162 DR. FRANK HORTON ON THE DISCHARGE OF 



observed results. The discrepancies may be due to the unsteadiness of the negative 

 leak from lime which has been mentioned above. The leak was steadier than with 

 the calcium cathode, but not nearly so steady as with the glowing platinum. The 

 chief alteration of the leak was the gradual decrease as the heating of the cathode 

 was continued. This may have been due to a diminution of the amount of lime by 

 spluttering or by peeling off from the surface of the platinum, although no such 

 phenomena could be observed. It is not due to a decomposition of the lime by 

 electrolysis, for the author has shown* that no signs of electrolysis can be detected 

 when a current is sent through a vacuum tube from a lime cathode, and further 

 experiments with other lime cathodes showed that the negative leak decreased with 

 continual heating of the cathode, whether the discharge was passed or not. The 

 discrepancies between the observed and calculated values in Table V. follow from the 

 discordant values found for the constant Q. 



It has already been mentioned that a large series of experiments with lime 

 cathodes all gave similar results. In order to see if these irregularities were peculiar 

 to the present method of experimenting, the values of Q were calculated from 

 WEHNELT'S values of the negative leak from lime given in the ' Philosophical Maga- 

 zine' for July, 1905, p. 87. The variations in the values of Q thus found were 

 somewhat greater than those shown in Table V. The mean value was Q = 50,900 

 considerably less than the value found in the present experiments. It thus seems 

 that the negative leak from lime is subject to irregular variations, and does not obey 

 the Wilson-Richardson law with anything like the accuracy of the leak from 

 platinum. 



The fact that the negative leak from calcium is greater than from platinum at the 

 same temperature we have seen to be due to a decrease in the value of the constant Q, 

 that is, to a diminution of the energy required for the liberation of the corpuscles. 

 The value of Q for lime as found in the observations tabulated above is greater than 

 the value for calcium, and the fact that the current is greater in the case of lime 

 than in the case of calcium is due to the enormously greater value of the constant A 

 in the former case. 



A theory to account for the negative leak from hot metals has been proposed by 

 RICHARDSON, t He supposes the negative leak to be due to the escape of the cor- 

 puscles which, on the ionic theory of metallic conduction, all conductors contain. 

 The corpuscles are supposed to move about freely inside the conductor, and to have 

 a distribution of velocities the same as the molecules of a gas. Corpuscles entering 

 the surface layers of the conductor with a normal velocity component greater than a 

 certain amount are supposed to escape into the surrounding space, and it is these 

 corpuscles which maintain the current forming the negative leak. From these 

 assumptions RICHARDSON has deduced a formula of the type x = A^e"^ 2 *, and has 



* 'Phil. Mag.,' April, 1906, p. 506. 



t O. W. RICHARDSON, ' Phil. Trans.,' A, vol. 201, p. 497. 



