PROPERTIES AND THE ELASTIC EXTENSION OF METALS. 
137 
The inference is that the material is in the overstrained condition. That this inference 
is correct is shown by fig. 26. A test bar of the same material was annealed by heating 
Fig. 26 (normal mild steel). 
to 550° C., and then cooling slowly within the furnace, and then it was found that it 
gave the diagram of fig. 26, which is the normal shape for the class of steel tested. 
I have tested many bars of steel rejected on shock test and giving low impact 
numbers, and I have always found that the shape of the load extension diagram 
discloses the abnormal state of the metal. Much work of a comparative kind must be 
done before the result can be widely generalized. The War Committee of the Royal 
Society did a considerable amount of work in this direction. 
The limiting fatigue stress may probably be found by inspection from a load elastic 
extension diagram. 
It is probable that the limiting range of stress in fatigue has for its positive value 
the stress equal to the limit of proportionality. 
Referring to the diagram for iron, fig. 7, it will be seen that the limit of 
proportionality is at about 3|- tons, corresponding to 11'45 tons per sq. inch. 
I prepared six test pieces of the material, and Dr. Stanton kindly applied his 
fatigue test to them at the National Physical Laboratory. He found, after applying 
alternating loads in the aggregate 24,000,000 times to the eight test pieces, that the 
approximate limiting range of stress in fatigue was between ± 10|- and ±13 tons per 
sq. inch. The average is ±11*75 tons per sq. inch. 
