372 
PROFESSOR J. A. EWING ON THE EFFECTS OF STRESS AND 
negative maximum on the “off” curve occurs later and later during the removal of 
load as the intensity of magnetisation increases. 
§ 30. Figs. 19-22, Plate 23, exhibit the results of another experiment, on the same 
wire, made with the object of comparing the changes of magnetisation and thermoelectric 
quality caused by changes of magnetising force with those caused by changes of load. 
The wire—free from load—having been demagnetised by reversals, was carried 
through a cycle of magnetisation by gradually applying a magnetising force <£> of 
nearly 17 c.g.s. units, gradually withdrawing and reversing, and finally re-reversing 
<§, while the magnetisation and thermoelectric quality were measured in the usual 
way. The consequent changes of 3 in their relation to are shown iu fig. 19. The 
simultaneous changes of the thermoelectric E.M.F. are shown, in their relation to 3 > 
in fig. 20. The precise form of the curves in fig. 20 is somewhat uncertain, in con¬ 
sequence of irregularities which were perhaps due to the difficulty of keeping the 
thermal condition absolutely constant during so long an experiment. The effect of 
the first magnetisation is shown by the full line (in fig. 20), of the first reversal by the 
lower broken line, and of the second reversal by the upper broken line. There is not 
much evidence of hysteresis in the relation of thermoelectric quality to intensity of 
magnetisation. 
Then the wire was demagnetised by reversals, the magnetising field was reduced 
to zero, and the process of loading and unloading (up to 12 kilos.) was performed. 
Fig. 21 shows the resulting thermoelectric changes to the same scale of E.M.F. as 
is used in fig. 20. During the process the magnetometer was observed, and a very 
slight residue of magnetism which had escaped the demagnetising process was seen 
to go through the minute changes shown in fig. 21, where 3 is drawn to the same 
scale as in figs. 19 and 20. It is obvious, by comparison of figs. 20, 21, and 22, that 
these changes of 3 were wholly incompetent to account for the changes of E.M.F. in 
fig. 21. The changes of E.M.F., caused even by the strongest magnetisation reached 
in the experiment, fell short of those wdiich occurred during the application and 
removal of load when the wire was almost wholly free from magnetism. 
§ 31. The effect of magnetisation on the thermoelectric quality of iron is that longi¬ 
tudinal magnetisation, in either direction, makes the magnetised part more negative 
in the conditions examined in fig. 20. In that case there was no load on the wire, 
and the question suggested itself: How is this effect of magnetisation modified by the 
presence of a constant stress ? To answer this question a group of experiments was 
carried out, whose results are exhibited in figs. 23, 24, 25, and 26, Plate 23. In each 
of these the wire, demagnetised by reversals to begin with, was alternately magnetised 
first in one and then in the other direction, as in figs. 19 and 20. The relation of the 
thermoelectric E.M.F. to the intensity of magnetisation 3 is shown by the figures. 
Fig. 23 is for the case of no load, and is virtually a repetition of the previous experi¬ 
ment of fig. 20. Fig. 24 is for the case where a constant load of 3 kilos, was applied 
to the wire. Fig. 25 is for a constant load of 6 kilos., and fig. 26 for 12 kilos. (Figs. 23, 
24, and 25 are drawn about the same axes; fig. 26 is drawn about a separate axis of 3 
