LAEVA OF THE DRAGON-FLY. 271 



mouth, so that the prey can be eaten without trouble. In 

 Plate VIII. Fig. 4, the end of the still folded mask is seen, the 

 jaws being opened preparatory to the darting forward of th»i 

 mask. The shape and details of this structure vary somewhat 

 in the different species. 



The larva is a voracious creature, and will eat almost any 

 U4uatic insect which it can master, and has even been known 

 to catcli and eat small fishes. In its own turn, however, it 

 falls a victim to the more powerful inhabitants of the water, 

 being attacked and eaten by the Dyticus and even by the 

 Acilius. Two specimens of Acilius which I kept in an 

 aquarium managed to eat, in little more than a fortnight, 

 three out of four Dragon-fly larvae which were in the same 

 vessel ; and in the waters where the creatures live I have often 

 found the dead and partly-eaten bodies of Dragon-fly larvse, the 

 destroyer having evidently been the Dyticus. 



We will now proceed to the other end of the creature, and 

 examine its tail. When taken out of the water, the tail 

 appears to end in a sharp, horny point ; but, when it is re- 

 placed in water, the pointed tail opens and shows that it is 

 composed of five sharp appendages, three being much larger 

 than the others. When they are opened, a tube is discovered 

 passing into the body of the larva. All the five spikes are 

 jointed together at their bases, so that if one be drawn aside 

 the other four immediately spread themselves apart. The 

 aperture which lies between them is just large enough to allow 

 an ordinary pin to pass easily through it. 



Within the abdomen is the respiratory apparatus, which ex- 

 tracts the oxygen from the water, and, when that gas has been 

 exhausted, the water is expelled and a fresh supply taken in 

 through the same aperture. If the larva be j)laced in a shallow 

 vessel and watched carefully, the process of respiration can 

 be easily observed. The five pointed projections are kept about 

 half open, and the water is ejected in gentle and tolerably 

 regular pulsations as long as the creature is undisturbed. 

 Siiould it, however, be alarmed, it suddenly" expels the water 

 with such violence that it drives itself forward with irreat 

 speed, the principle being exactly the same as that which 

 causes the flight of a rocket. 



If the bottom of the vessel in 'vhich the larva is kept be 



) 



