THE INTERNAL STRUCTURE OF THE LEPIDOPTEROUS PUPA. 71 



tongue, which, by an identical process, form the labial palpi. At the 

 anterior part of the head, where the organs are very close together, the 

 envelopes form many folds without any final use. The two layers 

 then unite and fall on the surface of the tarsus. In the mandibles and 

 the labrum there is only a cellular thickening without invagination. 

 On pupation, the trunk is three times folded on the head, and, before 

 raising the cuticle, one observes that a furrow already divides it into 

 two halves. By its size it prevents the approach of the mandibles, 

 whilst the maxilla? and filiere remain strongly depressed. The cimier, 

 before straightening itself as in the chrysalis, covers the head in the 

 fashion of a Phrygian bonnet (Gonin on Pieris brassicae). 



It will thus be seen that most of the head- appendages have under- 

 gone considerable differentiation by the time the pupal stage is entered 

 and the pupal structures formed. The histolytic processes are largely 

 confined to the tissues not involved in the formation of special imaginal 

 organs, and, to a great extent, the latter structures themselves have 

 only to complete their growth in the pupal stage. 



It is necessary, in order to obtain a clear understanding of the 

 changes that occur in the pupal wing, by means of which it is converted 

 into the perfect imaginal structure, to consider briefly the manner in 

 which the wing undergoes development during the larval period. The 

 details as to the facts have been worked out by many observers, and 

 we would here acknowledge our indebtedness to the work of Gonin and 

 Mayer for most of the facts in the following brief summary. Verson 

 has shown that traces of wings may be found in the embryonic larva 

 of Bombyx mori some days before it leaves the egg, when the wing 

 consists of a few cells in close propinquity to a tracheal branch placed 

 on the interior of the Avail of the body in the meso- and metathoracic 

 segments. They arise from four dorsal imaginal discs, placed two in 

 the mesothoracic and two in the metathoracic segments, which appear 

 towards the close of embryonic life. Landois (1871) and Pancritius 

 (1884) discovered the rudimentary winglets in young lepidopterous 

 larva? only 4mm. long. At this time they appear as infolded 

 hypodermal pockets, penetrated by tracheae. When the larva is full 

 grown it is evident that the wing is really a folded portion of the 

 hypodermis (tidrm., PL i., fig. 1) itself, enclosing a thin layer of meso- 

 dermal tissue (vibr. m., PI. i., fig. 1). The conditions, however, are 

 complicated. The wing-pad proper is a pocket-like o^folding of the 

 hypodermis, which is more or less folded upon itself. This pocket, 

 instead of lying exposed between the hypodermal covering of the larva 

 and its cuticula, is protected by being sunk into a deep saclike infolding 

 of the hypodermis, the walls of which are very much thinner than 

 those of the wing-pad, and, indeed, thinner than the rest of the hypo- 

 dermis. The walls of the infolded sac follow quite closely the foldings 

 of the wing-pad itself. In penetrating, from without inward, one 

 would traverse, in succession, in the region of the wing-pad, five 

 layers of the epidermis : (1) The outer and inner layers of the oper- 

 culum-like fold of the hypodermis which covers in the wing. (2) The 

 thick outer and inner layers of the wing-pad. (3) The thin inner layer 

 of the infolded sac (Mayer). 



In larva. 1 in the 1st stadium, 3-4mm. long, Gonin found the wing- 

 germs (discs) as a thickening of the hypodermis, with the embryonic 

 cells of Verson on the convex border. The two sides of the wing begin 



