198 Mr. J. Parker on the Theory of Magnetism and 



among other things, on the speed at which the change of 

 position is made. Suppose then, if possible, that the bismuth 

 is repelled by the pole P, and let the system be made to 

 undergo the following cycle of operations in two different 

 ways : — 



(1) When the piece of bismuth is in equilibrium in any 

 position X at the uniform temperature 0, let it be moved to 

 some other position Y, nearer P, so slowly that at every point 

 of the path the bismuth is magnetized to the full extent and 

 the temperature constant. The work done on the system in 

 this operation will be positive, and may be written W. 



Then let the piece of bismuth return slowly from Y to X 

 by the previous path reversed. The whole of the work W 

 will thus be returned by the system, so that, on the whole, 

 there is neither loss nor gain of mechanical work during the 

 cycle. 



(2) Let the piece of bismuth be moved from X to Y so 

 rapidly that the magnetization has scarcely time to change 

 before the change of position is complete. The work that 

 must be expended for this purpose is obviously < W ; in fact, 

 if the position X be far enough away from P, the work done 

 on the system during the change of position may be practi- 

 cally zero. Next, let the piece of bismuth remain at rest in 

 its new position until its magnetization has attained its per- 

 manent value. Then let it be brought rapidly back from Y 

 to X by the first path reversed ; and, lastly, let it remain in 

 the position X until the original state is attained. The 

 mechanical work returned by the system in this way will be 

 > W. Thus, on the whole, there is a gain of mechanical 

 work during the cycle in which heat is absorbed and evolved 

 at a constant temperature. This being impossible, by the 

 principle of energy and Carnot's principle, we conclude that 

 the assumption that the piece of bismuth is repelled by the 

 pole of the magnet is absurd. 



The foregoing cycle was described in my first paper on 

 diamagnetism ; but the inference then drawn was, not that 

 diamagnetism is impossible, but that Carnot's principle does 

 not hold for diamagnetic bodies. In my second paper, I 

 denied the existence of diamagnetism ; but long before that 

 time Duhemhad arrived at the same conclusion, from reading 

 my first paper, and had written a book on the subject. 



I must now point out that Dr. Lodge has objected, with 

 justice, to the preceding reasoning, that it contains a tacit 

 assumption. Thus, even if we admit, as we are bound, that 

 the magnetization of every body takes time to grow and to 

 die away, it is not proved but that, in the second way of per- 



