154 Prof. J. J. Thomson on the 



an additional reason for supposing that the processes con- 

 cerned in metallic conduction are the same as those in electro- 

 lytic. For the opacity of thin metal films is enormously less 

 than that theory would indicate, if the conductivity of the film 

 for the very rapid electrical vibrations which constitute light 

 were the same as for steady currents. In this respect the 

 metals behave very much like electrolytes, for these act as 

 dielectrics to the light vibrations and as conductors for steady 

 or oscillating currents, provided the period of vibration is very 

 much greater than that of the light- vibrations. On the view 

 we have taken of metallic conduction, since the process of disso- 

 ciation and recombination takes a finite time, if the polarization 

 is reversed in less than this time, the old polarization will not 

 have had time to disappear before the new is superposed, and 

 the metal will, under these circumstances, behave more like 

 an insulator than a conductor. 



We can easily find an expression for the time T taken by a 

 tube of electrostatic induction to disappear (that is, to contract 

 to a length comparable with that between the atoms of a 

 molecule). Let E be the electromotive intensity at any 

 point, K the specific inductive capacity of the medium ; then 

 the number of tubes of electrostatic induction passing through 

 unit area is -^ 



47T 



Since T is the time taken by one tube to disappear, the 

 number of tubes which disappear in the conductor in unit 

 time is KE/47rT ; the number of tubes which disappear in 

 unit time is equal, however, to the current c through unit 

 area. Hence -j™ 



C = 47rr 



Thus 47rT/K is the specific resistance a of the conductor ; 

 hence, if ^K} be the electrostatic measure of K, we have 



T - — — - 



4tt 9 x 10 20 * 



The following table contains the values of T/K for a few 

 substances : — 



Silver 1*5 xlO- 19 



Lead 1-8 xlO- 18 



Water with 8*3 per cent, of H 2 S0 4 . 3*1 x 10" 13 



The value of K for anything like a good conductor has 

 never been measured ; but since substances which show the 



