SOO Hon. &. J. Strutt on the Discharge of 



A thiid sample of the gas gave the value 



165 volts. 



It was not possible to submit this sample of gas to prolonged 

 sparking owing to an accident, by which air gained admission 

 to the tube. 



Taking the mean of all these measurements, we have as the 

 cathode-fall in argon : — 



167 volts. 



It was mentioned above, p. 293, that one condition for a 

 normal cathode-fall is that there should be no chemical 

 action between the cathode and the gas. This statement is, 

 however, of a somewhat hypothetical nature. It was found 

 by Warburg that the cathode-fall in hydrogen was nearly the 

 same (300 volts) over electrodes of copper, zinc, iron, and 

 platinum. But for aluminium and magnesium cathodes it was 

 much less (about 180 and 160 volts respectively). It seems 

 most natural to attribute the exceptional behaviour of these 

 latter metals to chemical' action o£ the gas, or of traces of 

 impurities in it, upon them. However this may be it 

 appeared to be of interest to determine whether argon would 

 show any similar effect. A few experiments were made to 

 test this point. 



A disk cathode of aluminium was employed. It spluttered 

 freely in the argon-filled tube. The values obtained for the 

 cathode -fall were 



97, 90, 85. 



By continual sparking the value rose somewhat higher. The 

 readings were 



100, 105, 105, 115. 



The mean value for the cathode-fall in argon over aluminium 

 electrodes is 



100 volts. 



Thus, so far as these experiments go, the cathode-fall in 

 argon appears to be diminished in the same ratio as that in 

 hydrogen, when aluminium electrodes are substituted for 

 platinum. 



Electrodeless Discharge. 



Interesting effects are observed when an endless ring dis- 

 charge is produced in rarefied argon by discharging a 

 leyden-jar through a coil of wire, inside which a bulb con- 

 taining the gas is placed. Discharges of this kind have been 



