RESEARCHES IN MAGNETISM. 
577 
have scarcely even a speculative interest if the basis of the formula is not an 
intelligible physical theory and its terms are not capable of physical interpretation. 
§ 65. Weber’s theory of the magnetisation of iron is incomplete, because it omits 
to take account of retentiveness. Maxwell’s extension of it, which ascribes 
retentiveness to something resembling the permanent set of an overstrained solid 
fails, it appears to me, to explain the hysteresis which is found in all changes of 
magnetism brought about by changes of the magnetising force. The idea of a 
resistance to rotation of the molecules resembling the friction of solids, which has 
been suggested by several writers, suffices to explain hysteresis, and is supported by 
other phenomena which I have already noticed in passing. It accords remarkably well 
with the low values of k which we find in the early part of the magnetisation curve ; 
also with the effects which vibration has in increasing the susceptibility of iron to 
small forces and in destroying its retentiveness. In fact, for soft iron it requires little 
modification to make it cover the facts. In hard iron and steel, on the other hand, 
where the residual magnetism is of a far stabler character, retentiveness seems partly 
due to some such action as Maxwell has suggested, though effects resembling those 
of friction are also present. Both notions—that of permanent set and that of 
frictional resistance—seem needful for anything like a full account of the phenomena, 
and if we are to attempt to form a mechanical conception of the process on such lines 
as these, we must assume : (1) An elastic tendency on the part of the molecules to 
recover their primitive position when displaced ; (2) a static frictional resistance to 
their displacement and to their return, removable by vibration ; (3) a limit for each, 
such that if the displacement of the molecule exceeds it, a permanent displacement, 
not removable by vibration, results ; (4) probably a viscous resistance to the displace¬ 
ment and return of the molecules ; (5) an unequal distribution amongst the molecules 
of the frictional resistance, such that in some of the molecules it is very (perhaps 
indefinitely) small. An alternative to this last supposition has already been men¬ 
tioned, but it is in any case necessary to assume that the friction varies somewhat 
widely amongst the molecules, in order to prevent the susceptibility from changing 
discontinuously. 
§ 66. A theory involving so many arbitrarily adjustable constants evidently admits 
of being brought without difficulty into general harmony with what we are able to 
observe of the process of magnetisation. Any examination of it by comparison with 
observed magnetic values could scarcely be conclusive. We might, however, by apply¬ 
ing these ideas, construct, or at least imagine, a mechanical model in which the pieces 
would move in such a manner as to give results closely imitative of those observed 
when we magnetise and demagnetise iron. It must be admitted that any mechanical 
analogue of this kind is probably a very crude representation of molecular movements. 
One cannot suppose that if we could obtain an insight into the real nature of magnetic 
induction we should find anything at all resembling miniature drums turning under 
brake-straps; nevertheless, it is certain that the internal movements which are 
