136 Prof. J. A. Ewing. On Measurements of small [May 16, 



Primitive Annealed State (3A). 



Mean extension from 1 to 4 tons = 94' 5 per ton ; 

 E = 12500 tons per square inch. 



Immediately after passing yield-point (3s) , the diameter was 

 reduced by the set to 0'693 inch, and the section to 0'377 square 

 inch. 



Extension for first ton = 97 ; E = 12310. 



Mean extension from to 2 tons = 97'5 ; E = 12250. 

 Mean extension from to 4 tons = 100'2 : E = 11910. 

 Mean extension from to 5 tons = 101*4 ; E = 11770. 



After elastic recovery by resting for 5 days (3c). 

 Mean extension from to 6 tons = 97 ; E = 12310. 



After a small amount of further overstrain (by applying 7 tons). 



Extension for first ton = 97 ; E = 12310. 



Mean extension from to 6 tons = 99 ; E = 12060. 



Similar results have been obtained in tests of other bars, of mild 

 steel, comparatively hard steel, and Lowmoor iron. In every case, 

 overstrain has produced a like fatigue of elasticity, and elastic 

 recovery has followed during an interval of some days or weeks of 

 rest. 



In several examples the process of passing the yield-point has been 

 watched in the extensometer and has been seen to take place in the 

 same manner as in test (3A). Creeping, which has been slightly 

 visible under lower loads, takes place at first slowly, then it gradually 

 gets faster, though no change is made in the load. This may go on 

 for a minute or two before the crepitation of oxide begins. The 

 following is an instance noticed in testing a bar of mild steel. The 

 load having been increased a step, and kept constant, there were 



3 scale-divisions of creep in the first minute 

 9 second 



320 third 



after which the index of the extensometer raced off the scale. In 

 another case the writer noticed creeping going on quite slowly, when 

 a sound of crepitation was heard. On looking at the bar it was seen 

 that the crepitation had begun near one end, beyond the clips. As 

 soon as it spread far enough to reach the clip there was, of course, a 

 rapid movement of the index. 

 The following test will suffice in further illustration of the 



