172 REPORTS ON THE STATE OF SCIENCE.—1913. 
PREVIOUS RESEARCHES. 
7. Early Engineering Tests. 
Reference has been made to the important experiments of Tresca,® 
which led him to support the stress difference theory of rupture for ductile 
materials, and his conclusions have been generally accepted. Most 
modern researches have differed from Tresca’s because attention has been 
concentrated on the yield point or the elastic limit of the material. 
Experiments on steel by the Committee of Civil Engineers are of 
interest because the tension and compression specimens were 15 inch 
diameter and 10 feet long, so that the longitudinal strains could be mea- 
sured accurately. There was little difference in the stress at the yield 
poimt in any case, a result which supports the contention that the co- 
efficient of internal friction for steel is zero, and that Theory (c) is a par- 
ticular case of (d) to suit ductile steels. 
Hodgkinson * found that cast iron fractured at stresses of 7 tons per 
square inch in tension, and 24 tons per square inch in compression. ‘lhe 
form of the fracture in compression in common with those of other brittle 
materials has led to the acceptance of I’heory (d) under these conditions, 
in the absence of direct experimental evidence. It is remarkable that the 
great differences in strength and form of fracture did not lead to an earlier 
recognition of the possibility of two limits for failure, at least for brittle 
substances. 
The tests of iron and steel in different ways by Appleby and Kirkaldy"® 
were of great engineering importance, but since results are given for 
fracture they lead to no definite conclusions for our purpose. ‘This applies 
also to the later work of Platt and Hayward," which included tests of cast 
iron in tension, torsion, and shear, because Scoble #4 has concluded that cast 
iron yields too much before fracture to allow the elastic formula to be used 
to calculate the true breaking stress in torsion. The results of the shearing 
tests are not acceptable. 
8. Hxperiments with Ductile Materials under Combined Stresses. 
Wehage * tested circular steel and wrought-iron plates supported round 
their edges and loaded at their centres. ‘I'he extension at the elastic limit 
was about half that in simple tension. 
Carus Wilson’s paper®® on ‘The Rupture of Steel by Longitudinal 
Stresses ’ describes an attempt to test the stress difference theory (c) 
employed by Darwin.'? Tension specimens of rectangular cross section 
were tested to fracture when plain, with a‘ V’* notch, and with a ‘ U’ notch 
oneach side. He used the true mean stresses calculated on the contracted 
areas at fracture. The ‘V’ notched specimens were weaker and the 
‘U’s’ stronger than the original bar. ‘he notches caused an uneven 
stress distribution across the breadth of a bar, which tended to weaken it ; 
they also reduced the tension area more than that which resisted the shear. 
He concluded that the material fractured by shearing. He also found 
that the shear stresses in tension, and in double shear tests agreed very 
well if the true stresses at tensile fracture were taken. 
Foppl™ tested materials under a uniform pressure of 50,000 lb. per 
square inch and also with compression in two directions, and a third 
principal stress absent. His experiments led to no definite conclusions, 
except that the uniform pressures applied did not cause rupture. 
