentate, with 



' The use of these I find explained by Mr. Hellins iu his description of the larva of C. bifida iu Buckler's Larvie 

 ■of British Butterflies .-ind Moths, ii, p. 142, .as follows : "At the tip of tlie an.al flap are two sharp points, and another 

 pair underneath, whicli are used to throw the jiellets of frass to a distance." Similar duugforlis are very generally 

 present iu Geometrid larvx, the paranal papilliform tubercles being well developed, thougli we have not .seen them 

 in use. 



I have noticed a caterpitlar of C. hoiraVia in the process of defecating, and with the forceps pulled oft' a pellet 

 which was held by tlie two spines of the paranal tubercles. Mr. Dyar tells me lie has both seen and heard the 

 ■caterpillars casting their pellets with the aid of their spine against the side of a tumbler. 



'P\. XXXVII, fig n. <;;iandular hairs of Dasylophia; a, of body; h, of the head; e, of |)rothuracic shield. 



