The Mode of Conduction in Gases. 657 



extract from it an electron apiece, thereby becoming neutral 

 again. 



' ; A constant diffusion of uncharged atoms towards the 

 anode and of positively charged ions away from it is therefore 

 essential to the passage of current." 



It is difficult to see how these effects at the anode are 

 essential, as electrodeless discharges are easily produced or 

 a discharge from a negative point in which there is but a 

 small force at the surface of the positive electrode, and conse- 

 quently no positive charge would be communicated to the 

 atoms ; but these difficulties are not the most formidable 

 objection to this part of the theory. 



It may perhaps be of interest to students who are confused 

 with the multitude of these theories, if I mention an expe- 

 rimental fact that is known to and accepted by the majority of 

 physicists. JVo positively electrified particles are given off from 

 a positively charged surface at. ordinary temperatures, either 

 molecules, atoms, or corpuscles of the gas in contact unth the 

 metal, or particles emanating from the metal, no matter how 

 great the electric force is at the surface. The only exceptions 

 occur in the case of a particles from radioactive substances. 

 It is possible to have a much higher electric force at the 

 surface of a metal than occurs in an ordinary discharge-tube, 

 and no appreciable current can be detected by sensitive 

 apparatus. An experiment on this subject by J. E. Almv 

 (Philosophical Magazine, September 1908) may be mentioned. 

 He found that a current does not pass through air at atmo- 

 spheric pressure between two electrodes at a distance apart 

 of the order of a wave-length of light when the potential 

 difference is 300 volts. If particles of the gas were to 

 become charged by contact with an electrified surface, the 

 phenomena should have been observed under these conditions, 

 as the force at the surface of the metal was of the order of 

 10 6 or 10 7 volts per centimetre. 



John S. Towxsend. 



Electrical Laboratory, Oxford. 

 26th August, 1U11. 



Reply hy Sir Oliver Lodge. 

 It is plain that Professor Townsend regards with favour, or, 

 shall we say, holds quite dogmatically, the opinion that 

 electric convection in a partial vacuum is effected entirely 

 by negatively charged carriers. That, or some approach to 

 it, is the natural supposition, and may possibly still turn out 

 to be true ; but recently I have not found it explanatory of: 

 the phenomena exhibited by electric vacuum-valves, and 

 accordingly I have suggested another hypothesis for consider- 



