ix DRAGON-FLIES 335 



do not depend upon their spiracles even when the}' 

 are open, but are able to breathe in other ways 

 also. 



The late Dr. Hermann Dewitz left behind him 

 some valuable experimental notes, which show, 

 among other things, that the respiration of Dragon- 

 fly larvae is effected in more than one way. 1 



He endeavoured, in the first place, to ascertain 

 whether the spiracles, and especially the large thoracic 

 spiracles, were as completely closed as had been 

 assumed. For this purpose he employed various 

 methods. Sometimes the larva was placed in more 

 or less dilute alcohol, when a train of bubbles issuing 

 from one (seldom from both) of the large thoracic 

 spiracles would often establish the fact that they \verc 

 pervious. Gentle warming, and even cooling, pro- 

 duced like results in many cases. When a thin glass 

 beaker, containing a live larva, was held in the hand, 

 the rise of temperature was often enough to cause 

 bubbles to issue from the spiracles. Useful informa- 

 tion was got also by placing living larvae in water 

 that had been boiled to get rid of dissolved air, and 

 afterwards cooled. The larvae were enabled to reach 

 the surface by climbing up a stick, and note was 

 taken of the way in which they came up. If they 

 came up head foremost and pushed part of the thorax 

 out of the water, this was a distinct indication that 

 they relied mainly upon their thoracic spiracles. 

 Some larvae, under these circumstances, come up 

 backwards, with the head down, and push the tip of 

 the abdomen out of the water an obvious proof 

 ZooL Am. Bd. XIII., pp. 500-504; 525-531 (1891;. 



