OF THE HEAD OF THE IMAGO OF CHIRONOMUS. '211 



a cloth to be spread out between two rails, then a hand grasping- the cloth at one place 

 may be made to push downwards and backwards until both liand and arm become 

 buried in a deep fold. The fist will correspond to the bulb of the antenna, the arm to 

 its shaft, and the fold in the cloth to the loniyitudinal invagination. This rude model 

 will also show how it becomes necessary to introduce a transverse fold, if the longi- 

 tudinal fold is to extend beneath an undisturbed surface of cloth or epidermis. 

 In all stages of larval growth the imaginal antenna* encloses the larval antennary 

 nerve, the invagination being, in fact, formed about the nerve, but in the pupa this 

 nerve becomes no longer traceable and new structures appear to take its place. 



The proportions of the male and female head differ materially in the adult fl.y. In 

 the male the antennary bulbs are larger and closer together than in the female. This 

 difference is already apparent in the antennary invaginations of the larva (Pi. XXIX. 

 figs. 21 & 22). We have found it possible to determine with certainty the sex of living 

 larvse by observation of the form of the incipient generative organs. Having marked 

 several specimens as male or female, we have cut sections through the growing heads of 

 the larvae so marked. In the female the invaginations are wider apart, and the antennary 

 bulb projects from the inner wall into the interior of the invagination. In the male 

 the invaginations are so close that they almost or actually touch behind, and the 

 antennary bulbs are at first connected with their i^osterior extremities. As the develop- 

 ment of the imaginal head advances, the antennary bullj, even in the male, becomes to 

 a great extent internal {i. e. facing the middle line) rather than posterior (PL XXX. 

 fig. 25). In this stage it may be distinguished from that of the female by its larger size, 

 and by its extending backward up to, and even a little beyond, the hindermost extremity 

 of the compound eye, which it never does in the female. 



In the compound eye of Chlronomus before pupation the epidermic cells of the 

 so-called vitreous layer are often much elongated and i-esemble fibres. Tbey retain their 

 power of forming cuticle to a late stage, ultimately producing lenses which are not 

 biconvex, but hollow, convex externally and concave internally. The retinal cells are 

 pigmented and form retinulne of seven cells each. No crystalline cones are formed, and 

 the eye of the fly is therefore aconic. 



Simultaneously with the formation of the compound eyes and the imaginal antennse, 

 new mouth-parts are developed. As in Corethra, they develop within those of the 

 larva. On either side of the salivary ducts and their common opening into the mouth, 

 the epidermis of the larval head becomes infolded, and the pouches ultimately extend 

 backwards to the back of the head (fig. 2G). Prom the inner side of each pouch, and 

 close to its hinder extremity, a secondary invagination pushes forwards and downwards, 

 and this ultimately gives rise to the labella t of the fly. In larva3 ready to change into 

 pupa) the tips of the labelhe are bent inwards, towards each other, at a right angle. 

 The invagination for the maxillary palp forms on the side of the larval head. The 

 mouth of the primary fold is at first nearly equidistant from the larval maxilla and the 



* We do not at present distinguish between the imaginal and the pupal antenna. 



t See Meinert, ' Fluernes Munddele,' 18S1, or Diramock, ' Anat. of Mouth-parta of Diptera,' 18S1. In Ortho- 

 ptcrous insects what is apparently the same part is named 2'a''«i??ossrt. 



