QUARTZ FIBRES AND ITS TEMPERATURE COEFFICIENT. 431 



that of a metal wire is crystalline. It is known that the effect of heat is to cause 

 the crystalline grains of which a metal is composed to grow, and this alteration of 

 the structure would mean an alteration in the rigidity of the material. The rate of 

 growth of the crystals at 1 00 C. would l)e very small, and it has been observed only 

 in the case of the softer metals, but it is probable that the present method can detect 

 changes which the microscope fails to show. If this view is correct, it would follow 

 that the larger crystalline structure has the greater rigidity, and this is supported by 

 some experiments on the effect of heat on the elasticity of steel recently performed in 

 the Engineering Laboratory of this University by Mr. F. ROUKRA, who found that 

 YOUNG'S modulus for steel was increased by cooking the specimen to high tempera- 

 tures a process which increases the size of the crystalline grains. On the other 

 hand, the observed increase of rigidity with time may be merely the result of a slow 

 annealing and due to the easing of the crystalline structure from contraction stresses. 

 In the present case of quartz fibres such small " time effect " as exists, is, I think, 

 undoubtedly due to a gradual annealing, and in the case of metal wires it is probably 

 due in part to each of the causes mentioned above. 



It is a curious fact that, although a quartz fibre is made quite brittle and useless 

 for most practical purposes by being heated to about 500 C., yet its rigidity at the 

 ordinary temperature of the laboratory is increased so long as it is not heated 

 above 760 C. 



Since the rigidity of quartz fibres is much more constant than the rigidity of any 

 metal wires, and as also the period and the logarithmic decrement of the torsional 

 vibrations are independent of the amplitude of oscillation so long as this is small, it 

 appears that quartz fibres are by far the best material to use for suspensions in all 

 forms of torsion apparatus where accuracy is required. 



In conclusion, I wish to express my sincere thanks to Professors J. H. PoYNTlNO 

 and J. J. THOMSON for their kind encouragement and valuable advice throughout the 

 course of these experiments. 



