410 DR. F. HORTON ON THE MODULUS OF TORSION AL RIGIDITY OF 



tubing, the other end of which is closed, and by gently squeezing this any desired 

 amplitude of oscillation can be communicated to the vibrator. 



The clamping rod E had been used in the experiments with metal wires, and, in 

 order to prevent leakage of heat along it, a piece of marble had been inserted between 

 the clamp and the rest of the rod. The rod can be fixed at any desired level in the 

 brass head D, which works smoothly in a collar and is capable of rotation about a 

 vertical axis by means of the screw C, thus giving a fine adjustment to the azimuth 

 in which the vibrator hangs. 



In determining the absolute value of the rigidity modulus, a rapid stream of water 

 was allowed to circulate through the jacket. It would then take about an hour for 

 the temperature, as indicated on the thermometer, to become constant. 



Professor POYNTING'S modification of the method of coincidences, which was used in 

 timing the torsional oscillations, has been described in detail in the paper before 

 referred to. A description of the apparatus used to obtain the second flashes is also 

 contained in the same paper. It will be sufficient to mention here that the accuracy 

 in timing obtained in the present experiments was about 1 in 100,000. 



The thermometer T was graduated to '1 C. and had been standardised for use 

 horizontally at the National Physical Laboratory. The part of the stem protruding 

 from the jacket was surrounded by a glass tube, through which a stream of water at 

 a known temperature was kept flowing, and thus the correction for the exposed part 

 of the mercury thread was obtained. It was only necessary to apply this correction 

 in the experiments at higher temperatures. 



The Order of Experiment. 



A fibre of the required size having been mounted in the manner already described, 

 its length was first measured by means of a cathetometer. For this purpose the 

 rod E was left in the position to which it had been lowered in order to clamp the 

 brass pin to which the fibre was affixed. The vibrator and fibre were surrounded by 

 a glass tube to keep off air currents, and the ends of the fibre were viewed in turn 

 by the cathetometer telescope. Owing to the very small coefficient of expansion of 

 quartz, it was unnecessary to take precautions to maintain the fibre at an accurately 

 known temperature whilst its length was being measured. The measurement of the 

 length was repeated before the fibre was taken down, but it was always found to be 

 constant, showing that the fibre had not drawn away from the cement, or other 

 material, used to fix it in position. 



After the length of the fibre had been measured, the clamping rod was carefully 

 raised until the vibrator was at the proper height inside the jacket. This was an 

 operation requiring much care and patience, for any irregular motion set the vibrator 

 in violent agitation, and with the finer fibres always ended in their breaking. 

 Vibrations were somewhat avoided by allowing the vibrator to rest against the side 



