SPHINGIM. 499 



originating in the Pterogonids, and having no direct ^descent from 

 the true Eumorphids, whose oldest traces are probably in the "Myron" 

 group. This branch would probably include Pachylia and Dilophonota, 

 and Manduca * might possibly be a branch here. This question 

 can only be resolved satisfactorily by a study of the young 

 larvae. All these divisions, originating practically together and 

 not one from another, have to be reconciled with the generalised 

 state of the young larvae in Eumorphids, otherwise one of the 

 most specialised groups. If we consider that the groups did so 

 arise, i.e., together, a certain confusion of characters in the lower 

 portions of the branches is not only very probable but practically 

 inevitable. A special group, like the Sphingids or the butterflies 

 probably arose, one might almost say, fully developed. For some 

 reason, say a rapid change of conditions — climatic, floral, parasitical, 

 or what not — or a sudden access of variability in some special 

 direction in the ancestral species or genus, threw open as it were 

 to competition, a whole new world, to which they were able to 

 adapt themselves ; this they colonised by rapid and abundant 

 variation in all directions, and, at this period, various distinct lines 

 might easily acquire similar structures quite independently of each 

 other. When this new world was fully colonised, so to speak, 

 evolution would fall back into its more regular and humdrum 

 routine." 



The ova of the Sphingidae are, in their broad aspects, 

 very uniform, and are very similar to those of the A?norphidae. 

 Such specialisation as there is, is found to be either in the direction 

 of small size, or in the reduction of the surface reticulations in 

 order to obtain a smooth surface. 



The larvae of the Sphingidae, as we have already noted (anted, pp. 

 367 — 368), although extremely specialised in colour, form, and resting- 

 position, are structurally, especially so far as the tubercular setae 

 are concerned, of a generalised type, the tubercles consisting 

 almost always of single-haired chitinous buttons, and placed in 

 generalised position on the abdominal segments, except that v is 

 obsolete, and an accessory prespiracular is present. One structural 

 peculiarity, however, occurs, and this is giving our phylogenists 

 much trouble. On the meso- and metathorax of the Eicmorphidi — 

 Eumorpha (elpenor), Theretra (porcellns), Celerio (gal Hi, euphorbiae ) — 

 i and ii are situated on separate subsegments, whilst in the other 

 subfamilies these are pushed up so as to be on the same subsegment 

 and either arise quite near each other (Amorphids), or else from a 

 common raised base (Hemarids), or on the same chitinous button 

 (Sphingids sens, strict.). Strangely, too, the Eumorphid larvae 

 have none of the bifid hairs, in their earlier stages, that characterise 

 the early larval stages of the Amorphidae, Hemarinae, Sesiinae, 

 and, to a slight extent, the Sphinginae. It is, we believe, true 

 that these bifid hairs are not present in all the Sphinginae, nor 

 are they noted as occurring in Manducinae, still, with some 



* One might point out here that the Manducid larva agrees with the Sphingid 

 and not at all with the 'Sesiid (Macroglossid) larva. Both the Manducid and Sphingid 

 larvae have tumid and pellucid-looking thoracic segments, whilst no Sesiid larva 

 that I have handled or seen figured shows any traces of this distinctive feature 

 (Bacot). 



