58 BRITISH LEPIDOPTERA. 



larvae may be stated to originate from a local transformation of 

 two portions of the subdorsal line into eye-spots, and the subsequent 

 transference of these two primary ocellated spots to the other seg- 

 ments. The eye-spots always originate on segments 4 and 5, and 

 from there the transference mostly occurs backwards, although, in 

 certain cases, it takes place at the same time forwards. In the origin 

 of the eye-spots then lies a great distinction between the Chcerocampids 

 and Deilephilids, which were formerly associated, but in which the origin 

 of a very similar kind of marking can be traced to quite another source. 

 With a better knowledge of the larvae of the Eumorph'id species, 

 one would expect considerable modification in their classification. 

 We are even now without a knowledge of the early larval stages 

 of many species, the adult larvae and imagines of which were figured 

 more than a century ago, and, until these are obtained, no real 

 advance in the grouping of these species can be made. No doubt 

 a complete detailed ontogeny of such species a.s Dap/inis nerii, Hippo- 

 Hon celerio, Xylophanes oldenlandiae and Fiorina japonica, &c, might 

 be obtained did lepidopterists know how important a knowledge of the 

 early larval stages of these and allied species would beat the present time. 

 The Eumorphid (sens, strict.) pupae (porcellus and elpenor) are 

 very close to those of the Phryxids (eupliorbiae and gallii) with the 

 very important exception that they have developed certain head 

 tubercles and certain abdominal spines, apparently with the object 

 of enabling the pupa either to move within its puparium or to 

 escape from it. Functionally and morphologically the abdominal 

 spines are the same as those that exist in so many pupae-incom- 

 pletae ; they are for moving the pupa, and they are special modifi- 

 cations of the secondary hairs (as seen in larva), though, as to 

 this view of their derivation, whilst they are undoubtedly modifi- 

 cations of hair-points, and these are again probably modifications 

 of the spiculate skin-points, it is more probable that in the pupae-in- 

 completae the spines are direct modifications of skin spicules, without 

 the intermediate condition of modification into hairs. Their actual 

 positions, as already noted, differ also. They are, in fact, a 

 precisely similar, if not identical, structure developed de novo, 

 without any genetic relationship to that of the pupae-incompletae. 

 The pupa of Eumorpha elpenor cannot, in any other respects, be 

 said to differ from that of Hyles euphorbiae to any important 

 degree except in the surface sculpturing. The maxillary keel is 

 a little more prominent and that is all. In that of Theretra 

 porcellus the maxillary keel is very definitely more fully developed. 

 The keel is more prominent, sharper, and has pushed the 

 labrum a little further to the dorsum, and itself projects to the 

 front of the pupa. Another point of distinction is that the surface 

 sculpturing, which, in all this group, is wrinkling in front, and 

 becomes pitting only on the later segments, is, in porcellus, 

 wrinkling right away down to the 9th abdominal segment. It 

 is possible that this is relative to the loss of the larval horn. In 

 the larva, the horn preserves a primitive (comparatively) skin 

 structure, long after it has disappeared from the rest of the surface. 

 This conservative tendency of the horn may have some similar 

 effect on the neighbouring cutaneous structure of the pupa. In 

 comparing llcmaris fuciformis with J^tcrogon proserpina, we find 



