AQUATIC COCKROACHES. 223 



their bonds had to be tightly fastened in order to keep them in 

 position ; even then the prisoners did not relax their efforts to 

 free themselves, and I believe that they died of exhaustion and of 

 injuries sustained in their struggles rather than from drowning. 

 The aquatic species, on the other hand, remained comparatively 

 quiet ; the reversed specimens, being fastened in a position more 

 or less natural to them, and being able to obtain their supply of 

 air in quite a normal and usual manner, were very little dis- 



Fig. 1. — End of abdomen of a larva of Rhicnoda natatrix, sp.n., dorsal 

 view ; s.p. spiracular tube ; c. cercus ; s. a. I. last dorsal tergite. 



Fig. 2. — The same, ventral view ; st. abdominal sternite ; s. spiracle ; 

 t. abdominal tergite. 



Fig. 3. — Spiracular tube, highly magnified, seen in optical section ; 

 v. vestibule; s. spiracle; m. muscle working &., the chitinous bow; t. 

 trachea. 



tressed by their bonds, and the other specimens were soon reduced 

 to a comatose condition by the difficulty of obtaining air. The 

 terminal spiracles of these aquatic Cockroaches are situated at 

 the base of two tubes visible on the dorsal side projecting from 

 below the seventh tergite, and external to the anal cerci (fig. 1). 



