370 
PROFESSOR J. A. EWING ON THE EFFECTS OF STRESS AND 
Fig. 12, Plate 21, shows the changes of thermoelectric current in relation to load 
in this experiment, and fig. 12a the (simultaneous) changes of magnetism, also in 
relation to load. The latter are not so perfectly cyclic as are the former, and in 
other respects, which will be evident on a close inspection, the two curves present 
noticeable features of difference, as well as a very striking general similarity. 
§ 23. The effect of magnetisation, separately, was then examined. While the wire 
was kept free from load a current was established in the magnetising solenoid, and 
strengthened step by step until the magnetisation was as great as the maximum value 
reached by loading in the previous experiment. This was found to produce very little 
change of thermoelectric quality—only about one-twentieth as much as occurred 
during loading. It was clear from this that the thermoelectric effect of stress was 
not a secondary result of the change of magnetisation. 
The same result was confirmed by other experiments, in which the wire was 
demagnetised as completely as possible by the method of reversals before the 
process of loading began. In such a case it was found that the thermoelectric 
effects of load were as marked as ever, although the wire was, and remained, sensibly 
free from magnetisation during loading and unloading. 
§ 24. To examine more particularly the influence of magnetism on the thermo¬ 
electric effects of stress, the same wire was used in a group of experiments, whose 
results are exhibited in figs. 13 to 18, Plate 22. In the experiments in this group 
the wire received more or less magnetisation to begin with, by the application of a 
magnetic field which was maintained constant during the process of loading and 
unloading. The changes of magnetism and the changes of thermoelectric quality, pro¬ 
duced by loading and unloading, were then observed, and both are shown graphically 
in their relation to the loads, and in absolute measure. The temperatures of the 
junctions were 160° C. and 20° C. 
In each case, after the magnetising field had been brought to a constant value, 
the load was raised from 0 to 12 kilos, (which was the stress originally employed in 
stretching the specimen, § 21). It was then reduced to 0, then reapplied up to 12 
kilos., and finally reduced to 0. The effects of the first application and removal are 
shown in each figure by full lines, and the effects of each second application and 
removal by dotted lines. 
First, the wire after being set in position at right angles to the earth’s lines of 
magnetic force was demagnetised by the method of reversals, and then loaded, while 
readings of the thermoelectric effect were taken. These are given, reduced to micro¬ 
volts, in Curve la, fig. 14. The magnetometer was watched during this process, hut 
it remained undeflected, there being neither residual magnetism nor inducing field. 
The second “ off ” curve in la, was so nearly coincident with the first “ off ” curve 
that it has not been drawn. 
§ 25. Next (the wire being free from load) a current was established in the magne¬ 
tising solenoid, of such strength as to give an intensity of magnetisation 3=160 
