THROW-TESTING MACHINE FOR REVERSALS OF MEAN STRESS. 
277 
The number of free vibrations per minute of the specimen vibrating transversely 
was calculated, considering the specimen as a bar fixed at its end, and found to be 
about 500,000 per minute, so that in this case also the free vibration need not be 
considered as influencing the results. 
The number of vibrations per minute executed by a specimen when oscillating 
torsionally was also estimated. The calculation gives for the heaviest and lightest 
loads used 1200 and 2800 vibrations per minute respectively, so that it is quite 
possible that the free period of torsional vibration of the specimen might coincide 
with the speed of the machine. 
The key attached to the vertical sliding spindle, which works in a key-slotted bush 
which can be adjusted and locked, prevents such vibration above a certain amplitude 
taking place, and they are damped also by the viscosity of the oil in the bearings of 
the spindle. Still, it is evident that this key and keyway cannot be fitted so 
accurately as to completely extinguish a twist in the specimen as it slides with the 
weight. 
With a view to eliminating the effect of the torsional vibrations of specimens, a 
number of tests were carried out under different conditions. It was found that when 
the speed corresponding to the free period of torsional vibrations was reached, a 
change in the moment of inertia of the load seemed at once to eliminate the 
vibrations, whilst the unlocking of the lower bearing greatly increased them, causing 
the specimen to break with fewer reversals. When the speed did not correspond 
with the free period, neither the change of moment of inertia nor the conditions ot 
locking effected the results (see footnote, Table III., set C). 
The results of these tests are given later (p. 289), and it is seen that only when the 
free vibration of the specimen coincides with the speed of the machine has this 
vibration any influence on the results. 
Preparation of the Specimens. 
The materials used in the tests, of which the results are given in this paper, were 
mild steel, best cast steel, and best Lowmoor iron. By far the greater number of 
tests were carried out on specimens 
having dimensions given in fig. 4. 
Bars f inch diameter were cut up 
into short lengths of 6 inches. The 
centres were marked and small holes 
were drilled up these centres for each 
piece, and the pieces were then square- 
centred in the ordinary w r ay. A rough 
cut was then taken over the whole length, and the cutting of the screw, over 
5 inches in length, was then commenced. This part was then finished by means of a 
