104 NATURAL HISTORY OF AQUATIC INSECTS CH. 



a pair of flaps which form an efficient swimming-fan. 

 The body of the pupa, like that of the larva, is 

 abundantly supplied with air-tubes, and a communi- 

 cation with the outer air is still maintained, though 

 in an entirely different way. The air-tubes no 

 longer open towards the tail as in the larva, but to- 

 wards the head. Just behind the head of the future 

 fly is a pair of trumpets so placed that in a position 

 of rest the margins of the trumpets come flush with 

 the surface of the water. Floating in this position, 

 the pupa remains still so long as it is undisturbed, 

 but if attacked by any of the predatory animals which 

 abound in the fresh waters, it is able to descend by 

 the powerful swimming movements of its tail-fin. 



Water is prevented from entering the trumpets 

 by the peculiar hairs which project in great numbers 

 from their inner surface. When the pupa first comes 

 up from below, the trumpets are overspread by a film 

 of water, which, becoming thinner and thinner by 

 evaporation, at length suddenly bursts. 1 



Why should the position of the respiratory organs 

 be changed from the tail-end in the larva to the head- 

 end in the pupa ? Not only the Gnat but Chi- 

 ronomus, Corethra and many other aquatic Insects 

 exhibit the same phenomenon. Evidently there must 

 be some reason why it is more convenient for the 

 larva to take in air by the tail, and for the pupa to 

 take in air by the head. We have already discussed 

 the considerations which seem to have fixed the res- 



1 Useful information respecting the respiratory trumpets and 

 other organs of the Gnat pupa will be found in Dr. C, H. Hurst's 

 papers, (See later,, footnote tQ p, J44-) 



