52 THE ROYAL SOCIETY OF CANADA 



300 °C. was directly measured, exterpolated values were given also for 

 temperatures below -10°C. The mean coefficient from 0°C. to 100°C. 

 as found by Callendar was -000182054. 



The object of this investigation was to directly determine the 

 expansion from 0°C. down to the freezing point of the mercury by 

 means of the method of the weight thermometer. 



A marked increase in the contraction of the mercury near its 

 freezing point was looked for in view of the theory that solid mole- 

 cules are mixed with the liquid. On this theory is based the explana- 

 tion of the anomaly of the point of minimum specific heat of water 

 at 40 °C, as found by Prof. Barnes. For mercury likewise it is assumed 

 that there are solid molecules present at all temperatures from the 

 freezing point where there is a maximum number to the boiling point, 

 for the specific heat of mercury decreases rapidly with temperature 

 between these limits. 



In the case of mercury, however, the solid is more dense than 

 the liquid, resulting in a contraction on freezing. The effect of the 

 solid molecules on the volume of the mercury would be to cause a 

 more rapid contraction in the neighbourhood of the freezing point 

 provided these particles were produced in greatly increased numbers 

 at the low temperature. The theory of solid molecules has also helped 

 to explain the maximum density point of water. We would therefore 

 expect an increase in the contraction of the mercury near its freezing 

 point. 



Apparatus and Method Used. 



Some special dilatometers of fused quartz were constructed for 

 us. These were approximately 100 c.c. capacity fitted with special small 

 overflow tubes of about 1 mm. bore. Two of these dilatometers 

 were filled with pure redistilled mercury. The final values were, 

 however, obtained with only one of these, as the second one filled 

 was accidentally broken before measurements could be made with it. 



A special thermostat was constructed after many trials of dif- 

 ferent forms. This worked so satisfactorily that we describe it here 

 in some detail. 



A diagram of the thermostat is shown in Fig. 1 . 



The containing bath V, was fitted with an auxiliary tube A, 

 fitted with a rotating worm N. This bath was packed in an outer 

 case W, with glass wool. Lead tubing K, was wound around the bath 

 and connected to the top with a reservoir containing liquid air L. 

 This reservoir was a Dewar bulb with an outlet at the bottom as shown. 

 The lower end of the lead tube was connected with a three way cock M, 



