152 



THE ROYAL SOCIETY OF CANADA 



II., column 3) 2 , and found to be in fairly good agreement with Chappuis' 

 values. 



Harlow concludes his paper as follows: — 



"The large discrepancy which exists between the results of the 

 absolute and weight thermometer methods, at the lower tempera- 

 tures, is difficult to explain. It seems that there must be some hitherto 

 undiscovered systematic error in one of the methods, and in view of 

 the importance of a knowledge of the correct thermal expansion of 

 mercury in thermometry of precision, the subject undoubtedly calls 

 for further investigation." 



In the discussion which followed, Mr. Sears called attention to 

 the values of the coefficient of mercury obtained by certain observers 

 (See Table II.) He thought it on the whole fairer, as Calendar's 



Table II. 



XlO-s 



and Eagle's 1 experiments were not made on the actual bulbs used by 

 Harlow and also were very near Kaye's mean values, to use the mean 

 of the two. Accordingly, he had calculated a quartic representing the 

 expansion of mercury on this basis (See Table II., column 6.) 



Mr. Sear^ thought that "Prof. Callendar would presumably 

 agree that, pending some further investigation, the Callendar-Moss 

 values at the lower temperatures, must be regarded as being affected 

 by some unexplained source of experimental error." 



-Table II is taken from the report of the discussion of Harlow's paper and is 

 due to Mr. Sears. 



1 Mr. A. Eagle has recently measured the expansion of fused silica over the 

 range 0° to 120°C. His results which have not yet come to hand, are in close agree- 

 ment with Callendar. (See Proc. Phys. Soc. London 26, p. 92.) 



