296 Hon. R. J. Strutt on the Discharge of 



exceedingly minute quantity would be sufficient to account 

 for this. The nitrogen spectrum was visible only in the 

 positive column, not in the negative glow. The spectrum of 

 the latter showed helium only. 



While the gas was in this condition the cathode-fall was at 

 first about 350 volts. The battery- wires were removed and 

 the discharge from an induction-coil sent through the tube, 

 using the aluminium wire s as cathode. This wire was 

 strongly heated by the discharge. The aluminium became 

 soft, and fused up into a knob under the influence of surface- 

 tension. At the same time the aluminium spluttered on the 

 walls of the tube. The nitrogen was gradually absorbed by 

 the heated aluminium, and the nitrogen bands faded out from 

 the spectrum. At the same time the cathode-fall, which was 

 observed at intervals, diminished till it reached a value of 

 about 226 volts. The nitrogen was now no longer visible, 

 and the cathode-fall remained stationary. 



Since this cathode-fall was below the lowest graduation 

 of the voltmeter, an E.M.F. had to be added to the volt- 

 meter circuit. For this purpose twenty storage-cells were 

 used. Their E.M.F. was independently determined by means 

 of a Cardew voltmeter. This E.M.F. was subtracted from 

 the reading of the multicellular. The first set of readings 

 obtained was as follows: — 



265, 268, 271, 265 volts. Mean 267 volts. 



Each of these readings was obtained with a different resist- 

 ance in the main circuit : — 



More helium was now admitted. The cathode-fall at once 

 went up to something like 350 volts. But on allowing the 

 discharge from the coil to run for some time, the aluminium 

 absorbed the nitrogen, and the cathode-fall returned to its 

 original value. 



Second set of readings : — 



265, 268, 271, 272. Mean 267. 



The same process was repeated. 

 Third set of readings :— 



270, 265, 270, 265. Mean 267. 



The E.M.F. of the battery (20 cells) added in the volt- 

 meter circuit was determined before and after the experiment. 

 The values were the same, 40*7 volts. 



Next morning readings were taken with the same sample 

 of gas, to see if the value had altered. 



