106 
MR. S. W. J. SMITH ON THE THERMOMAGNETIC ANALYSIS OF 
Further, the effects of temperature change upon the BH curves will be very 
different in the two cases. 
In the nickel-rich crystals, on account of their magnetic softness, increase of 
temperature (below the critical temperature) will not he accompanied by increase of 
induction (for a given held) unless the field intensity is small. In the nickel-poor 
crystals, however, 
(clB/cW), H constant, 
will not become negative until H is relatively large. 
Thus, while in each case the fx6 curve will be convex to the axis of temperature 
when the field is very small, it will become concave to the axis of temperature in the 
nickel-rich alloy under a field for which the /±6 curve of the nickel-poor alloy is still 
convex. 
These relations are shown in Hopkinson’s curves (‘Roy. Soc. Proc.,’ 1890; cf also 
Wills, ‘Phil. Mag.,’ 1900). 
§ 5. The significance of the results obtained by Honda and Shimizu may now be 
briefly considered. 
In the 36 per cent, alloy, for a field of 50 G.G.S., the permeability decreases 
continuously from the temperature of liquid air to that at which magnetism 
practically disappears (about 250° C.). The decrease is most rapid at the end, nearly 
half of the original magnetisability being lost between 150° C. and 200° C. 
Further, there is very little thermal hysteresis during cooling and, in sharp contrast 
with the behaviour of the alloys containing less than 30 per cent, of Ni, the magnetic 
properties at 15° G. are practically unchanged by cooling to — 186° C. This is clearly 
in agreement with the theory already given when it is remembered that 36 per cent, 
is approximately the nickel content of the least concentrated of the nickel-rich mixed 
crystals. 
§ 6. In the alloys containing between 30 and 28 per cent. Ni the permeability 
decreases as the temperature rises from —186° C. until a temperature of about 150° C. 
is reached. Beyond this the permeability rises for,H = 50, and diminishes slowly for 
H = 400, to a temperature of about 300° C. After this the permeability falls at first 
gradually and then very rapidly in both fields as the temperature approaches 500° C. 
These data are clearly explicable if we assume that, at —186° C., each alloy is a 
mixture of the magnetically soft 36 per cent, nickel-iron and of a magnetically harder 
alloy for which the permeability increases with rise of temperature, even when 
H = 50, and disappears at a higher temperature than that of the soft alloy. 
The character of the magnetism which first appears during cooling is significant. 
If the IH curves be plotted from the data given, it will be seen that the initial 
small amount of magnetism at the temperature of the air is that of a magnetically 
soft material, and that each of the curves closely resembles that, at the same 
temperature, of the 36 per cent, alloy already referred to. 
Further, as the temperature is reduced below 0° C-, the magnetisability gradually 
