1911-12.] The Molecular- Theory of Magnetism in Solids. 245 
magnetisation would have a very sharp apex and would correspond fairly 
well to Weiss’s lower curve. 
20. In the preceding investigation it has been postulated that the 
molecular magnets are all co-directed. This implies that the external field 
has been strong enough to turn them all so as to form a homogeneous 
crystalline assemblage, and a single infinite homogeneous assemblage has 
alone been considered. In the paper Proc. R.S.E., 1905, a brief considera- 
tion of the effect of plane boundaries has been given. But the process 
can readily be applied to cases in which sub-groups of the magnets are 
differently oriented. Thus magnets in alternate plane layers might be 
oppositely directed and be in stable equilibrium. If such a compound 
arrangement corresponded to a condition of greater exhaustion of internal 
potential energy than that above considered, we could have the system of 
magnets in a condition practically “ non-magnetic,” and, at a definite 
temperature, the two conditions might conceivably be of equal probability ; 
so that, above that temperature, the one might prevail, while below it the 
other might prevail. In this way the demagnetisation of iron, regarded as 
a congeries of crystals, may be explainable. Again, below that temperature, 
either condition might be possible, as, e.g ., is the case with some iron-nickel 
alloys. This is indeed probably more to be expected in the case of two 
interpenetrating and distinct, independently homogeneous, crystalline 
arrangements, than it is in the case of two interpenetrating and identical 
arrangements, together forming one single homogeneous arrangement, even 
if the magnetic moments of the molecules in each individual set were 
different. 
The factor M, the magnetic moment, which appears in the formulae, will 
effectively depend upon temperature ; and its form can readily be found 
from statistical considerations. It will also depend upon the inherent 
structure of the magnetic molecule. 
21. The persistence of the earth’s magnetism, in spite of the high tem- 
perature supposed to exist in its interior, may possibly be attributable to 
the maintenance, under extreme pressure, of a magnetic condition which, 
under ordinary pressures, could not persist at very high temperatures. This 
might very conceivably occur if the pressure stress were not isotropic. For 
the preceding discussion shows that the powerful terms L' 0 and T' 0 , which 
do not exist in cubic crystals, would then become effective. A numerical 
estimate shows that these terms would become as efficient as the terms 
L' 2 and T' 2 if the cubic crystals were made tetragonal under a stress 
difference amounting to some tens of tons’ weight per square cm. To 
account for axial polarity in this way, we see from § 8 that there must be 
