the Absurdity of Diamagnetic Polarity. 197 



absorbing or giving out heat except at some constant absolute 

 temperature 6. Also, for the sake of simplicity, we may 

 suppose the sphere connected with mechanism, on which it 

 does just so much work that its angular velocity of rotation 

 remains constant. Under these conditions the system will 

 assume an invariable state. If we then suppose that W is 

 the mechanical work done on the system in any time, and Q. 

 the amount of heat absorbed, the principle of energy will 

 give, since there is no change of state, 



W + Q=0. 



Now since work is obtained from the system, W is negative. 

 We see, therefore, that Q is positive. 

 But by Carnot^s principle, we have 



^<0, or Q<0. 



Thus Thomson's ideal experiment is inconsistent with both 

 the principle of energy and with Carnot ; s principle, and the 

 hypothesis of diamagnetic polarity is therefore disproved. 



The argument adopted in the two papers I have published 

 on diamagnetism in the Philosophical Magazine is only 9, 

 modified form of Thomson's, but possesses the advantage of 

 simplicity. It may be described thus : Take a piece of bis- 

 muth (or of any of the so-called diamagnetic substances) and 

 place it within the action of a strong steel magnet whose 

 poles are so far apart that only one of the poles, P say, exerts 

 any sensible force on the bismuth. Then let the steel magnet 

 be made immovable, and suppose the piece of bismuth held 

 by a pair of pincers, or in some other way, so that it can 

 either be kept at rest or moved about at will. Next, let the 

 air be exhausted in the space about the bismuth until the 

 " vacuum"" is perfect, and, lastly, let means be taken to pre- 

 vent the absorption or rejection of heat, except at some constant 

 temperature 6. 



If the piece of bismuth is held for some time in any posi- 

 tion X and then suddenly moved to a new position Y, 

 nearer to P, the magnetization of the bismuth will, of course, 

 become ultimately stronger in the position Y than it was in 

 the more remote position X ; but, since magnetization re- 

 quires time for its development, the change of position may 

 be effected so rapidly that the magnetization of the bismuth 

 has scarcely time to alter before the new position Y is reached. 

 We thus see that the amount of mechanical work which can 

 be obtained from the system of the steel magnet and the 

 piece of bismuth during a given change of position depends, 



