94 
MR. S. W. J. SMITH ON THE THEKMOMAGNETIC ANALYSIS OF 
If x — 1 and jj-± = 100, then, when yu 2 = 1, the permeability will be 
g = 50-25, 
i.e., the presence of 1 per cent, of a non-magnetic alloy would reduce the permeability 
to half the value it would have if the non-magnetic alloy were replaced by material 
of the same permeability as the rest of the ring. 
If g 2 = 5, then, when x = 1, 
g = 84 approx., 
i.e., the permeability in the case just considered would be increased nearly 70 per cent, 
if the permeability of the material in the gap rose from g 2 = 1 to g 2 = 5. 
The permeability of the kamacite in the field used in the experiments is more than 
100 (Section III, § 7, p. 40), and that of a 27 per cent, alloy when magnetic is at 
least 5 in the same field ( cf. Hopkinson’s curves, ‘Roy. Soc. Proc.,’ 1890). Hence 
the loss or gain of magnetism by the small quantity of the richer alloy known to be 
present, occurring transversely in the ring, would produce an effect of the kind and 
relative magnitude shown in fig. 14. 
§ 3. The curves of fig. 14, in conjunction with the other data, prove that, at the end 
of the experiments, the meteoric iron consisted almost entirely of two irreversible 
alloys, viz., a small quantity of an alloy containing between 25 and 30 per cent, of 
nickel and a relatively large quantity of an alloy containing between 6 and 7 per cent, 
of nickel. After the experiments the ring was polished and etched and showed a 
distribution of the kamacite and taenite which, as far as careful examination through 
a lens could decide, was identical with that existing in the core in its original state. 
Hence, although the internal structure of the taenite may have changed during the 
course of the experiments, its distribution with respect to the rest of the meteorite 
remained practically unaltered. At the end of the experiments it was an irreversible 
alloy containing about 27 per cent, of nickel. 
But a careful examination of the data obtained at different times after the beginning 
of the thermal treatment seems to show conclusively that the taenite was subject to 
alteration of internal structure during the course of the experiments and did not, at 
the outset, possess the same properties as the artificial alloy containing 27 per cent, 
of nickel. 
The evidence, as will be seen, supports the hypothesis that the taenite was 
originally a mixture, thermomagnetically discernible, of a nickel-poor alloy such as 
kamacite and of a much richer alloy containing not less than 37 per cent, of nickel. 
if; as before, we assume that the taenite contains t per cent, of Ni and that it 
is a mixture of kamacite containing k per cent, of Ni and of an alloy containing 
t' per cent, (where t! is greater than t), the number of grammes of kamacite in 
100 grammes of the eutectic will be 
x = 100 (t' — t)f(t'—Jc), 
