DEVELOPMENT OF THE DKAGOxX FLIES. 355 



very siugular mechanism. There is a large hollow in the bod}^ 

 extending nearly through the entire length of the abdomen, and 

 having an opening at the end of the tail. 



The primary object of this cavity is respiration, for the gills 

 of the larva open into it, and when the creature is at rest the 

 cavity is gently filled with and emptied of water, so as to keep 

 the gills constantly supplied. If, however, the larva be alarmed, 

 or desirous of darting rapidly on its prey, the enclosed water 

 is suddenly and violently ejected, so that, by its reaction, the 

 insect is driven forward on exactly the same principle as tliat 

 by which a rocket is driven through the air. It can fill and 

 discharge this chamber with much celerity, so that it traverses 

 the distance of a yard or so with very great speed. It seldom, 

 however, uses this mode of progression if it wishes to travel to 

 a distance, but prefers its legs. As a rule the larva does not 

 discharge the water-chamber more than three or four times in 

 succession. 



The appearance presented by the undeveloped wings of the 

 pupa has already been mentioned. When the pupa has finished 

 leeding and is about to pass into the perfect state, it crawls out 

 of the water by means of a reed or other aquatic plant, or, in 

 cases where no convenient plants exist, ascends the bank. 

 When in the air it climbs to some little height, mostly above 

 a foot, and then clings tightly to the object on which it has fixed 

 itself. The skin soon dries, and as the creature bends itself 

 backwards and forwards, splits along the back, and allows 

 the perfect Dragon Ry to emerge. As is the case with the 

 butterflies and moths, the wings are small, thick, and damp, 

 but are rapidly expanded by having air driven through the 

 vessels with which they are thickly permeated, and by being 

 constantly shaken in the breeze. As soon as they are dry, the 

 insect darts off in search of prey, and renews in the air the 

 predacious habits which it possessed in the water. 



As all the Dragon Flies are very similar in their habits, there 

 is little to be said respecting each species. I have therefore 

 selected only a few examples of these insects, so as to show the 

 different groups into which they have been arranged. Without 

 going deeply into systematic entomology, it will be sufficient to 

 state that the Dragon Flies fall naturally into two groups, which 

 are at once distinguished by the shape of the head. In the first 

 A A 2 



