104 



Prof. J. J. Thomson. 



[Mar. 9, 



I was not able to detect any change in the appearance of the spark 

 as the spark length passed through the critical value. My observa- 

 tion on the connexion between the appearance of the discharge and 

 the electrode at which the excess of hydrogen appears may be 

 summed up in the statement that when the discharge is plainly an 

 arc the hydrogen appears at the positive electrode, and when the 

 hydrogen appears at the negative electrode the discharge shows all 

 •the characteristics of a spark. However, before the spark length 

 reaches its critical value the discharge looks much more like a spark 

 than an arc. 



With regard to the quantity of hydrogen liberated from the steam 

 in comparison with that set free in the voltameter, I find that when 

 the spark length is a few millimetres greater than the critical length 

 the amount of hydrogen from the steam is very approximately the 

 same as that in the voltameter. The following table contains a few 

 measurements on this point : — 



Hydrogen Hydrogen 



Spark length. from steam. from voltameter. 



10 mm. 0*7 c.c. 0*8 c.c. 



12* „ 0-75 „ 0-9 „ 



14 „ 0-8 „ 11 „ 



When the sparks are longer than 14 mm. the amount of hydrogen 

 from the steam was no longer equal to that from the voltameter. 

 The results, however, were irregular, and, as mentioned before, there 

 was a considerable quantity of nitrogen (?) mixed with the hydrogen ' 

 and oxygen. 



When the sparks are very much longer, say about "22 mm., the 

 electrode at which the hydrogen appears reverses again, i.e., the 

 hydrogen comes off at the positive electrode, just as it does when the 

 sparks are very short. With these very long sparks the current is 

 extremely small, and it takes several hours to liberate 1 c.c. of 

 hydrogen in the voltameter. 



The proportion of hydrogen from the steam to that from the volta- 

 meter was with these long sparks too irregular to admit of any con- 

 clusions being drawn. 



The preceding results show that in the electrolysis of steam, as in 

 that of water, there is a very close connexion between the amounts of 

 hydrogen and oxygen liberated at the electrodes and the quantity of 

 electricity which has passed through the steam, and that this rela- 

 tion for certain lengths of sparks is the same in steam as in electro- 

 lytes. There is, however, this remarkable difference between the 

 electrolysis of steam and that of water, that whereas in the case of 



* In this experiment there was an air break 9 mm. long in series with the steam 

 tube. 



