APPARATUS FOR MEASURING STRAIN AND APPLYING STRESS. 



279 



was reached. The return curve was less so, but as soon as torsion was applied in the 

 negative direction the linearity disappeared, and the strains, though irregular, became 

 greater and greater as the torque increased. The material finally gave way under a 

 torque of about 1100 inch pounds. The torque was now reversed, and the stress-strain 

 curve became approximately linear until the zero torque was reached, from which point 

 the curve began to bend over to the left, until a torque of 1175 inch pounds was reached. 

 In order to roughly test the behaviour of the specimen still further, the applied 

 torque was continued, but no strain measurements were taken. 



Toryjuue, lmjch> ^S 



DcslarGuTii 



sctocrc 



Each complete cycle produced a hardening effect on the bar, widening its limits of 

 endurance each time until a final limit of 1750 inch-pounds of torque was reached after 

 fourteen reversals of the stress. The bar was now cracked in several places along the 

 minute seams of impurities, and further experiment seemed useless. 



This experiment demonstrates that the limits of elasticity do not remain in their 

 original positions, and, further, it shows that stress carried beyond the elastic limit in 

 one direction reduces the other limit to zero. The conclusion derived from the theory 

 above, that the bar is twice as strong to resist torsion in the original direction as in the 

 other, is also not borne out by the experiment. 



A second bar of wrought iron was next examined in the same manner, only four 

 cycles being performed, of which the first two are shown in fig. 23 and the last two in 

 fig. 24. These curves exhibit the same general properties as the one described above. 

 It is evident from figs. 23 and 24 that, after the first reversal of stress, there is no perceiv- 

 able yield -point, all such critical points being absent. The commonly received idea 



VOL. XL. PART II. (NO. 14). 2 T 



