COMPLEX STRESS DISTRIBUTIONS IN ENGINEERING MATERIALS. 203 
Three examples are given which illustrate the conclusion. These are 
(1) On the Pennsylvania Railroad, many fatigue failures of axles were 
obtained with acid open-hearth axles containing 0-25 to 0-28 per cent. carbon, 
with a tensile strength of 29 tons per square inch, and 25 per cent. 
elongation. The maximum calculated stress in the middle of the axle was 
6-8 tons, and in the journal 3 tons, per square inch. These axles failed 
in the journal. Steel of about 36 tons tensile was then substituted, and 
this cured the trouble. 
(2) 0-22 to 0-25 per cent. carbon steel rollers for a sugar mill used to 
break, and these were successfully replaced by rollers of 0-40 to 0-45 per 
cent. carbon steel. 
(3) A soft tough steel was successfully replaced by a higher carbon steel 
for use in the form of piston rods for steam hammers. 
On the above subject there is, and perhaps always will be, much diver- 
gence of opinion. The chief reason is that conditions vary so greatly. The 
treatment of the steel is obviously a very important factor. The axles, 
for example, if treated at all, could have been treated so as to give different, 
and probably much better, results, and to some users this would have formed 
a more acceptable solution of the problem. One has to allow, amongst 
other things, for considerable shock and some ill-treatment. This is em- 
phasised by the fact that the calculated stress in the journals is only 3 tons 
per square inch. No Wohler test results are given. 
22 Eden, E. M. 1910 Endurance of Metal under Alternating Stress and 
Effect of Rate of Alternation on Endurance. ‘ Univ. 
of Durham Phil. Soc. Proc.,’ 3, 5, 1910. ‘Sci. Abs.,’ 
1910, No. 1384. 
23 Eden, E. M., 1911 he Endurance of Metals. ‘ Proc. Inst. Mech. Eng.,’ 4, 
Rose, W. N., Oct., Dec. 1911. ‘Sci. Abs.,’ 1912, No. 1145. 
Cunningham, 
F. L. 
The main results of these well-known experiments were :— 
No ‘ speed effect? between 250 and 1,300r.p.m. ; agreeing with Nos. 43, 
80, 65 and 82. . 
Greater resistance of high tenacity steels ; agreeing with Nos. 93, 82, 62, 
90, 47 and 48. 
Tests with loaded rotating solid bars appear to give either a higher 
limiting range of stress (calculated by the usual theory of bending) or require 
a larger number of cycles to fracture than direct-stress experiments; a 
result in agreement with Nos. 47,48 and 93. The tests of these papers 
(Nos. 22 and 23) were carried to fracture or to 10° revolutions. In No. 43 
a small difference only was found in some comparative tests on machines 
of the rotating bar and reciprocating mass type. 
Rest intervals during a test appear to have little effect. This appears to 
be the case in all tests with cycles of equal + stresses. 
The effects of the kind of finishing process used in preparing the speci- 
men, and of the kind of finished surface, are found to be important. See 
also No. 51. 
A few cast-iron and copper specimens were tested. The relative resist- 
ance of certain forms of specimens was tested, with results in agreement 
with Nos. 74, 75, 82 and 93. 
24 Ercolini, G. 1906 Effect of Deformation upon Torsional Couple exerted 
bya Twisted Wire. ‘ Accad. Lincei,’ Atti 15, Sept. 2, 
1906. ‘Sci. Abs.,’? 1906, No. 1807. 
Some experiments with combined stress and with alternating combined 
stress on copper wire. The strains appear to have been considerably beyond 
the elastic limits. 
25  Ercolini, G. 1909 Recent Experiments on Elasticity. ‘Sci. Abs.,’ 1909, 
No. 965. 
The following is quoted from ‘Science Abstracts’: ‘It is concluded 
that the damping of vibrations is due to the dissipation of energy corre- 
