272 EXPERIMENTAL INVESTIGATIONS. 



in order to ascertain their voltaic relations. Hydrogen was 

 much more electro-positive than carbonic oxide, or rather 

 formed, with the oxygen of the atmospheric air in solution, a 

 combination which overpowered the opposite tendency of the 

 carbonic oxide and air. 



Experiment 18. Chlorine and olefiant gas gave a very 

 feeble effect upon iodide of potassium. After four hours the 

 liquid in the olefiant gas tubes had not risen more ^ in the 

 closed circuit than in the detached pair ; the chlorine was 

 nearly all absorbed in solution. 



Experiment 19. Chlorine and carbonic oxide gave very 

 notable effects ; ten cells decomposed water. From the ex- 

 treme solubility of the former gas the equivalent relationship 

 could not be ascertained. 



It now occurred to me that as oxygen and hydrogen are 

 evolved from water by electrolysis, and conversely form water 

 by electro-synthesis, so some other gases which are evolved 

 from certain electrolytes by voltaic action might, when ar- 

 ranged as a gas battery with the electrolyte from which they 

 are evolved, give rise to a current, although they would not do 

 so when arranged in circuit with a different electrolyte. To 

 test this view I tried 



Experiment 20. Oxygen and deutoxide of nitrogen, in 

 alternate tubes of the gas battery, with dilute nitric acid ; the 

 effects were, however, precisely similar to Experiment 9, viz. a 

 very feeble action for a few minutes, then a cessation, and no 

 continuous chemical action. 



Experiment 21. For the same reason oxygen and nitrogen, 

 with solution of sulphate of ammonia, were tried ; this arrange- 

 ment produced at first a slight effect upon the iodide, which 

 soon ceased, and after several days there was no more rise of 

 liquid in any cell of the closed circuit than in the detached 

 cell ; the rise of liquid in both was very trifling indeed (about 

 O'Oi cubic inch), and had evidently nothing to do with voltaic 

 action. In this experiment, and in every experiment that I 

 have tried, I have perceived a trifling action for the first few 

 minutes. This I should have attributed to accidental causes, 

 such as slight impurities in the gases, slight metallic deposits 

 on the plates, &c., but that it is always in the direction which 



