THE ELECTRIC AND LUMINIFEROUS MEDIUM. 
287 
stances, including gases, the coefficient of magnetization varies inversely as the 
absolute temperature, with a degree of accuracy which tends to j^erfection at high 
temperatures : that in strongly magnetic substances such as iron, nickel, and magne¬ 
tite, the same law is ultimately reached when the temj)erature is sufficiently high ; 
Avhile ill diamagnetic substances the coefficient is usually nearly independent of 
temperature and also of changes in the chemical state of the material. The inference 
is made by Curie that this points to diamagnetism being an affair of the internal 
constitution of the molecule, having only slight relation to the bodily motions of the 
molecules on which temperature depends; which is in accordance with the modified 
EBERIAN view necessitated by the present theory. On the other hand, paramsg- 
netization is an affair of orientation of the molecules in space without change of 
internal conlormation, so that alteration of the mean state of translational motion is 
involved in it, and we should expect a temperature effect. A striking and probably 
just analogy is drawn by Curie between (i) the simple law of expansion of a gaseous 
substance at high temperature, and the sudden chano-e which it undero’oes on lowerins’ 
the temperature beyond a critical point so that the mutual attractions of the mole¬ 
cules come into jilay and produce the liquid state, and (ii) the simple law of 
magnetization of a substance like iron or nickel at high temperatures, and the sudden 
change which it undergoes when the temperature is lowered beyond the point at 
which the material jiasses into its strongly magnetic or ferromagnetic condition. The 
lelation between paramagnetization and temperature in the former state proves to be 
so simple and univ^ersal that it must be the expression of a theoretical principle. The 
following considerations in fact derive it from Carnot’s principle : the argument is 
precise so long as the induced magnetization is so slight that the exciting magnetic 
force on the sejiarate molecules is practically that of the inducing field, but it loses 
exactness as soon as, owing to diminution of energy of agitation with falling tem¬ 
perature, the molecules begin to exercise sensible magnetic control over each other, 
and thus introduce the phenomena ol grouping and consequent hysteresis that are 
associated Avith the ferromagnetic state. 
72. Consider a mass of paramagnetic material, moved up from a place where the 
intensity of the magnetic field vanishes to a place Avhere it is JT. The aggregate per 
unit volume of the total magnetic energies of its molecules is thereby altered from 
null to IH or kH®. The mechanical work done by the mass in virtue of its 
attraction by the field is ^IH, for the magnetization is at each stage of its progress 
proportional to the inducing force. Thus there remains a loss in the total magnetic 
energy of tne molecules, equal to ; this can only have passed into heat in the 
material; for we can Avork on the hypothesis that the field of force H is due to an 
absolutely permanent magnetic system, so that no energy is used up in producing 
magnetic displacements in the inducing magnets. Noav let us apply Carnot’s 
principle to a reversible cycle in Avhich the material is moved up into the field at 
temperature T -f 8T and removed at temperature T, with adiabatic transition 
