268 DR E. G. COKER ON 



specimen M under test are secured in grips N N upon the projecting ends of the 

 cones B. 



In order to obtain a pure torque and a pure bending moment, both acting at the 

 same time, each casting is supported by a ring (fig. 9) encircling the spindles B, and 

 furnished with friction rollers P, running in grooves in the spindles B. The rings have 

 bearings Q, turning in stirrups R, these latter being hung from a horizontal bar S, by- 

 adjustable vertical hangers T. With this arrangement we get a pure torque of a known 

 amount throughout the specimen. 



Bending Moment. — Into the outer ends of the nuts C are screwed projecting arms 

 V, of known length and carrying weights at their ends. These put a bending couple 

 upon the specimen without shear, the arm of the couple being the distance of the 

 weight from the hanger T. With this arrangement simple twist and simple bending, or 

 any combination thereof, can be applied to a specimen with ease. The specimen is free 

 to take up its own position of equilibrium, since it is imperfectly constrained — (the 

 specimen can be easily rocked about even when fully loaded, but always comes back to 

 the first position after a few oscillations) — and the conditions of stress are accurately 

 known. 



Corrections : Twist. — The results of tests on the friction of the roller bearing show 

 that the friction is so small a quantity as not to introduce any sensible error. 



Bending. — The friction error is that due to the stirrups embracing the bearings Q, 

 which latter were made large purposely. No experiments were made in which the 

 bending moment varied during the experiment ; consequently it was sufficient in each 

 case to calculate the error due to friction for the particular load applied, and to make 

 the small corrections necessary. This has been clone in all cases. 



3. Apparatus for applying Torque and Tension. 



The apparatus for applying torque and tension is shown in general elevation by fig. 

 11, and in plan by fig. 12. A detailed section of some of the parts with the measuring 

 instrument fixed to the specimen is shown by fig. 13. 



The specimen A was screwed into a turned piece B, having a slotted hole above, 

 and tapped to receive a screw C, centred in a corresponding depression in a piece of tool 

 steel D resting upon a plate E, which latter was carried by four bolts F, depending from 

 a cast-iron beam G mounted upon two pillars H. Below, the specimen was screwed into 

 a turned piece J", carrying a sleeve and pulley L, the lower end of the piece being fitted 

 to receive a nut M and hanger N for weights. The torque was applied by weights 

 attached to fine bands made of clock spring steel, which latter were attached to the 

 pulley at convenient points, and passed over guide-pulleys mounted on ball-bearings. 

 The applied torque was balanced a,bove by a double-ended lever P, keyed to the piece 

 B, so that its axis passed through the point of suspension, and the lever being furnished 

 with screws at its outer ends, these latter pressed equally against the pillars H. The 



