52 MR, LEONARD BAIRSTOW ON THE ELASTIC LIMITS OF 



clearly indicated by both the Swedish iron and Bessemer steel. The slope of the line 

 CB is greatest for the axle steel, but otherwise the general features of the curves for 

 all three materials are the same. 



The remaining three curves have been drawn through WOHLER'S points of obser- 

 vation. The wrought iron differs from the Swedish iron only in the magnitude of the 

 ordinate. Curve 5 is almost identically the same as Curve 2 for axle steel, but 

 WOHLER'S points are not sufficiently numerous to complete the curve with certainty, 

 and for these observations GERBER'S parabola can be made to fit with equal accuracy. 



The last of the curves again completely represents WOHLER'S results, and con- 

 sidering the three diagrams together the agreement is remarkably close. 



The truth of BAUSCHINGER'S theories would now appear to be established, as the 

 elastic ranges have been shown to agree with WOHLER'S safe ranges. The experi- 

 mental results, however, show very different features to those obtained by 

 BAUSCHINQEE, and this is probably due to the more complete observation of the 

 extensions and, in part, to the different definition of elasticity adopted. 



The Condition of Specimens after Recovery from Fatigue. A suggestion made by 

 BAUSCHINGER as to a rapid means of determining the " natural" elastic limits seemed 

 worth investigation. He suggested that a specimen be made inelastic by overloading, 

 and then by increasing the range from a small value the maximum range of elasticity 

 be found as recovery proceeded. The only experiments given involved less than one 

 dozen reversals and little time was allowed for recovery. It is not surprising, there- 

 fore, that under these conditions the elastic ranges had no resemblance to the safe 

 ranges found later. 



MUIR lias since shown that very complete recovery of elasticity can be produced by 

 the immersion of a specimen in boiling water for a few minutes. Taking advantage 

 of this, a series of tests was made on some of the specimens already described. 

 Recovery reduces the permanent set at a given load, as will be seen from one of the 

 Bessemer steel specimens which was given a rest of 45 hours at the end of an 

 experiment ; almost exactly one-third of the set disappeared in the interval. A 

 specimen after fatigue therefore tends to return to its primitive condition, and it is 

 necessary to have experimental evidence to show the extent of the recovery. 



The detailed observations are given in Table III. At the stresses indicated by the 

 first column, extensometer readings were taken and the extensions are tabulated in 

 the succeeding columns. The main heading relates to the recovery produced 

 immediately before the experiment, whilst the letters T and C signify tension and 

 compression respectively. The elastic limit for each experiment is indicated by heavy- 

 type figures in the column of extensions.* The general features are most easily seen 

 from the plotted curves, as the errors of observation are then more easily allowed for. 

 These are never great in absolute magnitude, but they may make individual 

 differences appear unduly important. 



Axle Steel. The original condition of this material showed no want of elasticity in 



