BY R. .T. TILLYARD. 399 



gas should completely fill the respiratory system of the larva. 

 The second solution, that it might be oxygen, would require the 

 existence of some oxygenating substance in the blood, which 

 would act on the C0„ produced by metabolism in the developing 

 embryo, and release oxygen therefrom, in some manner analogous 

 to that of chlorophyll in the leaves of plants. But, though the 

 blood of Odonata is always of a yellowish or greenish colour, we 

 have it, on the high authority of Cuenot,* that the pigments 

 hcemoxanthine, h(emochlorvne, and hcemojwasme, which are the 

 only substances that might conceivably perform such a function, 

 are quite absent from the blood of these insects. 



It seemed necessary, in any case, to obtain some definite result 

 by experiment, and not to rely upon arguments alone, however 

 conclusive they might appear. Unfortunately, by the evening 

 of April 3rd, as the table already given will show, most of my 

 eggs had already hatched, and I could not expect to obtain many 

 more larvae. I was forced, therefoi-e, to limit my experiments, 

 both in variety and in the number of larvre on which they could 

 be performed. I decided to employ only two test-solutions, whose 

 combined results should indicate, with considerable certainty, the 

 composition of the gas in the tracheae. These wei*e 

 {a) a 10% solution of caustic potash. 

 (6) a 4% solution of Azol.f 

 As a stand-by, I also prepared, (r) a 4% solution of caustic soda, 

 in case solution (a) should prove too strong. 



Control Experiments :— As a control upon the experiments, two 

 live larvae were taken from the Petri dish (Series C), and killed 

 by being placed in a tube of water, in which a single drop of 

 chloroform had been well shaken up. They were then removed 

 to tap-water in a crystal dish, and the times taken for the tracheal 

 system to collapse were noted. 



* Etudes sur le sang et les glandes lymphatiques dans la serie animale, 

 2'"'' partie. Invertebres. Arch. Zool. exper. 2""* serie. ix., 1801. 



t Azol, the proprietary name for an excellent, one-solution, photographic 

 developer, which acts as a verj' strong reducing agent. The connect 

 strength for use on plates is 1 part in 24 of water, or roughly 4%. 



