THROW-TESTING MACHINE FOR REVERSALS OF MEAN STRESS. 
281 
attaining full speed to breaking was deduced and the total reversals were estimated 
from this and the mean speed. 
In the case of very short tests the speed was found either before or directly after 
the test, in which case the oscillatory weight was connected up by a specimen of 
large diameter specially kept for such work, and the machine was run for 10 minutes 
to the mark on the speed indicator at which the test was done. 
It is thus seen that the reversals of stress to which the materials were subjected 
during the rise of speed were not taken into account. This might make a slight 
difference in the short tests, but could certainly have very little effect on the results 
in the case of long tests, namely, those extending over a period greater than two or 
three hours. To reduce this effect as much as possible, short tests were never 
carried out without previously running the machine for some time at the speed at 
which they were done. 
If, during a test, the lubrication of any of the bearings of the engine, shafting, or 
machine failed, it was at once detected, since the speed indicator column on the 
engine fell and the pitch of the (unmusical) note given out by the machine was 
lowered. In such cases, the machine was at once stopped and the time taken ; the 
bearing was attended to, and a new start made. 
In the case of very long tests, the machine was run for five or six hours each day 
according to convenience. The engine could not be used for driving on Tuesday or 
Wednesday afternoons, as it was employed during those times by the students for 
the experimental trials, which form part of the College Laboratory course ; so that a 
test which required ten or twelve hours to complete could not be conveniently carried 
out without a stop which, in some cases, extended over two or three days. 
It was important to find the effects of these periods of rest on the number of 
reversals required for rupture, and after several preliminary tests were completed, a 
series of experiments were undertaken to investigate this matter. An account of 
these experiments is given later (see p. 283). 
Preliminary Tests for Mild Steel. 
A great number of preliminary tests were carried out during the time in which the 
oil-engine was used as the source of j^ower and also on first using the steam-engine. 
In some cases the oil-engine had an ordinary pulley and cord brake attached, so as to 
cause the engine to work at full load, thus reducing the fluctuations of velocity, 
whereas in other cases it was used to drive the machine without any brake. 
The ordinary statical test for this material gave the following results - 
Yield stress.18 ‘64 tons. 
Maximum stress. 25 - 83 ,,, 
Breaking stress.22'09 ,, 
Percentage elongation at maximum stress . . 22. 
rupture . . . . . 29, 
2 o 
VOL, CXCIX.—-A, 
