40 MR. LEONARD BAIRSTOW ON THE ELASTIC LIMITS OF 



this width having been obtained by subtracting the calculated elastic extension from 

 the total extension observed. 



In addition to the cyclical changes of length just mentioned, non-cyclical extensions 

 occurred, which took the form of a more or less rapid yielding of the material. It 

 was in order to observe these non-cyclical extensions that the extensometer was kept 

 in position during the whole test.* The change of length of the specimen which was 

 non-cyclical has been defined as " permanent extension." 



A detailed reference to a few actual readings will make the method of reduction 

 clear. - The new specimen was subjected to increasing loads, and for the first part of 

 the test the extension was proportional to the load. From the modulus so obtained, 

 the elastic extension for any range of stress could be calculated on the assumption 

 that no change of modulus occurred. For a particular specimen of axle steel, the 

 elastic extension was -'344, the unit being the 10,000th part of a millimetre. 



In the following table the actual scale readings for the same specimen are given at 

 10, 480, and 53:30 repetitions : 



At 10 reversals the scale reading for the left-hand mirror was 1068, when the 

 minimum load was applied and gradually increased during the cycle until it became 

 1240 at the maximum load. The change of reading for this mirror was therefore 172. 

 The change for the right-hand mirror was 171. The sum of these, i.e. 343, was the 

 change of length during the cycle. This was almost exactly equal to the elastic 

 extension of 344, and the effect of the reversals was not measurable. 



All four readings were repeated after 480 repetitions, so that no change had been 

 registered up to that time. 



The readings at 5330 repetitions showed great changes in every case. The left- and 

 right-hand mirrors now moved 190 and 188 divisions respectively, making a total 

 change of length of 378. Of this, 344 is elastic and the remainder, 34, has been 

 plotted in fig. 3 as the " cyclical permanent set " at 5330 repetitions. 



The reading of 1900, obtained from the left-hand mirror at the maximum stress in 

 the 5330th cycle, differed by 660 from the reading of 1240 obtained at the same 



* The small amount of creeping of the extensometer which occurred appeared as a change of length. 

 When suspected, the knife-edges of the extensometer were sharpened and a second specimen tried. The 

 creep was generally unimportant. 



