84 
MR. S. W. J. SMITH ON THE THERMOMAGNETIC ANALYSIS OF 
solution becomes labile is a single-valued function of its concentration. It remains to 
show how many of the results described in Section IY. can be explained on the 
assumption that crystallisation occurs in nickel steel from a solid solution, in accordance 
with a similar law. 
§ 2. The irreversible phenomena discovered by Hopkinson can be interpreted at 
once in terms of the theory outlined above. Thus, whilst magnetism is disappearing 
during heating, magnetic crystals are, by hypothesis, dissolving in solid solution, and, 
finally, when all the crystals have disappeared, the material consists of a solid solution 
only. It is a homogeneous mixture, of which the constituents may be crystalline, but 
are not magnetic. Further, just as in a mixture of crystals and water, the 
temperature at which the last of the crystals dissolve rises when the relative amount 
of crystals originally present is increased, so the temperature of the disappearance of 
magnetism, in a mixture of iron crystals and nickel, rises when the percentage amount 
of iron present increases. 
When the magnetism has just disappeared, a saturated solid solution remains. If 
this is allowed to cool it becomes supersaturated and metastable, but eventually it 
becomes labile. The temperature of lability, at which spontaneous recrystallisation 
begins, is definite and depends upon the relative amounts of nickel and iron present in 
the material. It decreases as the percentage of nickel in the alloy increases, just as 
the temperature at which lability is reached in a supersaturated salt solution decreases 
as the percentage of water which it contains increases. 
Thus, qualitatively, the manner of disappearance and reappearance of magnetism in 
nickel-iron alloys corresponds with that of disappearance and reappearance of crystals 
in fluid solution. 
§ 3. The applicability to the case of solid solutions of the fundamental hypothesis 
discussed in § I with respect to fluid solutions seems to receive remarkable confir¬ 
mation from the experiments already described. The general characteristics of the 
thermomagnetic curves and the quantitative relations between them, when a numerical 
comparison can be made, are in accord with the theory. 
The changes in the alloy when heated will differ slightly from those in the 
hypothetical case of an aqueous salt solution considered above. They will correspond 
more nearly with those which would occur in a crystallised salt which can go into 
solution in its own water of crystallisation. Referring to fig. 27, IV., the amount of 
crystalline material will remain constant (following the horizontal line) until a 
temperature is reached represented qualitatively by B. Gradual solution of the 
crystals will then begin and will continue until, at the point C, they have dis¬ 
appeared. 
If the percentage of nickel in the alloy is less than that contained by the eutectic, 
the changes that take place in the material in the region BC will be (according to the 
theory of mixed crystals) as follows :— 
Considering a small element of the material at first magnetically crystalline, when 
