Bending and Twisting Moments simultaneously. 499 



adjusted to any angular position. This arrangement measures 

 the twist on the whole specimen, and includes any motion 

 due to back-lash in the keys and key ways during a test, it 

 is therefore only suitable for measurements with long wires 

 having substantial ends, and for plastic strains in which the 

 end effects are negligible in comparison. 



For observations within the elastic limit the author prefers 

 to use an instrument* which is secured to the specimen and 

 is self-contained. 



This instrument was originally designed to measure the 

 angle of twist within the elastic limit, and with some recent 

 alterations it can be adjusted in a few seconds for measuring 

 the angular change due to bending. The calibration of 

 the readings is effected on the specimen and serves for both 

 bending and twisting. Fig. 9 (PI. XI.) shows the apparatus 

 in part longitudinal section. 



It consists of a graduated circle A mounted on the specimen 

 B by three screws C in the chuck-plate D. A sleeve E pro- 

 vided with three screws grips the specimen at a fixed distance 

 away from the first set. 



The spacing of these two main pieces on the specimen is 

 effected by a clamp, not shown in the figure, which grips the 

 double cones F, G, and maintains them at the correct distance 

 apart, until the set screws are adjusted. 



The clamp is afterwards removed, leaving the plane of the 

 graduated circle perpendicular to the axis of the specimen 

 and the sleeve correctly set and ready to receive the reading- 

 microscope H. 



The vernier plate carries a sliding tube I, on which a wire 

 J is mounted, and the movement of this latter due to bending 

 or twist is measured by a scale in the eye-piece K, the 

 divisions of which are calibrated by reference to the graduated 

 circle. It is found convenient to have the microscope-tube 

 pivoted about an axis perpendicular to its central line at L, 

 so that any slight difference due to imperfect centering can 

 be adjusted by the screw M to make the calibration value 

 agree for a series of specimens. 



The observation wire may be set at any convenient position 

 for calibration, but for observations of the angle of twist 

 when the specimen is also subjected to a uniform bending 

 moment the wire should be in the central plane perpen- 

 dicular to the specimen. For if the bending is in the plane 

 containing the axis of the specimen and the observation wire, 

 it has the effect of causing new parts of the w T ire to come 



* "On Instruments for Measuring small Torsional Strains," Phil. Mag. 

 December 1899. 



2 M 2 



