1896.] on Electric Besearch at Low Temperatures. 267 



state, and makes it almost impossible to obtain similar repeated 

 measurements. 



It is natural to inquire how far accepted theories of magnetic 

 action are able to reconcile the above-mentioned results. Some of 

 them undoubtedly are in accord with deductions from received 

 hypotheses. It is generally considered that the facts connected with 

 the magnetisation of iron indicate that each molecule, or perhaps 

 small groups of molecules, of the iron are complete micro-magnets, 

 and that in the unmagnetised condition of the iron these molecular 

 magnets arrange themselves in groups or in closed circuits so that 

 for each little group the external magnetic action or magnetic moment 

 is approximately zero. Magnetisation consists in arranging the 

 members of some or all of these groups so as to co-lineate the direction 

 of more or less of the molecular magnets and produce an external 

 resultant magnetic moment. 



Let us then consider one such little group by the aid of a model 

 made of small magnets, such as Ewing has suggested and used. 



SujDpose the members of this group to be at a certain distance 

 from each other, and we apply a given magnetising force which is 

 just sufficient to open out the group and co-lineate the magnetic 

 axes of the several members of it. 



Next, suppose we cool this iron, this would result in bringing 

 the members of the group into closer contiguity. The result of this 

 will be an increase of the interpolar magnetic forces of the different 

 members of the group ; and as we can see from the behaviour of the 

 model, it would require a greater magnetic force to effect the same 

 amount of co-lineation of the molecular magnets. This, therefore, 

 corresponds with what we find to be the case on cooling soft iron to 

 very low temperatures. Professor Dewar's experiments have shown 

 that the tensile strength of iron and steel is increased to about 

 double on cooling to — 182° C, and it is quite reasonable to suppose 

 that this is the result, in part at least, due to an approximation of the 

 molecules. 



As regards the behaviour of magnetised steel and iron when cooled, 

 it is highly likely, when the groups of molecular magnets have been 

 opened out more or less, that some of these are in a condition of insta- 

 bility, in which bringing the members of the group nearer together will 

 have the effect of making them close up again into magnetic circuits 

 of no external action. Hence, if this is the case, the first effect of the 

 sudden cooling will be to effect the observed change. These half- 

 hearted groups of molecular magnets constitute the subpermanent 

 magnetism which it is our desire to get rid of in ageing a magnet. 

 Then, as regards the effect of temperature changes on the magnet 

 when the stable condition of affairs is reached. In order to explain 

 this, I think we must consider the action of the molecular groups 

 upon each other. The approximation of molecular groups will in 

 general, after the magnet is aged, have the effect of co-lineating more 

 completely the different members of the groups, and hence increase 



