266 EXPERIMENTAL INVESTIGATIONS. 



in the exposed battery, but none was perceptible in the 

 inclosed one, although the liquid had risen a little more, 

 viz. O'l cubic inch, in the hydrogen tubes of the latter. After 

 the four days above mentioned the jar of nitrogen which 

 covered the battery was taken away, and the action of the 

 battery was tested by iodide of potassium. At first there was 

 no action, but after about fifteen minutes a slight action was 

 perceptible ; this gradually increased, and in two hours the 

 action was equal to that of the battery which had been from 

 the first exposed to the atmosphere. I cannot but regard 

 this experiment as a conclusive negation of that view which 

 regards hydrogen and water as the efficient agents in the gas 

 battery. The opinion appears to me to have arisen from the 

 circumstance of our working always in an atmosphere con- 

 taining oxygen, and also from the fact of this latter gas being 

 more soluble than hydrogen.* If we lived in an atmosphere 

 of hydrogen, and if this gas were equally or more soluble 

 than oxygen, I have little doubt that the converse effects 

 would be observed. A battery charged with hydrogen in one 

 set of tubes and acidulated water in the alternate ones, at 

 first gives an effect nearly equal to an oxyhydrogen gas 

 battery, but the action rapidly declines in the former, while it 

 is constant in the latter. Even the ordinary action of the 

 gas battery when charged with oxygen and hydrogen appears 

 to me unanswerable as to the point I am now discussing. 

 When we see a battery of a number of cells at work, and the. 

 liquid gradually rising in the oxygen tubes, just in the pro- 

 portion in which oxygen gas is eliminated in the voltameter, 

 and when in a similar battery placed by its side, similarly 

 charged, but not connected in closed circuit, not the slightest 

 rise takes place in any tube, it seems impossible to adopt 

 the conclusion that the oxygen has nothing to do with the 

 current. Here we have no slight galvanoscopic effects, but 

 chemical effects capable of quantitative admeasurement, 

 capable of being continued to an extent only limited by the 

 size of the apparatus, and equivalent to the chemical effects 



* The tendency of oxygen to combine with platinum may also have its 

 influence. See M. De La Rive's various experiments on this subject, Bibl. Univ, 

 passim. 



