200 Mr. S. Ratner on Polarization 



of 1200 v. put across the side tube produced a normal dis- 

 cbarge through it. Since the pressure and nature of the 

 gas in the side tube was essentially the same as in the 

 main bulb, these experiments furnish a definite proof that 

 the phenomenon of polarization is not brought about by the 

 gases in the bulb. The effect must be traced back to changes 

 in the nature of the electrodes caused by prolonged electronic 

 action. 



Later experiments have shown, however, that even small 

 quantities of water-vapour exercise a marked influence in the 

 phenomenon. When a small quantity of this gas is intro- 

 duced into a polarized bulb, the polarization is at once 

 considerably reduced, and the bulb has then to go through a 

 long run before the previous stage is again reached. It thus 

 seems as if the electrodes were in some way affected by the 

 water-vapour losing the properties acquired during the 

 process of polarization. 



An attempt was also made to study the polarization effect 

 with different gases in the bulb. These experiments, however, 

 proved to be impossible when bulbs such as those described 

 in this paper are used, since the gases which are constantly 

 evolved during the discharge render those initially filling the 

 bulb impure. 



When observed in the dark, it is seen that the shape and 

 colour of the luminous column in the discharge undergoes 

 considerable changes as polarization sets in, but a spectro- 

 scopic study of this effect has not yet been made. 



Effect of change of Cathode. 



5. In the experiments described so far, only aluminium 

 cathodes were used. Since now, as has been shown, polari- 

 zation is caused by changes in the nature of the electrodes, 

 it seemed important to repeat some of the experiments, using 

 other metals as cathodes. Before starting these experiments, 

 difficulties were anticipated from the disintegration of the 

 cathode, and consequently in the first place copper was 

 chosen, as being less liable to sputter. The first experiments 

 were unsuccessful. A thick layer of sputtered metal was 

 soon deposited on the walls of the bulb, and after a run of 

 several days a well-marked layer was formed also on the 

 neck of the bulb behind the cathode. Finally the current 

 through the bulb was found to be conducted entirely by the 

 metal film connecting the anode with the cathode, and the 

 •experiments had to be abandoned. Later on, however, a 



