58 THE ROYAL SOCIETY OF CANADA 



It will be seen that we are in good agreement with the deter- 

 mination of the quartz dilatometer of Callendar and Harlow, but 

 disagree with the determination of Callendar and Moss. The last 

 three results of Callendar and Moss are exterpolated values from 

 Callendar and Moss formula which was deduced from observations 

 extending to -10°C. only. 



Our results do not show any very large change in the coefficient 

 although it falls off more rapidly than the exterpolation formula of 

 Callendar. 



