II FLIES WITH AQUATIC LARV^ 205 



until only the tip of the respiratory tube projects 

 from the outer sheath, the channelled sheath disap- 

 pearing altogether. If further retraction is necessary 

 the muscles of the outer sheath itself come into 

 action, and this sheath shortens to any requisite ex- 

 tent. Its hinder end is furnished with a strong 



d> 



circular or sphincter muscle, which grips the respira- 

 tory tube at such times and prevents it from slipping 

 out. The base of the respiratory tube then be- 

 comes curved or even doubled up within the body, 

 reaching in some cases almost as far as the head 

 (Reaumur), and the coils of the tracheae become very 

 numerous and close. Reaumur and Batelli are in- 

 clined to suppose that the coiled tracheae play the 

 part of springs and help to effect the protrusion of 

 the tail, but this I cannot think to be correct. The 

 tracheae, even if it be admitted that they can act as 

 springs, have no base from which to act. The pro- 

 trusion of the tail is, I think, entirely due to pres- 

 sure set up within the body by contraction of its 

 muscular wall. If you take a dead larva with con- 

 tracted tail between the finger and the thumb, and 

 gently press the contents of the body backwards, the 

 tail will suddenly shoot out to its full length. 



" The tracheae are continued forwards into the 

 body, becoming dilated into large air-sacs. They are 

 in communication with a pair of anterior spiracles 

 which appear on the upper surface of the first seg- 

 ment. These spiracles are not functional in the 

 larva." 



Reaumur says that he often saw a bubble appear 

 at the tip of the tail in a submerged larva. At first 



