580 BELL SYSTEM TECHNICAL JOURNAL 



the gas? Without an impervious screen between the metal and the 

 gas this would, I suppose, be impracticable; but if there be such a 

 screen — if for instance the electrodes are outside of the glass-walled 

 tube containing the gas, instead of being inside — it can be achieved. 

 The ions in the imprisoned gas can be set into oscillatory motion by 

 an alternating voltage applied to electrodes outside; indeed, a self- 

 sustaining luminous glow-discharge can be maintained within the tube; 

 the charged particles which vibrate in the shining gas never enter 

 the metal parts of the circuit nor transfer their charges to these. 

 Only by invoking Maxwell's concept of the displacement current are 

 we able to contend that there is continuity of current-flow around the 

 circuit and across the gap between the anode and the cathode. 



The so-called "current through the gas" — i.e. the reading of the 

 galvanometer in series with the circuit — is therefore a datum decidedly 

 remote from the phenomena within the gas. Even in the simplest of 

 all cases, that of a direct-current discharge, it merely gives the sum 

 of the four terms aforesaid, not the value of any individually. In the 

 much more intricate case of the high-frequency discharge, it is a com- 

 bination of the four terms aforesaid and the displacement-current. 

 This is probably why it figures so little in accounts of the high- 

 frequency discharge. 



If the current through the gas is hard to interpret when measured, 

 the voltage across it may present an easier problem, or may on the 

 other hand be unmeasurable altogether. The simplest case of all is 

 that of a gas in contact with electrodes (as in the tube of Fig. 1) 

 between the two of which an undamped sinusoidal voltage is applied. 

 An alternating-current electrostatic voltmeter, of one type or another, 

 shunted across the electrodes, will give the "effective" or "root- 

 mean-square" value of the potential -difference between them. It is 

 often the maximum value of the voltage which the experimenter 

 wants, or thinks that he wants; to get it he must multiply the reading 

 of the voltmeter by ^^ This is not the proper factor unless the 

 voltage is truly sinusoidal; the observer must find out about this 

 (by using an oscillograph to make the waveform visible, or by hunting 

 with a wavemeter for harmonics of the fundamental frequency) and 

 use a different factor if the waveform is distinctly not a pure sine- 

 function. If the electrodes are outside of the tube containing the gas, 

 some part of the potential-gradient between them is located in the 

 insulating walls, and does not act upon the gas. If the voltage is 

 applied as a sequence of highly-damped short wavetrains with intervals 

 between, the problem of determining its maximum value is a serious 

 one. Perhaps the best available methods have never been applied in 



