the Absurdity of Diamagnetic Polarity. 267 



We can now explain Weber's hypothesis of magnetism. 

 He considers that a body which appears to be neutral is as 

 much magnetized as when it exhibits active magnetic pro- 

 perties, only that, in the former case, the magnetized mole- 

 cules have their axes pointing in all directions so as exactly to 

 neutralize one another. He then supposes that the act of 

 magnetization merely consists in giving the axes of the mag- 

 netized molecules a definite direction. In fact, if we suppose 

 an elementary magnet suspended freely by the centre of mass, 

 it is clear that it will set its axis in the direction of the external 

 magnetic force which acts upon it. 



In the common theory of magnetism it is admitted that in 

 a " perfectly soft " homogeneous substance, the magnetiza- 

 tion at any point is in the same straight line as the force ; 

 but it is supposed that in the so-called diamagnetic homo- 

 geneous "soft substances/' ^, or ^ is negative, or that 

 the magnetization is in the opposite direction to the 

 force. This gives rise to a difficulty in Weber's theory ; 

 for it appears to follow that when an elementary magnet 

 is freely suspended by the centre of mass, it may perma- 

 nently set its axis in the opposite direction to the external 

 magnetic force. To escape from this difficulty it might be 

 assumed that in every diamagnetic body a number of mole- 

 cules form a kind of lock-work, similar to that of a gun, 

 and that the first act of the external magnetizing force 

 is to set the lock. In this way, it might be thought, we 

 should have a means of setting the magnetized molecules in 

 the opposite direction to the force and keeping them there ; 

 but it would follow that a diamagnetic body could not be mag- 

 netized until the external magnetizing force exceeded a certain 

 value, and would not lose its magnetization when the force 

 was withdrawn. As this appears to be contrary to experi- 

 ment, we conclude that our explanation of the difficulty due to 

 the common theory of diamagnetism must be insufficient. 



In the case of a quasi-homogeneous substance, like air or 

 any gas, the mass-density will vary from point to point. 



Denoting the mass-density bv p, and putting 1= — I', where 



jPo 

 p is a standard fixed value of p, the state of the air or gas at 

 any point may be defined by the three variables {I', p, 0). 



If we put A=-£ A', &c. &c, it is easy to see that equations 



Po 

 (5), which hold for the stable distribution of magnetization^ 

 become 



T2 



