On the Kathode Fall of Potential in Gasea. 361 



difference between the measurements before and after the addition 

 was very small. 



It is not necessary to give a detailed list of all the observations 

 that were made. It will be sufficient to say that before the bichromate 

 was added twenty-five observations were made, the mean value found 

 for the kathode fall was 231 volts, the extreme values being 223 and 

 241. After the addition of the bichromate twenty measurements 

 were made, the mean value being 233 and the range 225 to 238. 

 The mean of the whole is 232. 



This is exactly the same as Warburg found for atmospheric 

 nitrogen, whence it appears that the presence of argon has no effect 

 on the kathode fall. 



Water Vapour. Distilled water from a clean silver still was 

 boiled to remove dissolved gases, and the apparatus was sealed up 

 whilst the water was still at the boiling temperature. 



The greater part of the air was removed from the apparatus by 

 means of the pump, which contained some sulphuric acid above the 

 mercury, and the whole was then allowed to stand for a week, the 

 acid absorbing the vapour and keeping up a steady evaporation of 

 the water, thus gradually sweeping out any remaining air. The un- 

 condensed gas was ejected from the pump, and the acid renewed 

 from time to time. Incondensable gas never entirely ceased to pass 

 over, but before any experiments were made it appeared from an 

 estimate of the volume of water absorbed by the acid and the volume 

 of the air bubble in the pump, that the ratio of the pressure of the 

 air to the pressure of water vapour in the apparatus was reduced to 

 about one part in five millions. Of course, more gas came over when 

 the discharge was passing, but even then the quantity was very 

 small. Presumably the hydrogen and oxygen set free near the 

 electrodes recombine at other parts of the tube. 



A compound gas naturally presents difficulties that are absent in 

 the case of an elementary gas, for as soon as the current is started 

 the gas becomes mixed with decomposition products, and is no longer 

 pure. The chief difficulty, however, arose from the intermittence of 

 the current. With elementary gases the current was seldom inter- 

 mittent, with the compound gases it was seldom constant. It was, 

 of course, useless to take any readings of the kathode fall when the 

 telephone was singing, and many [months were spent in a fruitless 

 attempt to find what circumstances determine the constancy or inter- 

 mittence of the current. 



The rate of intermittence was not usually constant, as was evident 

 from the variation in the note given by the telephone, and the 

 change from sound to silence was sudden, and often accompanied by 

 other changes, such as the appearance of striae in the tube, and a 

 sudden change in the kathode fall. When the telephone was silent 



