﻿of Ductile Materials under Combined Stress. 



539 



two pins branching from the central portion which fit into 

 bearings, these steel pins being carefully ground in place. 

 Drawings of the support at the other end are also given in 

 tig. 2. The bar rests on two rollers of large diameter which 

 are in turn supported on their side pins by the metal strips. 



Fie-. 2. 



rti 



rrm : (Tin 





■4- 



— i — 



A 



- r -- 



: | : ;;i 



-_--- 



i > i 



; 



ru 



fL7\] 



*=± 



This arrangement reduces the resistance to twisting at this 

 end to a minimum. The torque was applied through a light 

 wooden pulley having metal plates bolted to it, these having 

 j-inch square holes cut in them in line. Suitable wire ropes 

 were attached to this pulley, and each passing over another 

 pulley, weights could then be attached and considerable motion 

 was possible. The bending load was applied directly by 

 placing weights on a platform attached to a knife-edge resting 

 on the bar. Deflexions were measured by means of a scaled 

 strip which rested on the specimen and fitted easily in a slide 

 with a vernier cut on the side. This arrangement was quite 

 satisfactory, although it does not appear very sensitive, but it 

 should be remembered that the deflexion is large even before 

 the yield-point is reached, and it was possible to read 

 accurately to 0'005 of an inch. The twist was measured at 

 four points. At three of these, mirrors (see fig. 2) were 

 attached to the bar. Telescopes were fixed at approximately 

 a metre from the mirrors, and there were vertical scales; the 

 arrangement being similar to that in common use in connexion 

 with galvanometers. Outside the support, on the pulley side, 

 a pointer was fixed to the bar and moved over a fixed scale 

 as the rod twisted. This served to give fairly accurate 



