120 Prof. Thomson, On the theory of Electric Conduction 



On the theory of Electric Conduction through thin metallic 

 films. By J. J. Thomson, M.A., F.R.S., Cavendish Professor of 

 Experimental Physics. 



[Read 4 March 1901.] 



The study of the electric resistance of thin metallic films 

 affords a very direct method of testing a theory of electrical con- 

 duction through metals which was developed by the writer in a 

 report presented to the International Congress of Physics at Paris, 

 in 1900. According to this theory the current through a metal is 

 carried by means of corpuscles, those small negatively electrified 

 particles which constitute the cathode rays ; which are given off 

 by incandescent metals and also by metals when exposed to ultra- 

 violet light. These corpuscles are assumed to be distributed 

 throughout the volume of all metals, being produced by the 

 corpuscular dissociation of the molecules. These particles, like 

 the particles of a gas, are supposed to be moving rapidly in all 

 directions, their kinetic energy, like that of the molecules of a gas, 

 being proportional to the absolute temperature. Under the action 

 of an electric field these charged corpuscles acquire a drift in a 

 definite direction — the opposite direction to the electric force — 

 since their charge is negative. This drift of the corpuscles under 

 the electric field constitutes the current through the metal. If n 

 is the number of corpuscles per unit volume of the metal, u the 

 velocity of drift in the negative direction of x, e the charge on a 

 corpuscle, then the intensity of the current parallel to the axes of 

 x is equal to neu. If X is the electric force, m the mass of a cor- 

 puscle, t the average time between two collisions of a corpuscle, 



1 Xe 

 u is equal to - — t; if \ is the mean free path, c the velocity of 



1 726^ A* 



mean square, t = \/c ; thus the current equals X ; and the 



1 ¥1$ A- 



conductivity of the metal is therefore -=- . The conductivity 



thus depends on the free path of the corpuscle. In the case of one 

 metal, bismuth, we have data which enable us to find the value of 

 this mean free path, which turns out in this case to be between 



