!i FLIES WITH AQUATIC LARV^ 113 



arc glued together, so as to make a little concave 

 float, shaped like a shallow boat. The upper end of 

 each egg is pointed ; the lower end is provided with 

 a lid, through which the larva will ultimately issue 

 into the water. The Gnat in all stages, even while 

 still in the egg, requires an ample supply of air. It 

 is therefore necessary that the egg-raft should float at 

 tlie surface ; it is also necessary that it should always 

 float in the same position, so as to facilitate the escape 

 of the larva. This is effectually secured by a pro- 

 vision of almost amusing simplicity. Let us first 

 notice how efficient it is. If we take two or three of 

 these tiny egg-rafts, and place them in a jug of water, 

 we may pour the water into a basin again and again ; 

 the egg-rafts float instantly to the surface, and the 

 moment they come to the top they are seen to be as 

 dry as at first. The fact is that the surface-film 

 cannot penetrate the fine spaces between the pointed 

 ends of the eggs. The cavity of the egg-raft is thus 

 overspread by an air-bubble when accidentally sub- 

 merged. The eggs are kept from contact with water ; 

 the proper upper surface is made so buoyant that 

 the raft has great power of self-righting ; while, the 

 instant that it comes to the top, the excess of 

 water drains off, and the film bursts, leaving the eggs 

 perfectly dry on their upper surface. 



CORETHRA 



In still pools, especially such as are overhung by 

 trees, there may often be found in considerable abun- 

 dance the predatory larva of Corethra, also called the 



I 



