On the Kathode Fall of Potential in Gases. 363 



charge stopped. In the second experiment the current had been 

 running for some time through gas at a low pressure, the kathode 

 fall being 410 volts. A stream of pure gas sent it up to 570, but the 

 telephone soon began singing, and the kathode fall sank rapidly 

 to 480. The mean of the two observations is 582 volts. 



Nitric Oxide. A mixture of nitre and ferrous sulphate was acted 

 on by dilute sulphuric acid, and the evolved gas washed with potash 

 solution and sulphuric acid, and absorbed in a saturated solution of 

 ferrous sulphate. This solution was contained in a flask, sealed to 

 the rest of the apparatus. The gas was given off readily without 

 warming when the pressure was sufficiently reduced, and was 

 passed over potash and phosphoric anhydride before reaching the 

 discharge tube. 



In the case of water vapour and ammonia, when the current was 

 discontinuous the observed kathode fall was very variable and not 

 independent of the current strength, but with nitric oxide the 

 variation was within narrow limits, the readings always lying 

 between 340 and 380 volts, whether the telephone was silent or not, 

 and whatever the current strength, provided the kathode was not 

 covered with the negative glow. 



The readings of the kathode fall and the appearance of the dis- 

 charge showed that the gas is rapidly decomposed by the current. 

 When the discharge first started, the kathode fall was always near 

 370 volts, and the glow at both anode and kathode was white. In 

 a few seconds the glow round the kathode began to grow blue, and 

 that round the anode turned pink, whilst the kathode fall slowly sank 

 to about 345 volts. Meanwhile the glow spread backwards along 

 the kathode, showing an increase of current, and hence a decrease in 

 the resistance of the gas. 



The decomposition proceeded so rapidly that it was impossible to 

 get the kathode fall for the pure gas by taking the reading whilst a 

 stream of gas played on the electrodes, for the strength of stream 

 necessary to maintain the white glow of the pure gas was so great 

 that the pressure immediately rose high enough to stop the dis- 

 charge. Hence the only feasible plan was to allow the gas to 

 stream through the tube long enough to sweep out the products of 

 decomposition, then stop the stream by closing the mercury trap, 

 pump down to a suitable pressure, start the discharge, and take a 

 reading as quickly as possible. 



Twelve readings were taken in this way, when the telephone was 

 either silent throughout or became silent a few seconds after the 

 current started. There wa3 never any change in the kathcde fall at 

 the moment when the current became constant. The readings 

 varied between 366 and 378, the mean value being 373. 



This is so nearly the value for oxygen as to suggest a doubt of its 



2 D 2 



