74 Mr. J. J. Guest on the Strength of 



due to stresses or mechanical treatment, can be removed by 

 annealing, which is generally considered to render the material 

 isotropic. 



6. The Elastic Limit. — It would so far appear that the 

 yield-point is a definitely defined point with the material in 

 its normal condition, but a closer examination of the phe- 

 nomena involved brings to light a difficulty in its accurate 

 determination. In addition to the autographic curve in fig. 1, 

 a highly magnified edition of the same curve is shown, the 

 points on this curve being obtained by the use of an extenso- 

 meter ; the curve is referred to as the " extensoineter curve." 

 At some point D in the extensometer curve (fig. 1), before 

 the yield-point A, the stress-strain curve for the material 

 ceases to be a straight line, and at the same point a time-effect 

 manifests itself, the material yielding a little for some time 

 after the application of the load. This time-effect renders 

 the value of the strain just as the yield-point is reached a 

 somewhat vague quantity ; it, however, does not very materi- 

 ally affect the determination of the stress at the yield -point. 

 The point at which this deviation of the straight line and 

 time-effect appears is usually known as the elastic limit. I 

 shall refer to the effect of time upon the strain and to the 

 deviation of the stress-strain curve from linearity as the 

 elastic-limit effect. 



7. The Elastic-limit effect due to the existence of Local 

 Yield-points. — This phenomenon may be one sui generis, but, 

 partly as the departure from the straight line and the mani- 

 festation of the time-effect are simultaneous, I am inclined to 

 think that it is due to local variations in the material, which 

 partly by disturbing the uniform distribution of stress and 

 strain in their neighbourhood, and partly by their nature, 

 result in small volumes of the material reaching the yield- 

 point prematurely, and so cause the specimen as a whole to 

 exhibit a foreshadowing of the yield-point phenomena. On 

 this supposition, if a specimen be once strained, as a whole, 

 up to or beyond the yield-point, at the locations which give 

 rise to the elastic-limit effect, the strain would be carried well 

 beyond the yield-point ; and if the load were removed, on a 

 subsequent test they would contribute no time-effect until 

 their new yield-point was reached, which would occur simul- 

 taneously with the yield-point of the whole specimen ; hence 

 a specimen would display no elastic-limit effect, or at any rate 

 a very much reduced one, on a second testing. This agrees 

 with experiment, and illustrations of it will be pointed out 

 later. (See figs. 16, 17, & 18.) 



Furthermore, although the effect is removed from a par- 



