January 25, 1918] 



SCIENCE 



77 



and that here the teachings of the impact 

 test would not lead to economy. That cast 

 iron is strong enough for sash weights, fly- 

 wheels, and bedplates does not prevent our 

 heat-treating alloy steels for aeroplanes. 



Turning to our second subject, my asser- 

 tion that we do not determine the true re- 

 sistance of our material even to relatively 

 quiescent stress refers to fatigue strength, 

 the resistance to stresses applied repeatedly. 

 Of the four properties which we habitually 

 determine, the elongation and reduction of 

 area certainly throw little light on fatigue 

 strength, nor can we expect much from the 

 tensile strength, which represents the ex- 

 altation of the elastic limit induced by the 

 plastic deformation after the initial elastic 

 limit has been passed. This deformation 

 can not be tolerated in hypo-elastic serv- 

 ices, that is in those which demand that 

 each member must retain its initial size and 

 shape, and here the ability of the material 

 to undergo the resultant exaltation of its 

 elastic strength is useless, save as forming 

 a basis for calculating the ductilit}' from an 

 additional point of view, through the elastic 

 ratio, and thus supplementing the elonga- 

 tion and reduction of area as suggestions 

 of impact resistance. How is it then with 

 the fourth of these properties which we 

 determine, the elastic limit? 



Before answering this, it is well to re- 

 fresh our conception that, because the in- 

 numerable ferrite and cementite grains of 

 our steel lie with their slip planes inclined 

 in every possible direction, and because the 

 stresses are not distributed exactly evenly 

 throughout the member or test piece, there 

 is some one slip plane in some one grain 

 that is less favorably disposed than any of 

 the others towards the existing stress. 

 That is the weakest spot in the bar. If 

 we increase the stress gradually, slip occurs 

 along that slip plane before anywhere else. 

 The failure of this metal to bear its load 



tends to overload its neighbors. The stres? 

 which causes slip along this weakest plane 

 is strictly speaking the true elastic limit, 

 because this slip changes the dimensions 

 of the bar permanently. Perfect elasticity, 

 the power to return exactly to the initial 

 dimensions, exists only below this limit of 

 stress, which hence by definition is the 

 elastic limit. That this first slip can not 

 now be recognized and probably never can 

 be affects this inference as little as our 

 inability to see atoms and molecules inter- 

 feres with our belief in their existence. 



The elastic limit which we determine 

 with our extensometers may, for clearness, 

 be distinguished from the true elastic limit 

 by calling it the "observed elastic limit." 

 That is must needs recede as our extensom- 

 eters become more sensitive and must thus 

 approach the true elastic limit asymptot- 

 ically, is clear. 



If this stress which has caused the first 

 incipient slip is released, the elasticity of 

 the rest of the metal reverses the slip, and if 

 the stress is quickly reapplied new slip 

 recurs along this same plane, and so on 

 with quickly succeeding applications of 

 stress, according to our conception, which, 

 moreover, holds that tliis repeated slipping 

 back and forth causes local degeneration. 

 Moreover, the overloading of the adjoining 

 parts by the slip along this one slipping 

 plane leads them in time to begin slipping, 

 so that after many repetitions of the cycle 

 this degeneration extends through so much 

 of the section that the remaining sound 

 metal is unable to sustain the stress, and 

 hence breaks. This leads naturally to the 

 conception that the true elastic limit is 

 the fatigue strength. 



Without insisting on the accuracy of this 

 picture, it certainly helps u.s to understand 

 why our present observed elastic limit does 

 not measure the fatigue strength, and 

 stimulates us to determine much earlier 



