392 



Professor J. A, Eiinng 



[May 22, 



hours to be practically out of reacli of tlieir magnetic influence, we 

 might deduce many of the observed magnetic pro2:)erties, but not all. 

 In particular, we should not be able to account for so much residual 

 magnetism as is actually found. To get that, the molecules must 



Fig. 5. 



make new connections when the old ones are broken ; their relations 

 are of a kind more complex than the quasi-matrimonial one which 

 this experiment exhibits. Each molecule is a member of a larger 

 community, and has probably many neighbours close enough to atiect 

 its conduct. 



We get a better idea of what happens by considering four magnets 

 (Fig. 6). At first, in the absence of deflecting magnetic force, they 

 group themselves in stable pairs — in one of a number of possible 



Fig. 7. 



combinations. Then — as in the former case — when magnetic force is 

 applied, they are at first slightly deflected, in a manner that exactly 

 tallies with what I liave called the stage a of the magnetising process. 

 Next comes instability, 'i'he original ties break up,*^and the magnets 



