THROW-TESTING MACHINE FOR REVERSALS OF MEAN STRESS. 
279 
Method of Fixing the Specimens. 
A lock-nut was screwed on to each end of the specimen, a hardened steel ball, 
h inch diameter, was inserted in the upper chuck H (fig. 5), and that end of the 
specimen which contained the centre mark was screwed into 
this chuck, but not screwed home. The oscillatory weight 
was then next brought up and the chuck J, which contained 
another hardened steel ball, was screwed on the lower end of 
the specimen until the specimen bedded against the ball, the 
specimen being prevented from rotating by means of a pair of 
gas-tongs with whicli the short parallel part m was gripped. 
The lock-nut n was screwed tight, thus fixing the specimen 
to the oscillatory weight. 
J c> 
The weight was now supported, and the specimen screwed 
up so as to bed against the steel ball in the upper chuck, the 
small force necessary for this being supplied by gas-tongs, 
with which, in this case, the parallel part l was gripped. 
The lock-nut r was then screwed tight, thus fixing the 
specimen to the chuck H. 
By the above method, one was certain of getting the 
specimen into the machine without straining it, whereas, had 
the specimen been fixed first to the upper chuck, it is quite 
possible that the material would have been subjected to 
severe torsional strains in connecting up the oscillating 
weight. 
In the tests carried out with the apparatus driven by the 
oil engine, the steel balls were not used, and they were also 
discarded in the tests on wrought iron and cast steel, since 
it was found that the specimens could be easily locked in the 
manner described above without using the balls. 
In a great many of the tests the specimens were not 
prevented from rotating, for, as previously explained, it was found unnecessary to 
do so, but when necessary, to prevent oscillations, the lower bearing was locked by 
means of a small brass set-screw k, \ inch diameter, which was screwed so as to press 
against the outside of the loose bush which formed this bearing, thus by frictional 
force preventing torsional oscillations. 
Method of Conducting Tests. • 
The boiler fire was generally made about 8.30 in the morning ; steam was up and 
the engine was started a little before 9.30. During this time the oil supply pipes of 
