56 MR. O. W. RICHARDSON ON THE IONISATION 



diffusing through the wire would have been rather less than one-third of what it was 

 in the former. A calculation based on the results of RICHARDSON, NICOL, and 

 PARNELL* showed that the amount of hydrogen diffusing through each square 

 centimetre of surface of the platinum per minute must have been equal to about 

 2 cub. centims. at C. and 760 millims. 



Since the platinum tube in these experiments was giving the small negative 

 ionisations normally produced in air before the hydrogen was allowed to diffuse 

 through, the above experiments prove indisputably that the negative ionisation 

 produced by hot platinum in air and other gases is not due to residual traces of 

 absorbed hydrogen. It appears to be possible to go further than this and to say that 

 the effect, on the leak, of hydrogen inside the metal is not due to its direct action as 

 hydrogen, but to some change it produces in the properties of the metal surface. 

 This change appears to be inhibited when the metal is heated in an atmosphere of air 

 or oxygen. 



To substantiate this conclusion it is necessary to prove that there was enough 

 hydrogen inside the outer surface of the wire during the experiment to have 

 appreciably altered the value of the negative leak if it were exerting its full effect. 

 This may be done by finding a minimum value of the external hydrogen pressure 

 which would just stop the diffusion outwards for an instant. If the velocities of 

 the escaped hydrogen molecules were suddenly reversed, the diffusion would stop 

 momentarily, and the external pressure then occurring would give the minimum 

 external pressure which would keep the hydrogen inside the surface layer in equi- 

 librium. The equilibrium pressure might be greater than this, but could not be less. 

 In the experiment the mass of hydrogen diffusing through 1 sq. centim. per second at 

 1380 C. = 2'65xlO~ 7 gramme. In free hydrogen at 1 centim. pressure the mass 

 which is carried across an area of 1 sq. centim. per second = 2'3xlO~ 2 gramme at 

 1380 C. Hence the minimum value of the external pressure with which the 

 hydrogen instantaneously present inside the surface layer could be in equilibrium is 

 1*15 x 10~ 4 millim. It is necessary to show that a pressure of this amount of hydrogen 

 would have produced an appreciable increase in the value of the leak. According to 

 one table given by H. A. WILSON (loc. cit., p. 265), at 1350 C. hydrogen at 6x 10~ 4 

 millim. pressure increases the negative leak by a factor of 2500. The writer, 

 however, is inclined to think that a more accurate comparison with the present 

 experiments can be got by comparing the tables on pp. 260 and 269 of WILSON'S 

 paper. These show that hydrogen at a pressure of 13xlO~ 4 millim. increases the 

 leak at 1375 C. by a factor of 8, so that l'15x 10~ 4 millim. would cause an increase 

 by a factor of not less than 1'6. The experiments recorded in the present paper 

 (p. 52) also indicate an increase in the leak of about 100-fold at 1340 C. due to 

 hydrogen at a pressure certainly less than 10~ a millim. On the assumption that the 

 negative ionisation is nearly proportional to the pressure, this would give about the 



* ' Phil. Mag.' [6], vol. 8, p. 1. 



