16 Journal New York Entomological Society. [Vol. xiii 



forward so that in the larva it lies in the suture or even in the pro- 

 thorax. This is especially true of the Lepidoptera. The metathoracic 

 spiracle, when present, sometimes migrates forward as far as the suture, 

 while the abdominal spiracles as a rule, remain very nearly in their 

 embryonic positions. 



In the Coleoptera the mesothoracic spiracle usually lies in or near 

 the suture between the pro- and mesothorax (Pigs. 24, 25), but in 

 many cases it is distinctly in the mesothorax and in some species there 

 is little or no migration (Fig. 26). Its position corresponds very 

 nearly to that of the mesothoracic (Figs. 25, 26), but I have never 

 found it entirely in the mesothorax. 



The height at which the spiracles are situated on the pleurum is a 

 little below the place where, later, the suture is formed between the 

 dorsum and the pleurum and is always distinctly below the wing discs 

 (Fig. 31). The latter arise near the future sutural line and are always 

 centrally situated on their segments, while the spiracles lie in the front 

 part. Their relative positions can be determined by certain muscles 

 occurring in each segment. In certain Hymenoptera, however, the 

 spiracles are situated at some distance above the wing discs. 



Vestiges of the metathoracic spiracle are to be found in the larvae 

 of some Lepidoptera. An external examination of the silk-worm {B. 

 mori) will reveal the remains of this spiracle as a small, oval, faintly- 

 chitinized ring, occupying a position on the pleurum of the meta- 

 thorax near the suture between that segment and the mesothorax and 

 in a position very nearly corresponding to that of the first abdominal 

 spiracle on its segment, as will be seen by reference to Fig. 27 (drawn 

 from a larva killed after the second moult and cleared in cedar oil). 

 This shows the lateral tracheal system of the thoracic and first abdom- 

 inal segment and the forming wing buds. It will be seen that the 

 latter are centrally situated on the sides of their respective segments 

 and dorsad of the longitudinal tracheal trunk. A small tracheal 

 branch arising near the first abdominal spiracle pushes forward to the 

 metathoracic wing disc, where it is met and joined by another small 

 branch arising near the metathoracic spiracle. Two tracheae arising 

 similarly, near the meso- and metathoracic spiracles respectively, join 

 at the center of the mesothoracic wing disc (Fig. 27). These two 

 tracheal branches probably give rise, the one to the radio-medial and 

 the other to the costo-anal groups of wing tracheae. It is evident from 

 the position of these wing discs that they have not arisen from any 



