328 REPORTS ON THE STATE OF SCIENCE, ETC. 
table, from which it is deduced that fatigue occurs when the alternating stress A 
exceeds 14tons/in.”, in combination with the steady stress of 18 tons/in.?, corresponding 
to a maximum stress of 32, and a minimum of 4 tons/in.?. 
TaBLeE I.—FatTIGuE TESTS ON UNPIERCED PLATES. 
S = 18 Tons/In.? 
Specimen Endurance, 
Number Ai Dons/En Millions of Cycles. 
| 
1 12 2-000 Unbroken 
1* 14 | 1-208 Cracked 
2 14 0-722 Cracked 
3 13 0-872 Cracked 
4 13 0-526 Cracked 
5 13 8-572 Unbroken 
5* 15 0-872 Cracked 
6 14 0-964 Cracked 
7 13-5 7-778 Unbroken 
7% 15 0-910 Cracked 
8 14 16-826 Unbroken 
8* 14-5 1-826 Cracked 
9 14 | 3°040 Cracked 
10 15 9-818 Cracked 
Failure occurred in each instance by the development of a fatigue crack from some 
point on the parallel middle portion. In most cases the cracks started from the edge ; 
but in others the cracks started in the middle and ran outwards. 
In the second series of tests, the strips were in all instances pierced with a single 
circular hole, drilled on the centre-line and at mid-length. In the first few tests, the 
width of the strip was reduced at mid-length as in the first series, from 14 in. to 1 in. 
or to $ in.; but, when it was found that satisfactory fractures could be obtained 
without this reduction, subsequent tests were made on strips of the full width, 1} in. 
The reduction in fatigue strength was so great that the difficulty experienced in gripping 
the ends in a satisfactory manner was less than in the case of the tapered but un- 
pierced strips. 
The loads applied in different tests were such that, in all cases but one, the 
alternating component load bore a fixed ratio, namely 14: 18, to the steady tensile 
load. This being the ratio obtained in the limiting condition for the unpierced plate, 
it follows that if the material was perfectly elastic, this same ratio would hold between 
the alternating and steady components of stress throughout, and that at the point 
where failure starts, the actual stresses would be S = 18 tons/in.?, A = 14 tons/in.?. 
If s, a, were the components of applied stress—at a point remote from the.hole— 
which proved to be the limiting value causing fracture, then 18/s (= 14/a) would be 
the value of the maximum ratio of stress-concentration due to the hole, and if the 
elastic theory held this would be equal to 3. Thus the maintenance of the applied 
alae in this ratio enables the results to be checked at once against the theoretical 
values. 
In recording the results in the following table, the applied stresses have been quoted 
indirectly through the numerical factor N, such that 
a=Applied Alternating Stress = 14/N tons/in.?. 
s=Applied Steady Tensile Stress = 18/N tons/in.?. 
hus N affords a direct indication of the reduction applied in consideration of the 
presence of the hole. 
