﻿654 Mr. N. Eumorfopoulos on the 



and Moss' value for the expansion of mercury leads at low 

 temperatures to inadmissible values for that of quartz. I have 

 also plotted on the same diagram the expansion of the quartz, 

 making use of ChappmV value for mercury (curve IT.), 

 and also of Thiessen, Scheel, & Sell's value from the 59 m bulb 

 (curve III.). For comparison curves IV. and V. are added ; 

 these are obtained from Scheel's values for the linear 

 expansion of two different specimens of quartz, one of which 

 indicated a maximum density at —46° C., and the other at 

 — 84° C. In interpreting the differences between the 

 various curves, it should be remembered that an uncertainty 

 of 1 X 10" ' in the coefficient of expansion of the bulb * leads to 

 an uncertainty of "016 degree at the boiling-point of 

 sulphur, and of approximately "1 degree at 1000° C. With- 

 out laying too much stress on the exact shape of the curve 

 between 0° and 100°, it appears likely that the coefficient is 

 still changing at 100° f, and, therefore, the same value 

 cannot be used for the fundamental interval as for tempera- 

 tures above 100° C. Hence it is insufficient to know with 

 accuracy merely the fundamental interval of the weight 

 thermometer. It may further be pointed out that judging 

 from Scheel's two curves measurements on one specimen 

 cannot, even for silica, be used with confidence for another. 



With regard to the second point mentioned above, viz., 

 the calculation of the cubical from the linear expansion, we 

 must examine rapidly the various points of agreement 

 mentioned in Scheel and Heuse's paper, to see how they 

 affect the point we are considering. 



The agreement mentioned in §2 (p. 413) between the 

 expansion of mercury by the direct and indirect methods 

 throws no light on this question : it merely indicates that 

 Chappuis' measurements were carefully made, and that the 

 two glasses behaved almost identically (they were both 

 ft verre dm-") ; the experiment is, in fact, merely the 

 comparison (through the hydrogen thermometer) of the 

 judications of a weight with an ordinary thermometer. 



The next paragraph refers to the expansion of water. In 

 comparing the weight- thermometer method with the absolute 

 method, it will be better to take the values given by the latter 

 as the standard ones, as we then get a fairer test of the proper 

 allowance for the expansion of the bulb : 



* The difference between SclieeJ's two values is at 100° about 1| 

 times this quantitv. 



t Compare Randall, Phys. Rev. vol. xxx. (1910), p. 216. 



