[wheeler] CUBICAL EXPANSION OF VITREOUS QUARTZ 151 



and have concluded that Callendar and Moss had corrected a long- 

 standing error in the absolute expansion of mercury between 0° and 

 100°C. of nearly 0-3%. However, the reports of more recent in- 

 vestigations of Callendar 1 and Harlow, 2 communicated to the London 

 Physical Society in November, 1913, throw considerable doubt on 

 the conclusion that the cubical expansion of a fused quartz dilatom- 

 eter bulb between 0° and 100°C. is only twice the axial, and that it 

 vanishes between 0° and 15 °C. 



The more recent conclusions of Callendar may in part be summed 

 up as follows: — 



(1) That in attempting to deduce the expansion of mercury 

 by the weight thermometer method with silica bulbs, it was necessary 

 to determine the expansion of specimens of silica from the same source 

 over the entire range 0° to 300°C, 



(2) That, although differences of 5 or 10 per cent in the ex- 

 pansion in different directions appeared persistent, — and were not 

 removed by heating to 1,000°C. or cooling to the temperature of 

 liquid air, — the most probable result for the cubical coefficient would 

 be obtained by assuming it to be three times the linear; 



(3) That owing to the smallness of the expansion of silica, 

 and its comparative freedom from hysteresis, the possible uncertainty 

 with silica bulbs, in the expansion of mercury, was probably less 

 than 0-1 per cent in spite of imperfect annealing. 



Over the range 20° to 300°C, Callendar's recent results for the 

 linear expansion of fused silica are represented by the equation: — 



86-5 1 



•78 [ 10- 6 



t+175 J 



where a t is the mean coefficient between 0° and t°C. 



Harlow's recent paper 3 contains a more complete set of observa- 

 tions of the relative coefficient of expansion of mercury in silica, 

 than his earlier paper. Some readings taken with a spherical bulb 

 failed to support the earlier conclusion that the cylindrical bulbs 

 were not isotropic. 



Using Callendar's recent value for coefficient of expansion of 

 silica determined for specimens similar to those from which the bulbs 

 were made and assuming the cubical expansion to be three times the 

 linear, the absolute expansion of mercury was determined (See table 



1 Nature 92, p. 467, 1913. 



2 Proc. Plus. Soc. London 26, pp. 85-96, 1914; Nature 92, p. 467, 1913. 



3 Loc. cit. 



