6 MR. O. W. RICHARDSON ON THE IONISATION 



temperature. In respect of falling away with time, the initial leak offers a very 

 marked contrast to the part of the leak which depends on the surrounding oxygen. 

 This was found to remain constant, except for a temporary variability, under the same 

 conditions during the whole of the time the experiments were being carried out. 

 The experiments on this particular wire lasted about three months and during that 

 time it was heated at various high temperatures for about 150 hours. 



In the sequel it will be shown that as the pressure of the oxygen is increased the 

 current from the positively-charged wire asymptotically approaches a maximum 

 vulue. It is interesting to compare this value, which is independent of the pressure, 

 with the initial leak at the same temperature. The experiments show that at 804 C. 

 the greatest positive leak from an old platinum wire of the above dimensions in 

 oxygen = 3'6xlO~ n ampere, and is therefore about one five-hundredth part of the 

 initial leak from a new wire. 



Another source of trouble in these experiments arises from variations which take 

 place in the leak when all the controllable conditions are kept constant. These 

 variations, which will be considered more fully in the sequel, appear to fall into two 

 classes. The first are of a hysteretic nature and depend on the previous treatment 

 of the wire. For instance, if the gas pressure is suddenly lowered the leak does not 

 decrease immediately, but only gradually settles down to its final steady value. 

 Increasing the pressure gives rise to the converse effect. Sudden changes of 

 temperature, and in some cases of potential, will be shown to give rise to similar 

 hysteretic effects. These changes can be explained by supposing that the leak is due 

 not to the external gas, but to oxygen, which is held chemically or otherwise in the 

 superficial layers of the platinum, and that the amount necessary for equilibrium 

 takes time to adjust itself. 



The second kind of variation seemed to be of a purely irregular nature, and 

 manifested itself by sudden jumps in the rate of movement of the electrometer spot 

 across the scale. This effect was specially marked at high pressures, and at low 

 pressures was not so noticeable. It may be due to the pressure of the contained 

 oxygen becoming great enough to force a way through the overlying layer of 

 platinum, and so giving rise to a sudden evolution of highly ionised gas. 



To eliminate errors due to effects of the first kind, some time was always allowed 

 to elapse after each change had been made, and readings taken from time to time 

 until the leak became steady. At high pressures, where irregularities of the second 

 kind occurred in addition, it was more difficult to obtain the equilibrium value of the 

 leak ; but by neglecting all readings where the movement of the spot was noticed to 

 be jerky, by always taking the smallest values of the leak, and by waiting till two or 

 more of these were identical, consistent results could always be obtained. This 

 procedure was rather tedious in some cases, but it seemed to be the only method, as 

 the irregularities concerned were not affected by any change in the controllable 

 conditions (except by diminishing the pressure). 



