368 
PROFESSOR J. A. EWING ON THE EFFECTS OF STRESS AND 
to change of thermoelectric quality, and it is well known from the work of Thomson 
and others that stress affects the magnetism of iron, whether that be residual or 
induced. Herr Cohn has not overlooked the possibility that the thermoelectric 
effects of loading may be explained as secondary results of the changes of magnetisa¬ 
tion which it causes, but says that, while he has no conclusive evidence either way, he 
inclines to the view that the observed phenomena are not to be explained as results 
of changes in magnetism. 
The matter is one of much interest, for if the results are due to stress simply, and 
not to the magnetic changes caused by stress, we have here a novel instance of 
hysteresis, entirely distinct from the instances which are common in magnetic 
phenomena. 
§ 18. To put the question to a conclusive test the writer, in his second series of 
experiments, employed independent means to produce and to measure magnetisation 
in the loaded wire. In some experiments the wire was magnetised to begin with, or 
was exposed throughout to a constant magnetising force, while the effects of loading 
and unloading on the thermoelectric quality were examined. In others the wire was 
completely demagnetised before loading began, and was kept free from magnetising 
force during the process. In others still the effects of magnetisation alone on thermo¬ 
electric quality were measured, by varying the magnetic field while the wire was 
either without load, or was kept loaded with a constant weight. 
The results proved beyond question that the effects of stress on thermoelectric 
quality (though modified by the presence of magnetisation) are not secondary effects 
of the changes of magnetisation ; and that the hysteresis of thermoelectric quality 
with regard to stress is a thing distinct from the hysteresis of magnetism, though 
probably proceeding from the same peculiarity in the molecular mechanism. 
§ 1.9. In these experiments the loading was done by lead weights, instead of by a 
water-tank. The wire was held in much the same way as before, with the hot 
junction in a vessel of heated oil, and the cold junction simply exposed to the air. 
The wire was set in a sloping position at right angles to the lines of terrestrial 
magnetic force; and over it, covering the whole of the portion under stress, was 
slipped a tube, on which a magnetising solenoid was wound. Over against the upper 
end of the stressed portion was a mirror magnetometer, to measure the magnetisation 
of the part under stress. The continuations of the iron wire on either side of the 
part under stress were led away so that they scarcely affected the magnetometer, and 
were as much as possible prevented from becoming magnetised. 
§ 20. An arrangement was provided by which the wire would be demagnetised by 
reversals, in the manner described in the writer’s paper of Researches in Mag¬ 
netism,” by passing a continuously and gradually diminishing current through the 
solenoid, while the direction of the current was rapidly reversed many times. 
The magnetisation of the wire, and the magnetising force applied to it, were 
* Phil. Trans., vol. 176 (1885), p. 539, § 19. 
