4 Foster, Phenomena due to Repetitions of Stress. 



Evidently when the load is removed the extension 

 begins to disappear at first very rapidly, then more and 

 more slowly, somewhat as in Fig. 2. 



TIME 



Fig. 2. 



If now, before the bar has recovered, the cycle of 

 loading is repeated we should expect the lag at the end 

 of the second cycle {Fig. i) to be greater than at the end 

 of the first cycle (though probably less than twice the 

 first), and so on until ultimately the total extension 

 (which includes the total lag) is equal to the maximum 

 elastic extension of the bar. At this point any further 

 extension will be permanent and show itself by slipping 

 along surfaces in the crystalline grains. Professor 

 Ewing's results previously mentioned confirm this view. 



