34 THE QUANTUM [in. 3 
field there is first a small stable deflexion (this is the reversible 
stage), secondly a breaking-away from the position of stability 
through an unstable phase, and finally a settling down into a 
new position of stability. The theory also accounted for the 
main facts of magnetic hysteresis. Ewing * has recently pointed 
out that the old model fails quantitatively. Experiments show 
that the range of stable deflexion is very small. At the same 
time they show clearly that the control must be weak, for a 
small magnetizing force suffices to upset the equilibrium of most 
of the Weber elements. To account for the narrow limits of 
reversible deflexion it is necessary to suppose that the magnets 
are placed very close together; but this makes their stability 
far too great—the field which would be required to change from 
the reversible to the irreversible stage is many thousands of 
times greater than that actually observed. To meet these 
difficulties Ewing has modified the original theory. In the new 
model a narrow range of stable deflexion is secured without 
excessive stability. This is obtained by so devising the model 
that the control of the Weber elements in each atom depends on 
the inequality between strong opposing forces. These forces are 
exerted separately on different parts of it by other constituents 
of the atom. The actual models devised by Ewing consisted 
partly of systems of small magnets, each magnet representing a 
Weber element, and partly of systems in which electric circuits 
* J. A. Ewing, Proc. Roy. Soc., vol. ioo, p. 449, 1922 ; Phil. Mag., 
vol. 43, p. 49 3> 1922.