.Disappearance of Gas in the Electric Discharge. 587 



with the cathode fall of potential. The connexion, though 

 interesting and suggestive, does not seem a sufficient basis 

 for any theory of the process, although the workers named 

 argue that it proves that the absorption of gas cannot be 

 due to chemical action between the gas and the anode or 

 the glass walls. 



However, the connexion between absorption and cathode 

 fall cannot be universal, for the cathode fall cannot be 

 abolished by causing the thermionic emission of electrons 

 from the cathode ; and yet gas will undoubtedly disappear 

 when the discharge passes from an incandescent cathode 

 under a potential much less than the minimum spark 

 potential which would be required to send a discharge 

 through the gas if the electrode were cold. Indeed, it is 

 the disappearance of gas in the presence of a hot cathode 

 that has the greatest technical importance in connexion 

 with lamp and valve manufacture, and has led in recent 

 years to an increased interest in the matter. 



But it must be noticed that the introduction of an 

 incandescent cathode may effect the disappearance of gas 

 otherwise than by altering the electrical conditions in the 

 discharge ; in the absence of the discharge there may be 

 some direct interaction between the hot metal and the gas 

 which does not take place when the metal is cold. These 

 two influences of the incandescent cathode are closely 

 associated in all experiments, and may even be hard to 

 distinguish ; for since the cathode, usually in the form of 

 wire or strip, is almost always heated electrically, there 

 will be a potential difference between its ends, and this 

 potential difference may be great enough to cause a dis- 

 charge between the ends, one of which acts as cathode and 

 the other as anode. However, some observations described 

 below, the results of which are confirmed by other work, 

 seem to show that the two influences can be separated ; 

 for, if the potential difference between the ends of the wire 

 is sufficiently small, then, though it must cause some 

 electric discharge through the gas, the discharge which it 

 causes is not of a nature to influence the disappearance of 

 gas. Accordingly, if the filaments are so stout or so short 

 that the difference of potential between their ends when 

 they are incandescent does not exceed this limit (which is 

 apparently always greater than 20 volts), any action which 

 the filament may have when it is heated must be due 

 purely to its temperature and not to the electric field 

 sui rounding it. 



This purely thermal influence of an incandescent filament 

 2 Q 2 



