STRUCTUUE OP PUP^. 301 



Reaumur and De Geer proved this position by 

 numerous experiments. When a chrysahs, lor in- 

 stance, is suspended by a thread and immersed in oil 

 up to the tip of the winjr-cases, it does not seem to 

 be much injured, and the perlect insect is disclosed 

 in due time. If respiration, therefore, be essential to 

 the life of the chrysalis, it appears as if it could exist 

 with at least the greater number of its spiracles ob- 

 structed ; but this docs not happen with a chrysalis 

 just formed, which always dies. By immersing the 

 wiiole chrysalis in oil, it is certainly killed; and even 

 by immersing its head downwards as far as the lirst 

 pair of spiracles situated near the head. This seems 

 to indicate accordingly that this first pair is more es- 

 sential to the insect than all the rest; and in other ex- 

 periments it is also tbund to emit a much greater 

 quantity of air by this first pair. It seems of much 

 importance in such experiments to attend to the age 

 of the i)upa; for when near the change the function 

 of respiration is carried on more feebly, and at length 

 nearly ceases. 



Besides the decisive experiments of immersion in 

 oil, Reaumur placed pupae of various species in the 

 exhausted receiver of an air-pump, and at every 

 fresh stroke of the piston their bodies both bulged 

 out and became elongated: because, as he inferred, 

 the envelope is not pervious to the air contained in 

 the body, and the spiracles do not allow of its escape 

 with sufficient rapidity to keep pace with the ex- 

 haustion of the receiver; — contrary to what happens 

 when caterpillars are subjected to the same circum- 

 stances. Varying his experiments, he placed in the 

 exhausted receiver a vessel containing water deprived 

 of its air, and in this plunged a chrysalis, keeping 

 it immersed by means of a weight attached by a 

 thread. At the two or three first strokes of the pis- 

 ton, bubbles of air appeared at each of the spiracles, 

 VOL. VI. '26 



