88 
MR. S. W. J. SMITH ON THE THERMO MAGNETIC ANALYSIS OF 
’0039. It is curious that for a sample of nearly pure iron, examined in the same way 
as the nickel-iron alloys, the mean temperature coefficient between approximately the 
same temperatures was found to be nearly the same as those of the alloys, viz., ’0037 
(between 630° C. and 480° C., H = ‘43), although the mean permeability of the iron 
was more than four times that of the meteoric iron and more than double that of the 
nickel steel. 
§ 8. The data 63 to 69, &c., nickel steel, show that the reversible fully magnetic 
state of the material extends between the air temperature and 610° C. If the 
heating is continued beyond 610° C.—for example to 640° C. as in 72—irreversible 
effects are produced (shown by the magnetic behaviour on cooling). This is interesting 
as showing that the change of internal structure begins below the temperature of 
maximum permeability of the material as a whole. (Cf. Section IV., § 8, p. 63.) 
The fact that the loss of magnetism is very gradual in nickel-iron alloys instead of 
being very sudden as in nearly pure iron is clearly in accord with the theory of the 
process by which the alloys lose their magnetism. In iron the temperature at which 
the permeability becomes practically indistinguishable from that of the air is only a 
few degrees above that at which cZg/r/d first becomes negative. In each of the alloys 
there were about 130° C. between the corresponding temperatures. (Cf. Section V., 
§ 4, p. 69.) 
§ 9. If the heating is interrupted at some temperature between that at which the 
magnetism begins to disappear and that at which it has become inappreciable, and if 
the temperature is then varied between this temperature and some lower temperature 
above that at which the solution becomes labile (cf. fig. 26, III.), the changes in the 
amount of crystallisation can, as -already explained, be represented graphically by the 
curve r'r's's" (figs. 26, III., and 27, IV.). 
The agreement between this deduction and the permeability data is shown by the 
experiments r" (163), s' (164), s" (165) ; r" (173), s' (174), s" (175), meteoric iron, 
2nd winding, fig. 15. 
§ 10. The cases in which heating, interrupted at some temperature between B and 
C (fig. 27, IV.) is followed by uninterrupted cooling to the air temperature may next 
be considered (cf fig. 26, II.). 
The two cases which will differ most from each other are represented in fig. 27, V. 
In the first (<]f ), the heating is interrupted soon after solid solution has begun to 
form, i.e ., soon after magnetism has begun to decrease. 
In the second (rst'), the heating is not interrupted until the magnetism has nearly 
disappeared. 
In the former case, in accordance with the explanation already given (cf fig. 26, II.), 
the solid solution, rich in nickel, will deposit crystals (at a decreasing rate) as the 
temperature falls. It will become still richer in nickel during crystallisation (cf 
Section V., § 2, p. 67) and will itself not become labile until a low temperature is 
reached. Hence, over a considerable range of falling temperatures, the state of the 
