THE EXTERNAL MORPHOLOGY OF THE LEPIDOPTEROUS PUPA. 59 



hairs in the cocoon, and we find that these pupae are less hairy than those 

 of the other Liparids. The above facts appear to point to the pupal hairs 

 of the Liparids having a special and independent development, quite 

 apart from those of the larva, only we must not forget that Dasychira 

 fasceliiia, which has the stoutest cocoon of all the British species, 

 is the most hairy of all, and, further, that we have still the position or 

 arrangement, so similar to that of the larva, to account for. Is the 

 stimulus (if any) of the position of a tubercle or tuft in the larval skin 

 sufficient to determine where, on the pupa, the independent or ancestral 

 pupal hairs shall develop, or must we look for an explanation to a (?) 

 possible mixing of larval determinants with the pupal ones, and if so, 

 what prevents imaginal ones from doing the same ? Or, finally, are we 

 to consider the pupal envelope a hybrid structure, partly ancestral 

 imaginal and partly modern larval ? ' ' 



Chapman has noted that as soon as the final larval skin of Smerinthus 

 ocellatus has been moulted, the larval stripes are visible in dark green, as 

 well as the sites of the prolegs ; the lips and the oral spine of the 1st 

 spiracle are chitinous. In another specimen, the stripes and prolegs were 

 observed to be of a vivid blue, as well as the lips of the mandibles and 

 four spots at the base of the labrum. Poulton has also made obser- 

 vations on the persistence of the characteristic markings of the larva 

 on the pupa, immediately after the latter state is assumed, and he 

 states that it appears to be due to the fact that the hypodermic cells of 

 the larva and pupa are the same, so that any pigment contained in 

 them during larval life may remain unchanged after the pupal period 

 has begun, such colours, of course, being concealed in the living pupa 

 by the opaque cuticle. They may be fixed, however, by placing the 

 newly-formed pupa in spirit, and thus checking the darkening of the 

 surface. The persistence of the lateral larval stripes may also be 

 observed in the freshly-formed pupa of Sphinx ligustri, and the purple 

 borders of the stripes are seen to bear a relation to the segments 

 similar to that borne in the larval stage, whilst the relation of the 

 coloured borders to the spiracles is just the same as that of the larva. 

 Poulton proved conclusively that the constitution of the coloured 

 stripes in the pupa is similar to that in the larva, and entirely different 

 from, and independent of, the darkening of the pupal cuticle. Similar 

 facts are true of the pupa of Acherontia atropos, which, if examined 

 immediately after pupation, shows the purple stripes and small circular 

 patches (which probably spread from the bases of shagreen dots) of 

 the larva, through the undarkened pupal cuticle. Poulton adds that 

 he has also observed the light oblique stripes, with their dark green 

 borders, of Smerinthus populi and of S. ocellatus, conspicuously appear- 

 ing upon the surface just after pupation. One can readily understand 

 how important these observations are in homologising the larval and 

 pupal segments and structures. In this way, Poulton has shown 

 (Tram. Ent. Sue. Loncl., 1888, p. 566) that the terminal spine of the 

 pupa of Aijlia tan is the homologue of the anal flap of the larva. In 

 the freshly formed pupa of this species the markings of the larva are 

 especially distinct and the subspiracular line (which forms so prominent 

 a feature of the larva, and is continued along each side of the anal flap 

 to its extreme apex) is equally conspicuous in the pupa and occupies 

 an identical position in relation to the terminal anal spine, which, 

 in this species, is blunt, and covered with an immense number of 



