86 Dr. AY. F. G. Sw 



inn on 



conduction which explains the phenomenon of conduction as 

 due to a kind of polarization of the conductor by the electric 

 field, resulting in the atoms pulling corpuscles out of their 

 neighbours to a greater extent in one direction than in 

 another (see J.J.Thomson, 'Corpuscular Theory of Matter/ 

 ]>. #6). The expression for the electrical conductivity i comes 



out as £= — • — -,. , where e is the electronic charge, d is 



the distance between the two charges constituting the doublet 

 which represent the atom, n is the number of molecules per 

 c.C, aO is the mean kinetic energy of a gas molecule at a 

 temperature 6, b is the distance between the centres of the 

 doublets, and p' is the number of corpuscles emitted by a 

 doublet per second. Taking a0 = 5 X 10~ 14 , ^ = 10--° E.M.U., 

 d=10- 3 , />=10" 8 , >i=l-Gxl(F, e^O-GxlO- 3 C.G.S. unit 

 (for silver), we readily find p'=10 17 , i. e. a quantity 10 l> 

 times as great as that which we have allowed ourselves in 

 the case of the quantity p. 



The current may be accounted for without bringing the 

 question of the expulsion of electrons from the atoms into 

 the matter. Let us confine our attention to those free 

 electrons which exist between the atoms: in the case of a 

 piece of matter in the resting state the average velocity, and 

 the mean free path, are independent of the direction. If 

 the matter be in motion, however, we may expect that these 

 quantities will vary according as they are considered in the 

 direction of motion of the matter, or in the opposite directs n. 

 The chances are that any dynamical scheme applicable to 

 the electron will lead to this result. Truly, with the electrons 

 considered as hard spheres, and with ordinary Newtonian 

 dynamics as applied to them, there is no apparent foundation 

 for such a belief, but the application of Newtonian dynamics 

 in its simplest form, involving as it does force- between the 

 electrons depending only on their relative positions, and 

 masses independent of the motions, is only an approximation 

 to the true state of affairs. In fact, even in the c - 

 imiionn translatory motion, the principle oi' relativity itself 

 involves a want of symmetry in the velocities, i'ov example, 

 of the kind here suggested, as may readily be -ecu by con- 

 sidering the case <>f a particle which, in a resting BVStem, 

 oscillates backwards and forwards between two points whose 

 j coordinates are ./•, and r s . If the time /, taken by the 



panicle in going from Xy to ./•., is the same as the time taken 

 n. going from .r 2 to ./•,, and if .?-,' and ./•._.' refer to the corre- 

 sponding system moving parallel to the axis ^\' .>• with 

 velooity v % then, writing <. ' for 1 - >■■ t "■'. it can be shown 



