Model of Ferromagnetic Induction. 499 



is weak because it depends on a nearly perfect balance 

 between forces that exert opposing moments, though these 

 forces are separately strong. 



Suppose, as an ideal case, that the opposing actions within 

 the atom are exactly balanced : in other words that each 

 Weber element turns between fixed elements which exert 

 precisely equal magnetic forces on its poles. In that 

 -extreme case it will still be stable when the atom stands as 

 one of a row of atoms in the crystal. For the Weber 

 elements in the successive atoms of the row will still exert 

 mutual forces on one another — across a relatively large gap — 

 and will tend, as in my model of 1890, to set themselves in a 

 row. Their mutual action will give them a stability which 

 is feeble because of their wide spacing ; but there will be 

 this important difference from the model of 1890, that the 

 presence of fixed magnets in the gap, though by hypothesis 

 it contributes nothing to the resultant stability, limits the 

 range of stable deflexion. 



It is open to conjecture that in an ideal crystal of perfectly 

 pure and perfectly unstrained iron the action of the fixed 

 magnets may be perfectly balanced, and remain balanced 

 when the Weber element turns. In that case the stability of 

 the Weber elements would be due only to the forces which 

 they exert on one another from atom to atom. On the other 

 hand, there may be even in pure unstrained iron an excess 

 in the force between unlike poles over the force between like 

 poles. There may be an action such as occurs in the model 

 with steel magnets, by which the W r eber element itself disturbs 

 this balance, with the result that the unlike poles attract more 

 strongly than the like poles repel. In that case the Weber 

 element would be stably held by the fixed magnets of its own 

 atom, when in line with an}', pair. As it turns it transfers 

 to the pair between which it lies a slight inequality of forces 

 on which its stability depends. 



In any event it is probable that the atom is capable of 

 distortion, and is in fact distorted when the piece is strained. 

 It is well known that a homogeneous ferromagnetic metal 

 becomes magnetically seolotropic under strain. The model 

 suggests that such an effect is due to the setting up of 

 differences in the closeness of the controlling magnets along- 

 different axes. Extreme amounts of straining, such as those 

 that are experienced when a non-crystalline state is produced 

 (as in the experiments of Beilby) — with atoms forced out of 

 their places in the lattice — are probably associated with 

 considerable distortion of the atoms, We should expect this 

 • to disturb the balance of opposed magnetic forces, but 



2 K2 



