54 
MR. S. W. J. SMITH ON THE THERMOMAGNETIC ANALYSIS OF 
experimentally the observations were made at about 600° C. and 700° C., i.e., at 
temperatures where the permeability was not subject to rapid variation. The 
experiments 26, 27 and 28 seem to show that there is distinct irreversibility on the 
first cooling from 700° C., but the experiments 31 to 34 seem to show that, after 
the first reheating and cooling, the subsequent variations are practically reversible— 
the permeability being subject at most to a very slight rise. 
An attempt was made to determine how the curve of alternate heating and cooling 
just described departs from and returns to the steeper curve of continuous cooling. 
It was here that the uncertainty arising from the difficulty of maintaining a constant 
temperature became most pronounced, but from the experiments 40 to 42, in 
conjunction with those just quoted, it would seem that at the beginning of the first 
reheating the permeability rises more rapidly than it does subsequently (fig. 11), and 
that the changes in permeability variation are certainly less abrupt than in the 
qualitative diagrams given by Guillaume. (Experimental magnetic data are not 
given in any of the papers by Guillaume to which I have had access. It is worth 
notice that permeability measurements may reveal changes of internal structure too 
small to be detectable by other methods. Cf. Section VII., § 12, p. 101 below). The 
data of fig. 11 are not sufficient to show how the curve of falling permeability, passing- 
through 42 and 43, meets the curve of continuous cooling, but it is noteworthy, as in 
the first winding, that interrupted cooling and subsequent treatment as above is 
followed, in every case, by a lower permeability at ordinary temperatures than if the 
cooling had been continuous from beyond the critical point (see later, Section VI., 
§11, P- 91). 
It was next intended to test for reversibility, on cooling and reheating, from a 
point in the region (between 700° G. and 800° C.) in which the permeability falls 
rapidly with rising temperature. For this purpose the points 45 to 47 (see also fig. 10) 
were taken in order to get the general position of the region in question. The heating 
was not continued beyond about 770° G. (47) owing to want of time. It was noticed 
that, as the result of uninterrupted cooling from 770° C., the permeability at ordinary 
temperature was lower than any value previously obtained, and 20 per cent, lower 
than that obtained after uninterrupted cooling from temperatures above that at which 
ferromagnetism becomes imperceptible. 
The ring was then reheated to a point on the falling branch of the curve (about 
720° C.) and then allowed to cool without intermediate reheating. Measurements 
of permeability were taken from time to time until the air temperature was reached. 
The results are shown in Experiments 50 to 55. The permeability at ordinary 
temperature was again exceptionally low and practically identical with that obtained 
just previously after uninterrupted cooling from 770° G. 
The permeability at ordinary temperature was now raised, by treatment similar to 
that described under the first winding (see 67), and the ring was then heated to about 
755° C. and again cooled uninterruptedly (except for stoppages at three temperatures 
