[wheeler] CUBICAL EXPANSION OF VITREOUS QUARTZ 155 



The mean results, thus obtained, nearly all fit an even curve 

 very well as shown by Fig. IV. The values of l t /lo, obtained from the 

 curve of Fig. IV, plotted on a much larger scale, have been tabulated 

 for every ten degrees from —250° to 1,100°C. (See Table VI.) 



To find the average coefficient of expansion over any required 

 temperature range, it is only necessary to subtract the value of l t /l 

 for the lower temperature from that at the higher temperature and 

 then divide by the temperature difference. Thus to find the coefficient 

 of expansion between -40° and 100°C, we would subtract 0-9999884 

 from 1 • 0000488 and divide by 140, obtaining • 431 X 10- 6 . By inter- 

 polation, values for temperatures which are not exact multiples often 

 may be found. 



The absolute value of the thermal expansion of fused quartz is 

 so small that often its knowledge is not required to a high degree of 

 accuracy. In spite of the differences which appear to exist between 

 different specimens of fused quartz, it is thought that, over a range 

 of 100°C. or more (below 1,000°C), Table VI. can be relied upon 

 to give the linear or cubical expansion coefficient of an average 

 specimen of fused quartz, at least to an accuracy of within 4 or 5 per 

 cent, which corresponds to an accuracy in absolute expansion of 

 about 0-3 per cent for verre dur. 



In the light of the above evidence, it would seem that, where 

 a much greater accuracy is required, it is necessary to determine 

 the expansion of the actual silica dilatometer bulb or apparatus 

 used. Indeed where very great accuracy is required, it would seem 

 desirable to determine the linear expansion in more than one direction 

 when possible. 



X 



-Ot-lo)/, +i 



