4 
MR. J. MUIR ON THE TEMPERING OF IRON 
sq. inch of load until higher stresses were obtained, and then readings were taken 
after every half ton, and while each half ton was being slowly added the eye was 
kept at the microscope of the extensometer in oi'der to detect as accurately as possible 
the load at which large plastic yielding commenced. Curve No. 1, Diagram 1 (p. 6), 
shows that such yielding began at tlie high load of 38 tons per sq. inch, a well- 
defined yield-point being obtained at that stress. 
In plotting the curves of the j^resent exanqjle and of all other diagrams in this paper, 
tlie method of “ shearing back ” the curves, which was adopted on Professor Ewixg’s 
suggestion in the author's previous paper on Recovery from Overstrain, has again been 
employed."^ This method consists in diminishing all extensions before jdotting by an 
amount proportional to the loads producing them. By this means curves which 
would otherwise stretch far across the paper in the direction of extensions are brought 
more nearly into an upright position, so that a large scale for the measurement of 
extensions may he retained without an inconvenient amount of sj^ace being occupied. 
All the curves in this paper have been “sheared hack” by the same amount; 
of an inch have been deducted from the extension of a 4-inch length 
for every 4 tons of stress. For example, the extensometer readings for Curve 1, 
Diagram 1, corresponding to the stresses of 4, 8, and 12 tons per sq. inch, were 120, 
240, and 362 respectively. The numbers actually plotted were 10, 20, and 32. 
It should be remarked that in all the diagrams of this paper, the origin for the 
measurement of extensions has been displaced by an arbitrary amount between each 
curve and the next in the series. This was merely to keep the various curves distinct, 
and to facilitate comparison. 
The yielding which is shown by Curve No. 1, Diagram 1, to have begun at the 
stress of 38 tons per sq. inch soon became very rajDid under this stress. The load was 
therefore slightly reduced! until a rate of extension convenient for observation was 
obtained, and it was then found that the stretching continued at a more or less 
constant speed (the lever of the testing machine being kept floating) for a consider¬ 
able time, and then abruptly stopped ; or, to he more accurate, rapid extension then 
abruptly changed into very slow creeping. The replacement of the full load of 
38 tons was found to produce comparatively little further extension. 
Th is yielding which takes jdace at the yield-point does not occur simultaneously 
throughout the length of the specimen. The material at some point in the bar 
yields and the yielding is ol)served to spread. The yielding at any point increases 
the intensity of tlie stress at nelghhourlng points of the bar, so that adjacent 
portions of material 3 deld, and the action is transmitted piecemeal throughout the 
whole length of the specimen. With unturned material the progress of this 
yielding can he observed in the skin of oxide cracking and springing off as the 
* In ‘Phil. Trans.,’ A, vol. 193, p. 12, 1899, a full account of “shearing hack” is to be found. 
t With Lowmoor iron it was found that the load at the yield-point eould sometimes be i educed by 
almost two tons/in- without causing the yielding at the yield-point to eease. 
