Temperature on Spontaneous Ionization. 021 



those produced by heating in the feature which is most 

 important for the present experiments, namely, the velocity 

 of the current near the walls. On the other hand, it should 

 be pointed out that the difficulty of getting the air in so 

 laro-e a vessel free from dust reduces the conclusiveness of 

 the experiment. 



20. (2) In the vessel shown in fig. 1 was placed a similarly 

 shaped vessel of iron wire gauze with a mesh of '15 cm. The 

 gauze was supported on slate blocks, and was about 1*5 cms. 

 distant from the walls of the vessel at each point ; it could 

 be maintained at any desired potential by means of a wire 

 passing through the lid of the vessel. The gauze acted as an 

 almost perfect electrostatic shield but permitted the passage 

 of ions ; when the walls and the gauze were at the same 

 potential of + 500 volts, the leak was 30 per cent, greater than 

 when the vessel was earthed and the gauze at + 500 volts. 



When the vessel was heated with the walls and the gauze 

 at the same potential, the changes observed were exactly 

 similar to those when the gauze was absent except that the 

 value of Q w r as reduced by about 25 per cent. But when the 

 gauze was at +500 volts and the vessel earthed or at a 

 potential differing still more from that of the gauze, the sign 

 of the changes was reversed : the leak on initial heating was 

 Jess, and that during initial cooling greater than the normal. 

 The value of Q is only about one-third of that observed in the 

 absence of the gauze, and Q' is notably less than Q. 



21. This difference is in entire accordance with the ex- 

 planation offered. Since the superficial area of the walls 

 of the tin vessel is much greater than that of the wire in the 

 walls of the gauze vessel, most of the particles will come from 

 the former and their charge will be determined by the charge 

 on the former. Since the P.D. between the walls and the 

 electrode is less than that between the gauze and the electrode 

 (it is the latter which determines the leak), the approach of 

 the charged particles to the electrode will produce changes 

 equivalent to a decrease of the leak. It will be seen that there 

 must be some value of the potential of the walls for which the 

 sign of the effects changes and their magnitude becomes zero. 

 The value could only be estimated roughly ; but it appears 

 that the effects change sign when the P.D. between the 

 electrode and the walls is about 100 volts less numerically 

 than that between the electrode and the gauze. 



22. (3) It was hoped that some light might be thrown 

 on the subject by investigating the leak through the vessel 

 immediately after the field was put on. A great many 

 observations were taken, but the results were so inconsistent 



