IRON AND STEEL UNDER CYCLICAL VARIATIONS OF STRESS. 39 



steel, similar in characteristics to No. 2 of the same paper, was taken as representative 

 of hard material, whilst finally a bar of axle steel of intermediate resistance was 

 obtained to complete the tests. 



Both the harder materials were found to be exceedingly uniform in all respects, but 



- the yield stress for the Swedish iron varied with the position in the bar from which 



the specimen had been cut. This variation had no apparent effect on the values of 



the elastic ranges, but the tests were discontinued when the change had become 



considerable. 



The specimens, fig. IA, were necessarily of small diameter and length. The portion 



Fig. IA. 



under test was half an inch long in the parallel part, and the diameters, d, varied 

 from 0'170 inch to '2 50 inch. ^ 



The Observations. In the present experiments the measurement of the length of 

 the specimen under the actual conditions of test has constituted an important new 

 feature in the history of fatigue testing. Before the application of any stress the 

 extensometer was fixed in position and a reading taken ; this gave <i zero from which 

 all the changes of length were calculated. 



As the changes which occur due to repetitions of stress are somewhat slow, 

 observations were only taken at intervals of hours, when the cumulative effect of small 

 variations had become large enough for measurement. The extreme points in any 

 cycle could be observed without stopping the machine, and generally the readings at 

 these extremes, i.e. at the maximum and minimum lengths of the specimen, were the 

 only ones taken. 



Occasionally the machine was stopped, and, starting from the minimum stress, the 

 load was increased by successive increments to the maximum stress, the extension 

 being read at each step. The load was then decreased and the corresponding extension 

 measured in a similar manner. Plotting the extensions on a stress base then gave 

 the straight line or loop which indicated the condition of the specimen as to 

 elasticity. 



It will be shown later, when the actual figures are described, that the maximum 

 width of any loop, measured parallel to the axis of extension, added to the elastic 

 extension of the specimen, is the total change of length in each cycle. 



The complete plotting of the hysteresis loop was somewhat objectionable, as it 

 meant an interruption of the usual running of the machine, and only a few complete 

 loops were observed for each specimen. The intermediate readings are represented in 

 the diagrams by the maximum width of the loop, defined as " cyclical permanent set," 



