78 



SCIENCE 



[N. S. Vol. XL VII. No. 1204 



stages of slip, in the hope that as we ap- 

 proximate the true elastic limit we shall 

 thereby approximate the fatigue strength 

 as well. 



Two methods which might lead to much 

 closer approximations than can be expected 

 from refining our direct measurement of 

 the changes of dimension by means of ex- 

 tensometers, are the thermal and the mag- 

 netic. 



In the thermal method, the fatigue 

 strength is determined as the stress of 

 which rapid repetitions cause a detectable 

 rise of temperature, supposed to represent 

 the heat evolved by the friction along the 

 first slipping planes. This method owes 

 its attractiveness to our conception that the 

 thermo-electric measurements of tempera- 

 ture seem intrinsically capable of greater 

 sensitiveness than the direct measurement 

 of length. On further examination the 

 method loses some of its promise, for 

 though a slip over a very minute area might 

 yield enough heat to be detected if it lay at 

 the very outside of the specimen, if it were 

 deep-seated it might not, because the heat 

 in working outwards would spread itself 

 over a very large area of surface, and this 

 would lessen correspondingly the actual 

 rise of temperature. Hence if we took 

 the stress which causes the first detectable 

 rise of temperature as our best approxi- 

 mation to the fatigue strength, our results 

 might vary greatly with the position of 

 the first slipping area. 



The magnetic method would determine 

 the fatigue strength as the least stress of 

 which repeated applications cause a detect- 

 able change in the magnetic properties. 

 The developments of this method which are 

 now occurring at the Bureau of Standards 

 deserve our most careful attention. 



From the fact that the ability to endure 

 many millions of repetitions of stress is 

 much less than the observed elastic limit. 



and from our natural explanation that the 

 difference represents some form of progres- 

 sive degeneration under indefinitely re- 

 peated stresses, we naturally infer that 

 when the repetitions are in the hundreds 

 of thousands a like but smaller degree of 

 this same degeneration occurs. From this 

 in turn we infer that even when the repeti- 

 tions are only in the thousands this degen- 

 eration may occur, though in correspond- 

 ingly smaller amount. From this we gen- 

 eralized that the gap between the observed 

 elastic limit and the safe working load 

 should increase as some function of the 

 number of repetitions of stress to be ex- 

 pected during the life of the member, and 

 we ask whether this increment should not 

 be considered even in designing bridges, 

 and whether it should not increase with 

 their expected life and the frequency of 

 the passage of trains. 



A natural question which arises in this 

 connection is whether, as regards fatigue 

 strength, that part of the elastic limit of 

 steel which is caused by a low finishing 

 temperature in rolling is valid or fictitious 

 and removed during the exposure to re- 

 peated stresses. The great life of piano 

 wires, strained very severely and exposed 

 to a repetition of stress with each sound 

 wave, and of suspension bridge wires, both 

 of which owe much of their elastic strength 

 to cold deformation in the shape of wire 

 drawing, is reassuring. An interesting 

 ease is that of piano wire, which is credibly 

 reported to have sprung back into a helix 

 when cut after fifty-four years' service, 

 showing that it retains even through its 

 enormous number of repetitions of stress 

 the bending given it in coiling it. 



Struggling as we are for our national 

 existence, these pale thoughts move us 

 hardly more than the remembrance of his 

 last month's rent stirs the shipwrecked 

 swimmer in his landward struggles. 



