414 Mr. F. A. Waterman on an Improved Calorimeter 



varied from *0908 to '0969, the probable error of the mean 

 of the whole being much higher than is usual in the case of a 

 similar series of results obtained by Regnault's method. As 

 the above method appeared to be a desirable one for some 

 work in hand, I have endeavoured to perfect a calorimeter 

 by means of which it might be applied with greater accuracy. 

 The form of apparatus which I have devised has proved 

 very satisfactory after repeated trials, as is shown by the 

 results here submitted. It consists essentially of the following 

 parts (see figure) : — 



A glass jar which supports, by a wooden cover, the bulb R 

 of a glass air-thermometer, the manometric tube M of which, 

 in the form of a capillary U-tube, is supported in a vertical 

 position. The manometric tube is supplied with a stopcock 

 S. The glass jar is kept nearly full of water for the purpose 

 of maintaining the thermometer-bulb constantly at room- 

 temperature and preventing any sudden variation of tempe- 

 rature due to air-currents. For the purpose of increasing 

 the sensitiveness of the thermometer, the liquid selected was 



