272 MESSES. MIALL AND HAMMOND ON THE DEVELOPMENT 



occiput. The secondary forward-directed fold is long and narrow, and extends from the 

 back of the head into the larval maxilla. As it lengthens it liecomes coiled, and much 

 resembles one of the developing imaginal legs (fig. 27). The new parts thus formed are 

 those of the pupa, and the imaginal rudiments are enclosed within them. The pupal 

 integument of the head, like that of some other parts of the body, recedes considerably 

 from the larval cuticle, and the imaginal integument recedes again from that of the 

 pupa, so that in sections of the pupal head a tolerably wide space separates the mouth- 

 parts of the fly from the empty cuticle which represents the corresponding organs of 

 the pupa. 



The history of the invaginations which give rise to the head of the fly can be followed 

 in a series of larva? of different ages. They are not to be discovered even in a rudi- 

 mentary state until after the last larval moult*. Weismannf has given reasons for 

 sujiposing that invaginated imaginal rudiments could not come into existence before the 

 last larval moult in an insect whose life-history resembles that of Corethra or Chiro- 

 nomus. If the epidermis were invaginated in any stage befoi"e the ante-pupal one, the 

 new cuticle, moulded closely upon the epidermis, would become invaginated also, and 

 would appear at the next moult with projecting appendages like those of a pupa or 

 imago. This is actually the way in which the wings are developed in some larval insects 

 with incomplete metamorphosis. In Muscidse the invaginations for the head of the 

 imago have been traced back to the embryo within the egg J, but the almost total 

 subsequent separation of the disks from the epidermis renders their development inde- 

 pendent of the growth of the larval cuticle and of the moults that prol)ably take place 

 therein §. 



Very soon after the last larval moult, when the Chironomits larva is about half an 

 inch long, the first indications of the invaginations can be discovei'ed by means of 

 sections. They form rapidly, and among larvae quite similar in size and outward 

 appearance some are found to exhiliit tolerably advanced invaginations, while others do 

 not possess even the rudiments of such structures. Fig. 23, PL XXIX., represents a 

 moderately early stage. Here the invaginations are restricted to the larval head, and 

 form comparatively simple paired folds of the dorsal epidermis. Behind and on the 

 ventral side is a short extension {If"), which will subsequently give rise to the compound 

 eye and the antennary bulb. As the invaginations do not as yet extend into the thorax, 

 the transverse fold described above is wholly wanting. In this early condition the inva- 

 ginations of Chironomus are essentially similar to those of Corethra at the time of their 

 fullest development. 



The prolongation of the cephalic invaginations into the thorax gradually advances as 



* There are probably four larval moults in CJiirovnnuis, as in Corethra, but the burrowing habits of the insect 

 render it difficult to be quite certain of the exact number. 



t " Metamorph. der Corethra" loc. cit. p. 115. 



i Lowne on the Blow-fl)', new edition, pp. 2, 41 (fig. 7). 



§ Leuekart and Weismann have inferred the occurrence of at least two moults in the larva of the Blow-fly, from 

 the changes observed in the stigmata and the hooks. Weismann suspects that as many as four moults may take 

 place (Entw. der Diptercn. p. 104). 



