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



We must now consider the two layers on the common sur- 

 face of two soft bodies A, B, A being a gas or liquid. If the 

 surface-molecules of A were provided with sharp points and 

 were caused by the smallest amount of magnetization to stick 

 to B, it might be thought that both surface-layers would then 

 belong to B ; but a little consideration tells us that the 

 molecules which stick to B would take with them two layers 

 of opposite signs, and it is clear that the remainder of A 

 would still have a surface-layer of its own, adjoining the 

 modified surface of B. In order, therefore, to cause the two 

 surface-layers both to belong to B, we must make the follow- 

 ing assumptions: — It must be supposed that every molecule 

 of A is provided with a sharp point, and that the act of 

 magnetization causes each molecule to be broken into two 

 halves, on one of which is the positive magnetism, on the 

 other the negative. Then it must be supposed that those 

 half-molecules on which are the surface-layers of A stick to 

 the bodies B . . . , with which they happen to be in contact, 

 and lastly, that the other half-molecules stick together in 

 pairs in such a way that their magnetisms neutralize one 

 another. In this way, we should have both surface-layers 

 belonging to B, and we might treat the free part of the gas 

 or liquid A as unmagnetized. A difficulty would, however, 

 arise when the magnetizing force was withdrawn, unless we 

 had some means of reminding the half-molecules to take 

 partners. We might avoid the difficulty by imagining the 

 two halves of each molecule tied together by a piece of thread, 

 but then we should introduce the absurdity that magnetiza- 

 tion changed the gas or liquid A into a solid. Lastly, we 

 should be obliged to conclude that when a system is once 

 magnetized, it is impossible to increase the magnetization — a 

 conclusion which, of course, is necessarily false. 



There is one other way of treating diamagnetism which 

 requires to be noticed. This is the method of induced electric 

 currents used by Weber, by which it is thought to be proved 

 that k is negative for bismuth and some other substances. 

 To this I reply that I have already sufficiently disproved 

 the common theory of diamagnetism; and secondly, it will be 

 proved in a future paper that the common theory of induced 

 currents generally involves an absurdity, and can seldom be 

 correct. 



The rest of the paper will be occupied with a brief dis- 

 cussion of a few important problems in the light of the new 

 theory. 



I. If a body P, placed near a number of fixed bodies X, Y, 

 Z, . . . . , be subject to no actions at a distance, but those of 



