October 15, 1891] 



NATURE 



569 



and there is no residual eflfect on the configuration of the 

 group. But if we carry the process far enough to have 

 unstable deflections, the effects of these persist when the 

 force is removed, for the magnets then retain the new 



grouping into which they have fallen (Fig. lo). 

 And again, the quasi-elastic deflections whicli 

 go on during the third stage do not add to the 

 residual magnetism. 



stage is reached when instability begins, and then reversal 

 occurs with a rush. We thus find a close imitation of all 

 the features that are actually observed when iron or any 

 of the other magnetic metals is carried through a cyclic 

 magnetizing process (Fig. 12). The effect of any such 

 process is to form a loop in the curve which expresses 

 the relation of the magnetism to the magnetizing force. 

 The changes of magnetism always lag behind the changes 

 of magnetizing force. This tendency to lag behind is- 

 called magnetic hysteresis. 



\ye have a manifestation of hysteresis whenever a mag- 

 netic metal has its magnetism changed in any manner 

 through changes in the magnetizing force, unless indeed 

 the changes are so minute as to be confined to what I 

 have called the first stage {a. Fig. i). Residual magnetism 

 is only a particular case of hysteresis. 



Hysteresis comes in whatever be the character or 

 cause of the magnetic change, provided it involves sucli 

 deflections on the part of the molecules as make them 

 become unstable. The unstable movements are not re- 

 versible with respect to the agent which produces them ; 



Notice, further, what happens to the group if after 

 applying a magnetic force in one direction and removing 

 it, I begin to apply force in the opposite direction. At 

 first there is liitle reduction of the residual polarity, till a 



NO. I 146, VOL. 44] 



Fig. 12. — Cyclic reversal of magnetiiation in soft iron (aa), and in the same iron when 

 hardened by stretching (bb). 



that is to say, they are not simply undone step by step as 

 the agent is removed. 



We know, on quite independent grounds, that when 

 the magnetism of a piece of iron or steel is reversed, or 

 indeed cyclically altered in any way, some work is spent 

 in performing the operation— energy is being given to the 

 iron at one stage, and is being recovered from it at 

 another ; but when the cycle is taken as a whole, there is 

 a net loss, or rather a waste of energy. It may be shown 

 that this waste is proportional to the area of the loop in 

 our diagrams. This energy is dissipated ; that is to say, 

 it is scattered and rendered useless : it takes the form of 

 heat. The iron core of a transformer, for instance, which 

 is having its magnetism reversed with every pulsation of 

 the alternating current, tends to become hot for this very 

 reason ; indeed, the loss of energy which happens in it, 

 in consequence of magnetic hysteresis, is a serious draw- 

 back to the efficiency of alternating-current systems of 

 distributing electricity. It is the chief reason why they 



