the Absurdity of Diamagnetic Polarity. 201 



experiment. If the substance S be brought from X to Y, a 

 positive quantity of work, W\ say, will be obtained. Then if 

 S be turned round without altering the magnetization, the 

 force between P and S will become repulsive, and therefore 

 if 8 be brought back from Y to X without any change in the 

 magnetization, another positive quantity of work, W 2 say, 

 will be obtained. la this way, it might be thought, we could 

 prove that S could not be attracted by P, but we must 

 recollect that work will have been expended in turniug S 

 round. In fact, we may be sure that the work expended in 

 turning S round, first in the position Y, and then in the 

 position X, will be > W x + W 2 . 



II. Generally, if the substance M, when removed to a suffi- 

 cient distance from the pole P, is in the neutral state, it can 

 easily be proved that when M is near P, and its magnetiza- 

 tion in stable equilibrium, M must be attracted by P with a 

 force which increases as the distance from P decreases. Such 

 a substance is defined to be " perfectly soft/' and since 

 unstable states of magnetization do not occur in practice, we 

 may say that a " perfectly soft " substance can only be 

 attracted by a magnet pole. 



Of course, a substance S may be repelled by the pole P, 

 but in that case, it can easily be shown, the magnetization of 

 S could not be zero when it is at a great distance from P. 

 The substance S would then be a permanent magnet, or its 

 magnetization would consist of a permanent part combined 

 with a temporary part induced by the influencing pole P. 



III. If we now suppose the pole P replaced by a positively 

 electrified body and the piece of bismuth by a negatively 

 electrified body B, the force between the two bodies will be 

 attractive, and when B is moved from X to Y in such a way 

 that the distributions of the electrifications are in stable 

 equilibrium, the attraction on B in the position Y will be 

 greater than if the distributions had not been changed. 

 Hence if W be the positive quantity of work yielded by the 

 system when B is moved from X to Y in such a way that at 

 every instant the distributions are in stable equilibrium, then 

 when B is moved from its initial position X, in which the 

 distributions are stable, to the second position Y, without any 

 change in the distributions, the work yielded by the system 

 will be < W. Also if, after causing the distributions in the 

 position Y to become stable, we bring B back from X by the 

 previous path reversed without any change in the distribu- 

 tions, the work so done on the system will be > W. Conse- 

 quently, if the system is now caused to resume its original 

 state, a complete cycle will have been undergone, and a posi- 



