1921-22.] On Models of Ferromagnetic Induction. 
97 
IX.— On Models of Ferromagnetic Induction. By Sir J. Alfred 
Ewing, K.C.B , F.R.S., Principal of the University of Edinburgh. 
(With Two Plates and Fifteen Figs, in Text.) 
(MS. received February 20, 1922. Read February 20, 1922.) 
1. In 1890 I published a theory of ferromagnetic induction in which it 
was suggested that the equilibrium of Weber’s elementary magnetic 
particles was due only to magnetic forces.^ It was shown that when the 
elementary magnets are made to turn by applying a magnetising force 
which is progressively increased, the conditions of equilibrium must be 
such that there is first a small amount of stable (reversible) deflection, 
then a break away with irreversible deflection into a new position of 
stability, and finally a reversible approach to the position of complete 
parallelism which corresponds to saturation. A model was constructed 
showing that these conditions could be satisfied by a purely magnetic 
control. It was made up of little magnets, pivoted on fixed centres which 
were uniformly spaced. The magnets were all free to turn, but controlled 
one another by their mutual magnetic forces. They tended to form rows, 
and when an external field was applied the rows broke up and fresh rows 
were formed more nearly in the direction of the field. With this simple 
model the known characteristics of the magnetising process were, qualita- 
tively, well reproduced. A recent study of the stability of such rows of 
magnets has, however, led me to abandon this model, and to design instead 
a model in which each atom forms a magnetic system comprising a Weber 
element capable of turning, but controlled by the magnetic forces exerted 
on it by other parts of the atom which are taken as fixed.f As in the old 
model, the control is wholly magnetic. Various forms of the new model 
will be described in a later part of this paper, but in the first place it may 
be useful to give some account of the investigations of stability which 
have convinced me that the old model fails, quantitatively, to represent 
the process. 
2. When a magnetising field is applied to a substance such as soft iron, 
the amount of magnetism that is taken up during the initial or quasi- 
elastic stage, before hysteresis begins to show itself, is a very small part, 
* “Contributions to the Molecular Theory of Induced Magnetism,” Proc. Roy. Soc.^ 
xlviii, p. 342 ; Phil. Mag.., Sept, 1890. 
f Proc. Roy. Soc., Feb. 1922 ; Phil. Mag., March 1922. 
VOL. XLII. 7 
