; 
COMPLEX STRESS DISTRIBUTIONS IN ENGINEERING MATERIALS. 173 
Guest 3 conducted by far the most important early research. He dis- 
tinguished between brittle and ductile materials. Thin tubes of steel, 
copper, and brass were tested to yield under combinations of tension, torsion, 
and internal fluid pressure. The principal stresses were two tensions, or 
one tension and one compression, with the third always very small. An 
abstract cannot do justice to Mr. Guest’s paper, which raised the experi- 
mental side of our subject to a higher level, and has directly suggested 
much of the more recent research. He concluded that the condition for 
initial yielding of a uniform ductile material is the existence of a specific 
shearing stress, and that the intermediate principal stress is without effect. 
Coker ® studied iron and steel under torsional stress, and included some 
data in relation to torsion with tension or bending. This paper indicates 
the general character of the effect of tension and bending on a specimen 
subjected to torque. 
Wehage* presented no new experimental results. He contended that 
although two tensile stresses at right angles counteract one another when 
the extensions are considered, their destructive effect on the material is 
really superposed. The previous experiments of Guest prove him to be 
wrong. Mohr’s Theory is quoted as only considering the stress normal to 
the plane of maximum shear. 
Hancock *® 38; 48, 46, 55 first tested solid steel rounds, and then steel 
tubing, in tension and torsion. His results have been adversely criticised, 
and certainly supported neither hypothesis, although the author favoured 
the shear stress theory. Later tests under tension or compression with 
torsion indicated that the maximum tension was seldom greater than the 
tensile strength of the steel, but the maximum shear stress was often greater 
than its shearing strength. 
Izod*" tested materials to fracture in double shear. The discussion on 
his paper makes it clear that shearing tests of this type are complicated by 
cross stresses after yield. The ratio of shear to tensile strength was 0°62 
to 0°78 for iron and steel, the tensile strength being the maximum load 
divided by the original area of the cross-section. The results were con- 
firmed by Goodman. Lilly held the surprising view that only under ex- 
ceptional conditions was the shear strength less than the tensile, and that 
isotropic materials were strongest in compression, next in pure shear and 
weakest in tension. There was a rough indication that ductile materials 
followed the shear rather than the maximum stress law of failure. 
Frémont *® modified the usual shearing test by filing away the sheared 
face from time to time to eliminate the friction between the steelings and 
the sheared faces. He then found that the shear stress at fracture was 
about 0-4 times the maximum tensile stress for irons and steels which had 
a range of tensile strength from 19 to 65 tons per square inch. 
Scoble *° *! employed combinations of bending and torsion on solid 
round steel bars. Yield was taken as the point of failure. The maximum 
shear stress varied from 29,170 to 33,500, and the maximum principal 
stress from 29,170 to 64,600 Ib. per square inch, having the low values in 
pure torsion. The bending moment was not constant over the length of 
a bar, which would tend to mask the yield and give a high stress 
under bending. It was concluded that the maximum shearing stress was 
approximately constant at yield, but it was also shown that engineering 
materials ave not perfectly isotropic, and consequently have different 
shearing strengths in different directions. Later tests included steel and 
