164 REPORTS ON THE STATE OF SCIENCE. — 1915, 



ratio between the maximum stress of the tensile test and the hardness 

 is 0"40 to 0'38, according to the value taken for the tensile test. The 

 specimens used in the tests which are now described were not annealed, 

 but were tested in the condition in which the metal was received. 



Alternating stress tests were carried out in a machine of the type 

 described at the Dundee Meeting of the Association, 1912.^ In this 

 machine the alternating stress applied to the specimen is obtained by 

 combining two pulsating forces (derived from two-phase magnets) 

 acting upon opposite sides of a single armature to which one end of the 

 specimen is rigidly coupled. The inertia force required for the oscilla- 

 tory acceleration of the armature is cancelled by the force of de- 

 flection of a spring which is adjusted to suit the frequency of the 

 test. The load applied to the specimen is determined by measuring 

 the voltage induced in a fine wire secondary coil, wound close to the 

 pole faces of the magnets. To standardise the machine, a phosphor- 

 bronze specimen, fitted with an extensometer, and previously tested 

 under steady stress, is inserted in the machine and stressed by the 

 application of an assigned voltage at the desired frequency. The range 

 of extension measured under alternating stress may be used as the basis 

 of measurement, or, alternatively, the alternating stress may be con- 

 verted to a steady stress by reversing the pull of one of the two mag- 

 nets. Stroboscopic observations showed that the stress applied to the 

 specimen varies very approximately in a sine wave, and it is estimated 

 that the load range of stress can be determined with an accuracy within 

 1 per cent. Combinations of alternating and steady stress are obtained 

 by extending or compressing the spring used for compensating the 

 inertia of the armature. The stiffness of this spring having been 

 measured, the magnitude of the steady stress is determined by measure- 

 ment of the deflection of the spring. 



The form of the specimens used in the alternating stress tests is 

 shown in Pig. 3. The maximum stress is developed in the central 

 cylindrical portion of the piece, about i inch in length, which is joined 

 to the conical ends by gradual transition curves. The ends of the 

 specimens are screwed, and the necessary precautions are taken to 

 ensure that the test-pieces are free from initial stresses, either torsion 

 or bending, due to the tightening of the grips. 



The tests may be arranged in seven series, and are summarised m 

 the table given below. In the first series the stress alternated between 

 equal intensities of tension and compression. In three series, II. to 

 IV., steady stresses of tension were applied in combination with the 

 alternating stress, and in the final three series, V. to VII., steady 

 stresses of compression were combined with the alternating stress. The 

 frequency of stress employed was the same in all series — viz., 2,000 

 per minute. The duration of the tests followed by fracture varied from a 

 very few cycles up to 8 million. Other tests were continued beyond 

 this number, in one instance to seventeen million cycles, without 

 fracture. 



In Fig. 3 the range of stress is plotted on a base representing the 



" See also Engdneering, November 22, 1912. 



