HAKDENED BY OVERSTRAIX. 
27 
at the stress of 42 tons per sq. inch. A stress of 40^ tons was, however, found 
sufficient to cause the yielding, once it had started, to spread throughout the length 
of the specimen. This result obtained with a specimen winch iiad been largely over¬ 
strained by a single loading is, like the result of the experiment last described, in 
practical agreement with Diagram 6. Curve No. 7 of that diagram shows that, 
with a specimen which had been largely overstrained by the passage of several yield- 
points, a temperature of 500° C. brought the yield-point to a stress of 40 or 41 tons 
per sq. inch. 
The Tempering of Loivmoor Iron Hardened hy Stretching. 
The tempering of Lowmoor iron which has been hardened by stretching now 
remains to be considered. A specimen of the soft iron whose properties were 
described in a preceding section of this jmper, was annealed by heating to 750° G. 
with slow cooling, and was then subjected to the treatment illustrated by Diagram 6. 
The procedure adopted will he readily understood hy reference to the diagram : it 
was practically identical with that adopted to obtain Diagram 6. 
Tlie specimen was first hardened Ijy ovei-sti'ain in the manner illustrated hy 
Curves 1, 2, 3, and 4, Diagram 7. A temperature of 320° C. was then shown 
(Curve No. 5) to have had no tempering action on the hardened material, while the 
amount hy which tlie yield-point was lowered in consequence of heating the 
specimen to 350 to 400 and to 500° C. is shown by Curves 6, 7 and 8 respectively. 
Perhaps reference should again he made here to the method adopted to detect 
yield-points. The load was very slowly applied as its critical value was approached ; 
the extensometer reading was continually observed in order to detect “creeping” within 
the 4-inch length under observation, and the lever of the testing machine was 
watched for any indication of yielding starting outside tlie measured length of the 
specimen. When “ creeping ” had been detected, the load was immediately reduced 
so as to keep the material yielding slowly. A considerable lowering of tins load at 
the yield-point could often be effected, as is shown hy the dip in various curves of 
the present Diagram No. 7, without causing the extension of the yield-point to 
stop ; but since the rate of extension fluctuated somewhat as the stretching action 
travelled along, the load had sometimes to he slightly increased again in order to get 
the yielding to spread at a reasonable rate throughout the whole length of the 
specimen. The exact position of a yield-point is thus a little indefinite. By slightly 
reducing the load as the yielding at a yield-point was occurring, the danger of fracture 
supervening at the higher yield-points was somewhat reduced. 
After curve No. 8 of Diagram 7 had been obtained, the specimen was again heated 
to 500° C., and Curve No. 9 shows that the yield-point had been brought to very 
much the same position as before. The loading in tliis test was carried far beyond 
the yield-point, and then the specimen was once more raised to 500° C. Curve No. 10 
E 2 
