692 PROCEEDINGS OF SECTION H. 
ence points into spaces of one quarter of an inch, and con- 
nected to the autographic apparatus, and tested to destruc- 
tion. Theyield point recorded in the autographic diagram is 
consequently higher than would have been the case if the 
test piece had not been previously strained in obtaining the 
true elastic limit. The results of these tests are shown in 
Tables I. to V., and in the Summary Tables X. and XI. 
Compression T ests.—These consisted of the determination 
of the elastic limit and coefficient of elasticity as in the 
tensile tests, the compressions being observed with the 
Marten’s mirror extensometer. The length of the specimen 
was 10 inches, and the diameter 1 inch. One set of readings 
was taken on a specimen 2 inches long. The compression 
strength was determined by using cylinders 1 inch in 
diameter and 1 and 2 inches long respectively. The com- 
pression strength was taken as the yield point of the test 
piece. See Tables VI. to IX., and Summary of Results, 
Table XII. 
Alternating Bending Tests.—These tests were made on 
turned specimens, reduced along the length over which the 
bending moment and fibre stress was constant, and having 
rounded shoulders. The machine was designed and made 
in the author’s laboratory, and is fully described and illus- 
trated in Mr. Madsen’s paper, “ Investigation of the Effect 
of Alternating or Universal Repetitive Stresses upon the 
Physical Properties of Materials.’”’ The extreme fibre stress 
produced in the specimens of nickel steel tested was 54,085 
pounds per square inch, and the number of repetitions 
varied from 13,800 to 29,050. The bars were afterwards 
tested in tension in the usual way, and the elastic limit 
determined with Marten’s mirror extensometer, as in other 
tests of the same material which had not been rotated. 
The results are recorded in Table XVI., which also shows 
the effects of raising the temperature of nickel steel on its 
elastic limit, strength, and ductility. The lowering of the 
elastic limit by repetition of stress is clearly marked in the 
mild variety of nickel steel; but in the two harder qualities 
the effect is less noticeable. Probably the form of the test 
piece is not the best for the purpose. Experiments on the 
effect of repeated stresses on nickel steel are still in progress. 
Impact Tests —These consisted of tension tests in which 
the load was suddenly applied, and in transverse tests on 
nickel specimens supported on bearings 6 inches apart. 
The impact machine used Was similar to the one in the 
Royal Testing Laboratory, Berlin, for testing specimens. 
The weight of the hammer used was 122°5 lbs. for tension 
