PRODUCED BY HOT PLATINUM IX DIFFERENT OASES. 57 



same increase due to 10 ' milliin. as that obtained above. Reasons bave been adduced 

 i Mi-lier in this paper (p. 52) for believing that the assumption that the negative 

 iouisation is nearly proportional to the pressure of the hydrogen is incorrect. It has 

 been retained in the present argument because it is the assumption which is most 

 unfavourable to the view advocated. On the most unfavourable view, then, the 

 hydrogen diffusing through the wire should have produced an increase of at least 

 60 per cent, in the leak. Allowing an exj>erimental uncertainty of 20 per cent., no 

 change could be detected in the leak due to the diffusion of hydrogen through the 

 wire. This strongly supports the view, which also seems required on other grounds, 

 that the hydrogen does not act per se, but produces some change in the platinum 

 surface, and this change is prevented from taking place if the platinum is heated 

 in air.* 



In contrast to the negative ionisation, the positive ionisation produced by the hot 

 platinum tube was found to be altered when hydrogen was allowed to diffuse through 

 from inside. In fact, at constant temperature an additional amount of ionisation is 

 caused thereby which is proportional to the amount of hydrogen diffusing through 

 the tube. 



At high temperatures the positive ionisation was found to be readily saturated. 

 Thus at 1200 C. the leak with -|-80 volts was equal to 64 divisions, and with 

 + 400 volts 75 divisions with air inside the tube ; with hydrogen diffusing through, 

 the values under these voltages were 88 and 95 divisions respectively. These 

 proportions were much the same, so the leak was generally measured with +80 volts, 

 as higher voltages were not always available. On changing from a low to a high 

 voltage, a big leak was often noticed at first, but this always fell away in a few 

 minutes, until, approximately, the above ratio was obtained. Effects of this kind 

 have already been described in detail (see 9). At low temperatures the positive 

 ionisation obtained with this apparatus seemed to be different in character, for it was 



[* Note added Septemlrr 7, 1906. It seems advisable to indicate the exact bearing of this argument 

 more precisely. It is intended to confirm the conclusion, which has Ixsen drawn from direct experiment 

 on page 52, that the high value of the negative ionisation in a vacuum containing traces of hydrogen is 

 due to the fact that the hydrogen keeps the surface of the wire in a certain state, rather than that the 

 wire contains a certain amount of hydrogen. The writer does not wish to create the impression that 

 hydrogen never exerts a direct influence on the magnitude of the negative leak. The numbers on 

 p. 53 show that the value of the ionisation, at constant temperature, increases with the pressure of 

 the hydrogen at high pressures ; so that it is probable that at high pressures there is a negative leak 

 which is a definite function of the pressure of the hydrogen. This is also demanded by the fact that the 

 constant o_, which enters into the temperature formula, is dependent on the pressure. 



A comparison of the table at the top of page 63 with that on page 52 would seem to indicate that the 

 ionisation in hydrogen at 1'9 millims. pressure is much greater than in a hydrogen vacuum. The writer 

 considers, however, that these experiments are not comparable with one another, owing to the wires having 

 lM?en differently treated before the two experiments. The direct testa made on pp. 50 and 51 showed that 

 decreasing the pressure from about 1 millim. to -001 millim, only reduced the ionisation in hydrogen by 

 about 40 per cent, of its value.] 



VOL. CCVII. A. I 



