ATTACIDES. 277 



tuna, &c, adopt a Sphinx-like attitude more marked in some stadia 

 than others, whilst the development of the double tubercle (i) on the 

 8th abdominal segment suggests strongly the Sphingid caudal horn. 

 This is more distinctly divided in Telea polyphemus than in any other 

 American Attacid, whilst it shows no such distinct trace of its double 

 origin in the ist stadium of Actias luna, as in the ist stadium of 

 Callosamia promethea or Telea polyphemus, yet in A. luna it is more 

 duplex than in Platysamia cecropia. Packard distinctly states {Proc. 

 Am. Ac. Arts Sci., 1893, p. 57) that, in his opinion, this single median 

 tubercle of the 8th abdominal in the more specialised Attacids is the 

 homologue of the caudal horn of the Sphingids, of Bombyx mori, and 

 of the Notodont genus Pheosia, and believes it to result from the fusion, 

 before the end of embryonic lite, of what were originally two separate 

 tubercles, like the two separate ones of Saturnia. Packard's opinion, 

 therefore, confirms that of M tiller {Siidamerikanische Nymphaliden- 

 raupen, 1886, pp. 249, 250). [See also Poulton {Trans. Ent. Soc. London, 

 1885, p. 30), Packard {Proc. Post. Soc. Nat. Hist., xxv., pp. 99, 103, 

 footnotes 1, 2 and 3), and Grote {North Amer. Lepidoptera, 1886, pp. 

 16, 54).] It is difficult to form any definite opinion as to how far the 

 caudal horn of the Sphingids is foreshadowed in the Attacids as an 

 ancestral character, and how far it is a matter of convergence, but 

 the similarity of structure certainly is entirely due to convergence in 

 the case of those Notodonts that have developed this peculiar character. 

 The armature of certain of the spined larvse of the Citheroniids 

 and Hemileucids is figured by Packard (Proc. Am. Phil. Soc, xxxi., 

 pi. v — x). Some of the structures (enlarged) here exhibited are very 

 remarkable. Barrett states {Can. Ent., xxxii., p. 236) that the larva 

 of Copaxa multifenestrata is most strikingly beautiful, and that in that 

 of Automeris -Janus the spine defence system is carried to an extreme; 

 the length of the profusely branching spines is 15mm. — 25mm., or 

 twice the diameter of the body, and so abundant are they that the 

 larva looks like a bunch of moss a few yards away, while the quantity 

 of poison contained in these spines is so great that during the process 

 of inflating, the fumes which are driven off with the vapour are 

 positively dangerous to the operator. The fullgrown larva of Sphingi- 

 campa bicolor is remarkable, not only on account of the two dorsal 

 rows of acutely conical spines, with three or four blunt spinules, 

 on abdominal segments 1 — 7, which are recurved and directed 

 backwards, but also because the spines themselves are externally 

 coloured with a pearly, silvery-white hue, which gives off all the 

 colours of the rainbow during the movements of the larva, the 

 corresponding spines on the other segments being similarly but less 

 brilliantly coloured. Packard says : " The general colour of the 

 body is of nearly the same hue as the undersides of the leaves of the 

 honey-locust, and, thus coloured, it is partly assimilated and protected 

 by its colour, whilst the horns are in general like the spines of its 

 foodplants ; on the other hand the gleaming silvery spines certainly 

 render the creature conspicuous, as well as the lateral parti-coloured 

 band. It would appear probable that the formidable spines of the 

 grown-up caterpillar save it not infrequently from being swallowed 

 by birds, though the horns are probably of greater use in the earlier 

 stages, when they are much longer and much more movable in 

 frightening away ichneumons and Tachinae, e.g., even when 20mm, 



