Magnetic Properties of Iron and Nickel. 23 



the molecule of iron itself, is capable of motion of rotation and 

 of motion of translation. At ordinary temperatures the mag- 

 netic molecules have their centres »i 1? m 2 , m z so close together 

 that though they are endeavouring to form a closed magnetic 

 circuit they are unable to do so. If now we apply a magnetiz- 

 ing force in any direction, each magnetic molecule will be 

 rotated about its axis, and the whole iron molecule will be 

 magnetized and rotated about its axis. When the magnetizing 

 force is removed both the magnetic molecules and the iron 

 molecules will return more or less completely to their original 

 positions. On raising the temperature the centres of both the 

 iron molecules and the magnetic molecules will increase their 

 distances from each other until a certain temperature is reached, 

 when the opposite poles of adjacent magnetic molecules will 

 come into such close proximity that they will be able to rush 

 together so that the system of magnetic molecules forms a 

 close magnetic circuit, as in fig. 2. When the magnetic 



Fig. 2. 



molecules are in this position a small magnetizing force in 

 any direction will have no appreciable effect ; but by employ- 

 ing larger and larger magnetizing force we may obtain 

 sensible evidence of magnetization, as indeed Faraday, Baur, 

 and the author have found. The suddenness of the loss 

 of magnetic properties is easily comprehended when we 

 remember the rapidity with which magnetic attractions 

 increase as the distance between magnetic poles diminishes. 

 We can also understand that heat at this point should become 

 latent, for it is evident that the sudden bringing together of 

 N and S poles, which takes place on heating, will cause a 

 sudden expansion of m x , ra 2 , an d m 3- Since the action goes 

 on inside the iron molecule there may be no sensible sudden 

 increase of distance between the centre of one iron molecule 

 and another, and no sensible change in the molar dimen- 

 sions of the iron. This agrees with the fact ascertained by 



