﻿536 Mr. W. A. Scoble on the Strength and Behaviour 



material is satisfactory, the stresses at the elastic limit and 

 yield-point are nearly proportional and it makes little differ- 

 ence which is taken. Faulty specimens will usually have a 

 low elastic limit whereas the yield-point is little affected, and 

 the same applies to changes in the metal due to any special 

 treatment to which it may have been subjected. Taking these 

 facts together, it is evident that the yield-point is much more 

 nearly constant than the elastic limit, and in making a simple 

 test it is correct to consider both points in relation to each 

 other. The yield-point, unfortunately, is not quite definite, 

 but that taken for the purpose of these tests will be precisely 

 stated. Having decided exactly how the yield-point is to be 

 found, it is more easily obtained accurately than the elastic 

 limit. The error in proportionality between the stress and 

 strain is small to the yield-point. The view is now generally 

 held that the lack of proportionality immediately after the 

 elastic limit is due to local yielding, parts of the material 

 being either weak or under internal stresses. The con- 

 siderable lowering of the elastic limit with poor specimens 

 supports this view very strongly; and if it is accepted, the 

 yield-point is more important than the elastic limit, being less 

 dependent on unique conditions. It will be seen later that 

 when a specimen is tested to yield under combined stresses, 

 there is yield in both ways. We therefore have definite 

 loads of both kinds. Mr. Hancock has shown that when the 

 elastic limit is reached under one loading, a large stress is 

 needed to reach it with the other. A bar tested to the elastic 

 limit in tension needs a considerable torque before this point 

 is reached in torsion. For comparative work a difficulty 

 arises. The first load reaches the elastic limit and the second 

 can do no more except reach it with the other yield, therefore 

 the first load can be taken with any fraction of the second 

 to give the combined loads. No doubt the first kind of stress 

 affects the elastic limit under the other kind, but the con- 

 ditions are not so definite as when the yield-point is taken. 

 Unfortunately more than one position can be ascribed to the 

 yield-point, the point at which the curvature of the stress- 

 strain curve becomes large, or the stress causing considerable 

 yield. As the distribution of stress due to bending varies 

 from a maximum tension to a maximum compression, and 

 due to the torque from zero to a maximum shear, elements 

 will yield in succession, this being especially the case under 

 combined stresses with irregular distribution, the maximum 

 shear stress only occurring at certain places. On this account 

 Mr. Guest used thin tubes to obtain nearly uniform stresses. 

 Since the elastic strain is so small a little yield at certain 



