RESEARCHES IN MAGNETISM. 
525 
The fact that iron possesses retentiveness, or, in other words, that it remains 
magnetic after a magnetising force has been applied and removed, is of itself sufficient 
proof that there must be hysteresis in the relation of magnetism to magnetising force, 
when the changes of the latter are such that forces of opposite signs are alternately 
applied. But it does not follow that hysteresis is necessarily present when the changes 
of magnetising force are restricted to one direction and to one sign. The case may be 
supposed similar to that of a strained solid. When a stress along any axis is alter¬ 
nately applied and reversed, we know that (provided the resulting strain exceeds the 
body’s limit of elasticity) there is hysteresis, of a simple and obvious character, in the 
relation of strain to stress. But if, instead of being reversed, the stress is merely 
applied and removed periodically, the greater part, if not the whole of the resulting 
change of strain (after the first application) is purely elastic, and we have no ground 
for asserting that there will then be hysteresis in the relation of strain to stress. In 
fact, if the process of magnetisation is strictly analogous to the straining of a solid 
whose limits of elasticity are exceeded by the strain, we should expect to find little 
or no hysteresis in the changes of magnetism which occur when after applying a 
magnetising force we (wholly or partially) remove and reapply it. 
§ 3. The point under consideration has an important bearing on the theory of 
Weber, which endeavours to explain the process of magnetic induction by supposing 
that the molecules of iron and other paramagnetic substances are always magnets, 
whose axes point indifferently in all directions, until, in consequence of applied mag¬ 
netising force, they are turned more or less towards the direction in which the force 
acts. Weber supposes that each magnetic molecule, when deflected from its initial 
position, tends to return to that position with a force which is the same as that which 
a magnetising force D, acting in the initial direction of its axis, would produce. If 
then deflection of the molecule be produced by the action of an applied magnetising 
force X, the direction which the molecule takes up while the force is at work is such 
that its magnetic axis points along the resultant of X and D. The force X exerts on 
it (per unit of its magnetic moment) a couple X sin 6, where 6 is the angle made by 
its axis, in its deflected position, with the direction along which X acts. And the 
assumed restoring force D exerts a couple Dsin/3, where /3 is the angle through 
which the molecule has been revolved. The molecule remains in equilibrium under 
these, and only these, forces, and when the magnetising force X ceases to act, it 
returns to its initial position through the action of D. Thus, the theory, in this 
form, takes no account of residual magnetism, and fails to explain retentiveness. 
To remedy this defect, Clerk Maxwell (‘ Electricity,’ ii., Part 3, chap, vi.) has 
suggested a further assumption, based on the analogy of magnetism to mechanical 
strain. He supposes that when a molecule is deflected by a magnetising force X, it 
returns to its primitive position on the removal of X, provided the angle of deflection /3 
has been less than a certain limit /3 0 ; but if the deflection has exceeded /3 (J , then when 
X is removed the molecule does not completely return, but remains deflected through 
MDCCCLXXXV. 3 Y 
