THE MOLECULAR PROCESS IN MACxNETIC INDUCTION. 259 



turning at once into the exact direction of the applied force, for there 

 was nothing (beyond a trifling friction at the pivot) to prevent it from 

 turning. 



Now try two magnets. I liave cut off" the current, >so that there is at 

 present no fiehl, hut you see at once thattlio pair has, so to speak, a will 

 of its own. I may shake or disturb them as I please, but they insist on 

 taking up a position (Fig. ."J) with the north end of one as close as i)os- 



H 



Fiii. :i. Fig. 4. 



sible to the sontli end of the other. If disturbed they return to it; this 

 configuration is highly stable. Watch what happens when the mag- 

 netic field acts with gradually growing strength. At first, so long as 

 the field is weak (Fig. 4), there is but little defection; but as the de- 

 flection increases it is evident that the stability is being lost, the state 

 is getting more and mcn-e critical, until (Fig. 5) the tie that holds them 



together seems to break, and they suddenly turn, with violent swing- 

 ing, into almost perfect alignment with tlie nuignefic tbrce II. J^ow I 

 gradually remove the force, and you see that they are slow to return, 

 but a stage comes when they swing back, and a complete removal of 

 the force brings tliem into the condition with which we began (Fig. li). 

 If we Avere to picture a ])iccc of ir>n as formed of a vast number of 

 such pairs of molecular magnets, each pair far enough from its neigh- 

 bors to be ])ractically out of reach of their magnetic intluence, we 

 might deduce many of the observed magnetic ])roi)crties, but not all. 

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



Fk;. 0. 



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

 new coniu'ctioiis when the old ones a-i'e l)rokcn; their lelations are of 

 a kind more complex thau the (|uasi-matrimoniai one which the experi- 



