PROFESSOR OSBORNE REYNOLDS AND MR. J. H. SMITH ON A 
29 G 
The range of stress for one million reversals, which these tests of Baker seem to 
point to, lies between 27 and 28 tons. 
Taking the results given in Table V. along with those of \\ ohler and Baker, we 
o O o 
see that in the case of cast-steel there is a great lowering of the range of stress—for 
rupture with a given number of reversals—as the speed is increased. 
In the following table the author’s results are added to those of Wohler and 
Baker in order to show the relation between the range of stress and the reversals 
per minute for rupture with one million reversals :— 
Cast Steel. 
Range of stress for 
rupture 
with 10 u reversals. 
Reversals per minute. 
Ratio of range for 
10 IJ reversals to yield 
stress. 
30-9 
60 to 80 (Wohler) 
27-5 
50 to 60 (Baker) 
— 
20-1 
1320 
•50 
18-3 
1660 
•46 
16-8 
1820 
•42 
13-1 
1990 
•33 
In the case of cast-iron the range of stress for one million reversals obtained in the 
author’s experiments is approximately 5'5 tons at 1300 revolutions per minute. 
Wohler obtained 478 tons as the range for one million reversals for bars subjected 
to repeated tensions, the limits being 0 and 478 tons; if we assume that cast-iron 
behaves in the same general way as wrought-iron and steel, Wohler’s limit would 
have been much greater if the stress range had been between equal and opposite 
limits, pointing possibly to the same lowering of the range as the speed increases. 
It is, therefore, impossible, in the case of cast-iron, to say definitely whether the 
range is diminished as with the other metals experimented upon. 
Conclusion. 
There are many points which the author would have liked to investigate, but was 
unable to owing to the great amount of time which would be required. The only 
satisfactory method of procedure with experiments of the kind dealt with in this 
paper is to carry out a large number of tests bearing upon any particular point, in 
order to eliminate the effects of irregularities or inequalities of the materials of which 
the specimens are composed. It is only in this way that one can be certain of 
avoiding the inclusion of anomalous results among those from which the deductions 
are made. 
A little time had been spent on the effect of annealing specimens after subjecting 
