188 



Mr. H. E. Schmitz. On the Determination of [Feb. 2, 



of mixtures give appreciably too high a result if the liquid air is 

 allowed to enter the calorimeter. But in this method the corresponding 

 error will probably be of trifling amount, as the bulk of the liquid air 

 rises to the surface of the water immediately on immersion of the 

 object in the calorimeter ; it sometimes forms independent ice capsules 

 floating on the water. There is no necessity from this point of view 

 to attach much importance to the time of transference from the bath 

 of liquid air to the ice calorimeter. 



(ii) A minute quantity of liquid air is always vaporised in contact 

 with the object, or in its immediate proximity, and is imprisoned within 

 the ice coating. This causes the chief difficulty in the method as prac- 

 tised by the author. The difficulty arises, not from any calorimetrical 

 effect, but from the fact that the vaporised air slowly and gradually 

 escapes in small bubbles through one or more holes near the top of the 

 ice coating, its place being taken by water drawn in some way into the 

 interior of the ice coating and therefore necessarily weighed with 

 the ice. 



(iii) For the reason mentioned in the last paragraph, the shape as 

 well as the size of the immersed object is of importance. The object 

 should be large and should have a small surface. Its base should be 

 well rounded, otherwise a bubble of air of considerable size may form 

 .at the base in the earliest stages of the formation of the ice coating ; 

 this may produce a large error in the manner above described. 



(iv) On account again of the gradual suction of water into the 

 ice coating, the result depends on the time of immersion in the 

 calorimeter. 



The statements of the three preceding paragraphs are illustrated 

 by Tables V and VI. In Table V each of the numbers given is the 

 mean of three, four, five or six. In Table VI the results of the 



Table V. Variable Immersion. 



* This number is not quite comparable with the others in the table. In the 

 experiments of which it gives the mean result, the calorimeter was in a more 

 efficient state than in the other long-period determinations. 



