PRODUCED MY HOT PLATINUM I\ DIFFKRKNT GASES. 



21 



This apparatus, in which effects due to heating of the walls were prevented, was 

 found to give exactly the same kind of results as the earlier experiments. For 

 instance, the saturation current in oxygen at different pressures was found to have 

 the following values : 



The temperature in this experiment was 1)76 C. and the unit of current 

 7xlO~ 13 ampere. The observations were taken in the order of the numbers in 

 the table. 



Several other points were tested with this apparatus, one of which was to see 

 whether the lag in the leak behind changes the pressure still held. The wire was 

 giving a minimum leak of 36 divisions under a pressure of 0/64 millim. when the 

 oxygen was pumped out as rapidly as possible to a pressure of 0*004 millim. The 

 temperature was then adjusted to its former value, and the following minimum values 

 of the leak were observed at the times stated, the time being reckoned from the 

 point at which the temperature first became steady. 



Evidently the wire requires time to adjust itself to the changed conditions, so that 

 this effect cannot be attributed to anything given off owing to the walls becoming 

 heated. 



The irregular changes in the leak previously noticed seemed to occur in this vessel 

 to about the same extent as in the others, so they also cannot be ascribed to anything 

 from the walls of the tube. 



Some puzzling effects which have been observed may, however, probably be assigned 

 to this cause, and it seems advisable to mention them for the benefit of other workers 

 iu this subject. The writer has several times obtained a large increase of the leak 

 with the pressure at high pressures, especially with the wire at a high temperature. 

 This effect has, however, only been found to occur when the whole tube became very 

 hot and it could be reduced to a small value by simply blowing cold air on to the 

 outside of the tube. Another effect which probably arises in the same way is an 

 increase in the leak at a given temperature produced by heating the wire for a short 

 time to a c>nsi(lfral)ly higher temperature (see p. 13). Both these effects appear to 

 lie really due to the walls of the tube becoming heated. 



