MUS. COATES'S BATH. 77 



extends for several inches behind the larva. But when 

 it desires to propel itself quickly the larva expels the 

 water violently, and so, on the principle of the ' direct- 

 action ' machine, drives itself forcibly in the opposite 

 direction. Thus the progress of the dragon-fly larva 

 is necessarily a series of jerks, as some appreciable space 

 of time is required in which the hollow can be filled 

 with water. The nautilus, the common cuttle-fish and 

 their kin, propel themselves in the same manner, 

 which is exactly identical in principle with the flight 

 of the rocket, and, in the creature called the Flying 

 Squid, produces much the same effect. 



If the larva be placed in a shallow and flat vessel, 

 in which some very fine dark sand has been scattered, 

 the whole process is rendered plainly visible. When 

 the larva remains quietly in one place, the sand is 

 gradually washed away in a direct line with the insect, 

 leaving a track about a quarter of an inch wide, and 

 some three inches in length. This track is very clear 

 and well-defined near the insect, but becomes vague 

 and broad in proportion to the distance from the larva. 

 Now, if the larva be touched, a very different appear- 

 ance is shown. The larva darts suddenly through the 

 water, and, instead of the simple narrow track, a broad 

 fan-shaped track is left, the water having been expelled 

 with such force as to drive away the sand on both 

 sides. 



Its mode of eating is as strange as its progression. 

 The lower lip, instead of being, as it mostly is, a mere 



