N;o|>rrKI> l:V HOT PLATINUM IN DIFFKKKNT GASES. 



47 



was observed time after time with two different wires under varied conditions ot 

 temjKjratuiv and pressure. On the other hand, the absolute values of the leaks 



linallv "I. tainccl seemed very capricious. The ^em-nil rharactrr of these time changes 

 is exhibited by tig. 15. 



250 



200 



150 



c 

 O 



100 



I 



V = 80 



V=19 



V=286 



50 100 150 200 250 300 



. Time: Minutes 



Fig. 15. 



In order to make the experiments comparable, the wire was maintained at a 

 potential of 286 volts for 120 seconds before the readings under any assigned 

 voltage were commenced. For obvious reasons the voltage on the filament during 

 any one set of readings was never changed. 



The reduction in the negative leak by applying a big voltage does not occur under 

 the following conditions: (1) At very low pressures (<0 - 1 millim.), (2) at high 

 pressures (200 millims.), (3) when the wire is positively charged. The subsequent 

 increase in the ionisation occurs if the wire is either uncharged or charged positively 

 as wt'll as under a low negative potential. The time required for the establishment 

 ol'the linai equilibrium appears t<, decrease fairly rapidly as the temperature increiises. 

 Tin 1 voltages employed in the previous experiments were never great enough to 

 produce a luminous discharge, though the effect does occur if a luminous discharge 

 passes. The reduction in the ionisation is greater if the wire is made the cathode 

 than it' it is made the anode. 



