1921] 



on Elasticity 



the lower end of the vertical rod and clamped again to a fixed sup- 

 port. To avoid conduction the clamps are mounted on glass rods. 

 The period of the torsional vibrations depends chiefly on the stiffness 

 of the test-piece, but in part also on that of the suspension wire, and 

 this has to be allowed for in calculating the true period due to the 

 reactions of the test-piece itself. (Fig. 4.) 



The periods of the test-piece and suspension wire combined 

 ranged from five to ten seconds, and the oscillations when once 

 started lasted long enough without maintenance to allow of their 

 being accurately timed. As 

 a matter of convenience, 

 however, in recording each 

 oscillation on to the chrono- 

 graph, maintenance was 

 applied in the form of an 

 impulsive couple acting for 

 an instant while the oscil- 

 lating system was passing 

 through the zero position. 

 This was effected by the 

 falling of two small weights, 

 which were released at the 

 proper time by an electro- 

 magnet actuated by a cur- 

 rent which flowed only 

 when a fine wire point 

 carried by the oscillating 

 parts touched a globule of 

 mercury, the same current 

 being used to work the pen 

 on the chronograph. 



The variations of the 

 rigidity of most of the 

 metals used in the previous 

 series, with the addition of 

 tungsten, were examined in 

 this way, and the results, 



as will be shown presently, did not differ much from those found for 

 the variations of Young's Modulus. ' 



What was noticeable, however, was the great difference which 

 the change of temperature produced in the amplitude of the main- 

 tained oscillation and in the rate at which the unmaintained oscilla- 

 tion died out. In both cases the difference is caused by the variation 

 of the viscosity, or internal friction of the material of the test-piece. 



To get some measure of this internal friction I arranged a further 

 series of experiments which, though still proceeding, has yielded some 

 results. 



Tn Chronograph 

 and Relay . 



-^Battery 



Fig. 4. 



