8 SCHMITZ, On a MetJiod of Ice Calorimetry. 



not used in the earlier experiments, but was found to be a 

 distinct improvement. 



The average temperature of the water in the calori- 

 meter was slightly above o^'C. When the tray of ice was 

 used the thermometer gave a reading of O'Oi'' to 0*05''. 

 When it was not used the temperature was slightly higher, 

 but probably not often above O'l". The fact that the 

 temperature is not zero makes the initial weight of ice 

 slightly too small (say by one part in one thousand in 

 extreme cases), and further diminishes it by a gradual 

 melting. The correction for initial deficiency was too 

 small to take account of. The correction for gradual 

 melting was also too small to take account of ; this is 

 shown by the following experiments, (i) A piece of 

 ice of mass 14 grammes was weighted with a bullet and 

 weighed on a sensitive balance while immersed in the 

 calorimeter, the ice-tray not being used ; the rate of 

 melting was 4 per cent, per hour. (2) A piece of ice 

 of mass 5^ grammes was similarly weighed, the ice-tray 

 being in position ; there was no appreciable melting in 

 one hour. 



The Drying Apparatus. 



In my earlier experiments I dried the ice-jacket by 

 the application of filter-paper cooled by means of melting 

 ice. As it was desirable to make the drying process as 

 mechanical as possible, I ultimately adopted the form of 

 apparatus shown diagrammatically in Fig. 2. 



A deep cylindrical metal vessel A is firmly supported 

 (supports omitted in the diagram) within an earthenware 

 jar containing ice and water. A second metal vessel B, 

 of the shape indicated in the figure, fits loosely in A, and 

 contains ice. The lower part of the vessel A contains 

 filter-paper suitably arranged for the purpose of drying 



