KIISIIKs \MI ITS TKMI'KKATURK GOKFFICIKNT. 



123 



logarithmic decrement gives the togMltluauo decrement due to the viscosity of the 

 quartz fibre : 



It thus appears that the internal viscosity of a quartz fibre varies in an irregular 

 manner with increase of temperature. It must, however, be noticed that the internal 

 viscosity plays but a small part in the damping of the torsional oscillations, and that 

 i> small error in the observed logarithmic decrement would make a large difference in 

 the value obtained for the logarithmic decrement due to the viscosity of the fibre. 

 These numbers can only be taken as showing that the viscosity of a quartz fibre is 

 very small, and that it does not increase to any great extent when the temperature 

 of the fibre is raised from 15 C. to 100 C. 



PART III. THE VARIATION OF THE MODULUS OP RIGIDITY BETWEKX 20 C. 



AND 1000 C. 



In these experiments the apparatus represented in fig. 5 was used. The fibre A 

 was much thicker than those used in the previous experiments, and was drawn down 

 by hand from the two quartz rods B and C. These are about 2 millims. in diameter 

 and 8 centims. in length. The upper end of B was silvered, coppered, and then 

 soldered into a hole along the axis of the brass rod E, from which the vibrating 

 system is suspended. E can be clamped at the desired level in the torsion head by 

 means of the milled-headed screw H. The lower end of C was similarly soldered into 

 the brass rod D, which, with the cylinder F, forms the vibrator. On the rod D a 

 small plane mirror M is fixed, and this serves, as in the former experiments, for 

 observing the torsional oscillations. The fibre, quartz rods, and part of the brass 

 rods E and D are surrounded by a platinum tube 7 millims. in diameter and 

 35 centims. in length, which was kindly lent to me for this experiment by Dr. H. A. 

 Winsox, of Trinity College. The centre portion of this tube is heated by means of 

 an alternating current from a transformer, and its temperature is obtained by means 

 of a thermocouple of wires of platinum and an alloy of platinum with 10 per c*ent. 

 of rhodium, welded on to opposite sides of the platinum tube at its middle, as shown 

 in fig. 5. These wires were about '04 millim. in diameter, and their other ends, 



