362 BRITISH LEPIDOPTERA. 



specialisation proceeds in the larva (i) by the atrophy of tubercle 

 ii and the development of i on abdominal segments, (2) by the 

 development of i, iii, iv + v into a series of projecting spine-bearing 

 processes on either side of each segment from the mesothorax to the 

 7th abdominal, (3) by the special development of i (and ? ii) on the 

 meso- and metathorax as a pair of highly-developed spine-bearing pro- 

 cesses, (4) by the union of bases of i on abdominal segment 8 into a 

 caudal horn. He instances the various stages indicated by means of: 

 (1) Platysamia cecropia and Callosamiapromethea, which show the usual 

 ringed series on each segment commencing from the mesothorax, 

 characteristic of the Attacids (sens, strict.), and with no defined 

 caudal horn on the 8th abdominal. (2) Eacles imperialis and 

 E. penelope, in which i on the meso- and metathorax are small and 

 inconspicuous. (3) Citheronia wardii, C. principalis, C. var. argyra- 

 cantha, which have i on meso- and metathorax, and the caudal horn 

 rather better developed than the others. (4) Rhescynthis erythinae, 

 Citheronia phoronea, C. ixion, C. regalis, with large spined i on 

 meso- and metathorax, well-developed caudal horn on 8th ab- 

 dominal, and, in R. erythinae, a smaller on the 9th abdominal, 

 whilst his description of Ceratomia taken alone would show it 

 to be almost certainly Attacid, but Dyar writes {Ent. Rec, xii., p. 21) 

 that these horns are perfectly secondary and merely a special adaptation, 

 the ordinary oblique lines being replaced by rows of teeth, and the 

 " horns" only another manifestation of the peculiar conversion of mark- 

 ings into structural characters (see also, posted, p. 364). Poulton also 

 notes that the horns, &c, of these specialised Citheroniid larvae, are 

 dropped at the last moult, the larvae becoming entirely smooth, with 

 atrophied tubercles, and with a coloration exactly suited to their 

 environment. [For figures of the larvae noted by Poulton, see Bur- 

 meister's Atlas of the Lepidoptera<f the Argentine Republic, pi. xix et sea.] 

 Bearing on this last point is the fact that the essential facies 

 of the adult Sphingid larva is similar to that just noted as being 

 characteristic of certain adult Attacids. In the early stages of 

 the Sphingid larva, how r ever, we find an amazing difference from 

 the Attacid, and the mode of evolution of a similar facies is entirely 

 different. The newly-hatched Sphingid larva has, in its most 

 generalised form, an almost primitive first stage, i and ii trapezoidal, 

 iii supraspiracular, iv subspiracular, each simple and bearing one seta. 

 Three specialisations (?) occur — (1) the atrophy of v; (2) the develop- 

 ment of a special prespiracular (common in some Lachneids) ; (3) the 

 presence of a more or less developed caudal horn on abdominal 

 segment 8. Gradually, and never suddenly as in certain Attacids, 

 the larva loses its simple primitive tubercles and becomes a smooth 

 larva with only the caudal horn developed. Then the Sphingid larva 

 bears a close superficial resemblance to the adult Citheroniid and 

 Agliid larvae, and Poulton makes the former a direct evolutionary branch 

 of the latter and attempts to support his view with the African, 

 spined, so-called Smerinthine, larva, LopJiostetluis, which, apparently 

 from the description, is not Sphingid at all, and certainly not 

 Smerinthine. The Sphingid larva is very specialised in colour and 

 markings, but in a very simple, clear, and definite direction. The 

 Attacid (sens, laf.) larva is even more specialised on distinctly 

 different and independent lines and in a much more complicated 



