86 
MR. S. W. J. SMITH ON THE THERMOMAGNETIC ANALYSIS OF 
§ 4. To show how closely the permeability measurements agree with the view that 
the phenomena assumed above really occur, consider first fig. 27, IV. 
Assume that the nickel-iron alloy contains about 6'5 per cent, of nickel. Crystals 
will begin to disappear at a temperature which has not yet been accurately fixed. 
Their solution will be complete at a temperature not far from 750° C. 
The change may be represented qualitatively by the curve «BC (the ordinate giving 
the percentage of the material that is still crystalline at the temperature corresponding 
to the abscissa), although the curve for extremely gradual heating is more probably 
of the form represented diagrammatically by abc. 
The curve of crystallisation during cooling can be calculated roughly from the data 
at disposal. Thus, at any temperature d, the percentage of nickel in the labile solid 
solution is known approximately, since it is, by hypothesis, the same as the total 
percentage of nickel contained by that alloy in which magnetism reappears at 6. 
It may be assumed as a first approximation from the experimental data that 6 
decreases linearly from 0 — 780° C. when the percentage of nickel is negligible to 
0 = 0° C. when the percentage of nickel is 27, i.e.. that the curve AB of fig. 23 on 
p. 67 is approximately a straight line. 
The percentage of nickel in the crystals in equilibrium with the labile solid solution 
in question is dp (fig. 23), 6q being the percentage of nickel in the solution. Again 
assume, as a first approximation, that the curve AE showing the variation in the 
percentage composition of the crystals is a straight line. For reasons already given, 
suppose that E corresponds roughly to 6 A per cent, of nickel. 
If a is the total percentage amount of Ni in the alloy, and if x is the percentage 
amount of the alloy which is crystalline at the temperature 9, we have 
6p 
Too 
x 
too 
(100 — x) = a, 
from which, assuming that Op and Oq can he calculated roughly as above, it is possible 
to find the value of x for any value of 0. 
The curve CfB'A of fig. 27, IV., is plotted from numbers calculated in this way for 
a = 6 A. It is assumed that the rate of cooling is sufficiently slow to ensure that all 
the crystals at any temperature contain the same percentage of nickel. 
§ 5. The first conclusion from the curve CfB'A is that, in an alloy containing 
6 A per cent, of nickel, the amount of magnetic iron will increase rapidly at first as 
the temperature falls below 600° C., and then more gradually. The final portions of 
the alloy will become magnetic very slowly, and the return to the magnetic state of 
all the iron in the alloy will not be complete until the temperature is below the 
ordinary temperature of the air. This accounts for the fact, carefully examined in the 
first experiments with the meteoric iron (1st winding) and corroborated subsequently, 
that the permeability during cooling, e.g., at 550° C., is considerably below that shown 
at the same temperature during reheating from 15° C. The fact is less obvious in 
