378 
PROFESSOR J. A. EWING ON THE EFFECTS OF STRESS AND 
in the thermoelectric behaviour of iron, with any assigned junction temperatures, 
which have now been fully exhibited by means of curves. 
§ 43. Experiments were made with other pieces cut from the same bundle of iron 
wire (diameter =0'64 mm.), but not annealed, to test the changes of thermoelectric 
quality which are caused by longitudinal pull in iron in the hard-drawn state. The 
results of two representative experiments of this kind with two distinct pieces of wire 
are shown in figs. 32 and 33, Plate 23. In both cases the portion of the wire between 
the hot and cold “junctions ” was, for some accidental reason, rather strongly negative 
with respect to the adjoining portions when there was no load, but became positive 
while under a strong pull. 
In the experiment of fig. 32, the temperatures were 135° C. and 21° C. The wire 
was loaded with 13 kilos., then unloaded, and during the next loading the observa¬ 
tions shown in fig. 32 were taken. The form of the curve is similar to that given by 
Cohn for hard wire, except that the loop which hysteresis causes at the upper end 
appears to have escaped his notice. The general result is that loading has a positive 
effect; and, further, that a mechanically hard wire behaves very similarly to a strongly 
magnetised but much softer wire. That it is so will be seen by comparing fig. 32, 
where the wire is hard-drawn but not magnetised, with figs. 17 and 18, Plate 22, where 
the wire was only slightly hardened by stretching, but was strongly magnetised. 
Obviously we should expect that iron in a state of hardness intermediate between 
these would give results resembling those of figs. 15 and 16, which refer to moderately 
magnetised wires; and this corollary has been verified by experiment with a piece of 
the same iron which was annealed, but much less perfectly than in the former experi¬ 
ments, and was then loaded and unloaded without magnetisation. In fact the series 
of figures 14 to 18, which exhibit the effects of various degrees of magnetisation, serve 
also to illustrate how under various degrees of mechanical hardness there is continuity 
in form between the familiar type of fig. 4, or 12, or 14, or 21, and the very different 
looking curve of fig. 32. The phrase, “ mechanical hardness,” must be understood here 
to refer only to the kind of hardness which comes of permanent strain, such as stretch¬ 
ing or wire-drawing; for the writer has made no experiments to determine whether 
a like relation holds between the various degrees of hardness which can be given to 
steel by variations of temper apart altogether from strains. Hardness of the latter 
kind may confidently be expected to reduce hysteresis, which has now been shown to 
be reduced by the presence of load, by the existence of a magnetic condition, and by 
the existence of a state of non-elastic strain or “set.” 
§44. In fig. 33 the temperatures were 143^° C. and 19°C. The wire—an un¬ 
annealed specimen—had a load of 14 kilos, applied to it; this was removed and a load 
of 12 kilos, was applied and removed while the observations were taken whose results 
are exhibited in the figure. This figure, along with the last, shows that the curves for 
hard-drawn iron differ from the curves for softer wire chiefly in this, that the negative 
maximum, familiar in the earlier figures, is now shifted so close to the line of zero 
