3G4 
PROFESSOR J. A. EWING ON THE EFFECTS OF STRESS AND 
draw, they became steady, and the curves of E.M.F. and load were well defined 
throughout the remainder of the experiment, and have the same character as those 
already described. In this instance the whole diagram lies, to an unusual degree, on the 
positive side, probably because of some local peculiarity in the metal at the hot junction. 
Here, again, the effect of each step in the process of drawing out is well marked 
(first at 30 kilos., then at 37 kilos., and last at 42 kilos.). At each extension the 
stretched portion becomes more and more positive. The want of coincidence between 
each “off” curve (or curve of E.M.F. and load during removal of load), and the 
succeeding “on” curve (or curve of E.M.F. and load during application of load), is as 
conspicuous as in fig. 2. 
§ 9. This last is the most noteworthy of the phenomena now under review. In 
drawing attention to it in the paper submitted to the Royal Society in 1881, the 
writer was not aware that it had been previously noticed by any observer. He is 
now glad to be able to refer to an important paper by Herr Emil Cohn," who has 
anticipated him in the discovery of this very interesting characteristic of the curves 
of stress and thermoelectric quality. In the figure which Herr Cohn has given to 
illustrate his experiments with annealed iron wires, the process of loading has not 
been carried far enough to pass the negative maximum of E.M.F. shown in figs. 1, 2, 
and 3, and for this reason the curves have a comparatively simple form. But the 
“on'’ and “oft”’ curves differ from one another in just the way which is shown by 
the independent evidence of the present writer’s experiments. 
§ 10. For the next group of experiments another piece of the same wire was taken, 
and, after annealing, was loaded with 21 kilos., which caused some permanent 
extension. This load was left on for a time, and was then removed. In all the 
subsequent loadings of this wire the load never exceeded 21 kilos., so that subsequent 
loadings caused no other than elastic strain. 
The load of 21 kilos, was then repeatedly applied and removed. This brought the 
changes of E.M.F. to a cyclic state. They are shown in fig. 4, Plate 21, for the cycle 
of loads 0 — 21 — 0. Numerous experiments with other specimens have shown that 
the curves of fig. 4 are thoroughly characteristic of the behaviour of a piece of iron 
which, after annealing, has been stretched beyond its limit of elasticity, and is then 
alternately loaded and unloaded. 
The great difference between the “ on ” and “ off” curves will be realised when it is 
noticed that the negative maximum occurs at about 18 kilos, on the “ on ” curve, but 
at near 8 kilos, on the “off” curve, and that a load such as 4 kilos., when reached 
from a lower value, has associated with it a strongly positive quality on the part of the 
stressed piece, while if approached from a considerably higher value it is associated 
with an even more strongly negative quality. 
* “ Ucber das thermo-electrische Verhalten gedehnter Drahte,” Wied. Anna!., vol. 6 (1879), p. 385. 
A few similar observations bare more recently been published (three years after the present writer’s first 
paper on the subject was read before the Royal Society) by R. Overbeck, Wied. Annal., vol. 22 (1884), 
p. 344. 
