8 Journal New York Entomological Society. [VoI.viii. 



what is suggested in Figure 14, T. These evaginations are extended 

 into the lumen of the wing, which process takes place while the 

 tracheoles are becoming completed and straightened out. The lumen 

 of these small tracheal in the wing, has no connection with the main 

 lateral tracheal trunk (Fig. 9, //), since they are closed by the pre- 

 existing intima, therefore these tracheae are not functional. As the 

 tracheae lengthen the tracheoles straighten out and extend to the mar - 

 gin of the wing. Tlie tracheoles occupy the space, the vein cavity, 

 along the course of the tracheae as they push out into the wing. Ver- 

 son thought that the tracheoles never reached the margin of the wing, 

 but my observations bear out those of Gonin who says they do extend 

 to the margin of the wing unless hindered by the more dense por- 

 tion of the edge of the organ. It is probable that the pressure of the 

 air is the main agent in straightening out the tracheoles. 



The Fjkth Larval Stage. (Plate II, Figs. 12-14, 16, 17. Plate 



III, Fig. 23.) 

 In the study of this period two methods were used, that of direct 

 observation of the entire wing-bud and that of the study of sections. 

 The external lateral trachea sends out branches to the wing and 

 these enter it at about one third of the distance from the hypodermal 

 attachment to the margin and give the appearance, as Gonin states, 

 of the stem of a nasturtion leaf (Fig. 12). The tracheae seem to 

 enter as into an arm pit and the tracheoles extend from this point to all 

 parts of the organ, even to near the margin where the cells are denser 

 and do not allow them to penetrate. The tracheoles extend in masses 

 of bundles from the trachea as previously described. Tracheae, be- 

 ing filled with lymph at this stage, are not easily seen in the entire 

 wing-bud, but the photograph for Figure 12 was taken under very 

 favorable conditions and they are faintly outlined. 



Upon the study of sections (Fig. 14) it will be observed that the 

 process of evagination has gone on from that described in the stage 

 preceding this until the v^'ing is nearly surrounded by the peripodal 

 membrane (e). The evaginated portion is better shown in Figure 16 

 which is a frontal section of the larva and gives a longi-section of a 

 part of the wing. In cutting the section shown in Figure 14 five 

 layers of hypodermal origin were passed through, viz., the outside 

 body wall or hypodermis (Ji), the external part of the peripodal mem- 

 brane (if'), the external boundary of the wing-bud (//), the internal 



