168 Prof. J. J. Thomson on the 



field is in a steady state we have the positive tubes moving with 

 equal velocity but in the opposite direction to the negative 

 ones ; in this case 2w = 0, and there is no electromotive force 

 due to the motion of the tube. When the current is first 

 started the positive tubes move in before the negative ones 

 begin to move out, so that in this case %u is positive, and 

 hence the electromotive force is negative, i. e. in the opposite 

 direction to the current. When the current is suddenly 

 stopped, the inward flow of tubes is stopped, but the out- 

 ward one continues for some time ; in this case %u is negative 

 and the electromotive force is therefore positive, i. e. in the 

 direction of the original current. 



We shall now consider, from the point of view of this theory, 

 experiments such as those made by Rowland on the magnetic 

 effects produced by the rotation of electrified disks. Let us 

 consider a parallel plate-condenser rotating with an angular 

 velocity co about its axis, which we shall take as the axis of z. 

 Let h be the electric displacement parallel to z; then the com- 

 ponents of the magnetic force due to the motion of the tubes 

 whose ends are on the plates of the condenser are given by 



a = 4:7rcoxh, 

 /3 = 4z7ra)yh, 

 y = 0. 

 These values do not satisfy the solenoidal condition 



dx dy dz 



So that inside the condenser we must have, in addition to the 

 motion of those tubes which end on the plates of the con- 

 denser, a system of closed tubes in motion such that, while 

 they do not alter the electric displacement, they alter the 

 magnetic force. Thus we must have positive tubes moving 

 in one direction and an equal number of negative ones 

 moving in the opposite. If the motion of these tubes were 

 confined to the inside of the condenser there would be no 

 magnetic force outside the condenser, as the tubes of electro- 

 static induction outside are then all at rest, and it would follow 

 that in this case there would be no magnetic force outside or 

 inside. As this is not consistent with the results of Rowland's 

 experiments, we shall suppose that these closed tubes pass 

 right through the condenser, their motion being continuous 

 from the inside to the out. The magnetic force due to the 

 motion of these tubes will therefore be continuous as we cross 

 the plates of the condenser. In consequence, however, of the 



