and Chemical Change. 517 



the conditions. In general a considerable increase (4 to 5 

 times) was observed : but on other occasions no change 

 occurred or a slight decrease was noted. The determining 

 condition appeared to be the amount of dust in the vessel ; 



it was impossible to get the air of a vessel of the size and 

 construction used quite dust-free, but when all precautions 

 against dust had been taken the change on heating was very 

 small. 



If the bunsen was left in action the leak gradually re- 

 turned to its normal value : when that was reached no further 

 change occurred whether the heating were continued or the 

 flame removed and the vessel allowed to cool down. 



The effect of chemical action was investigated by placing 

 one of the reagents in a dish on the tray and pouring on it, 

 through a tube, a solution of the other reagent. Gases which 

 are evolved in chemical action are known to be strongly 

 ionized, and hence in these experiments no actions in which 

 gases are evolved were tried. 



The following actions which are attended by no considerable 

 heat-change were found to have no perceptible influence on 

 the leak. (Class A) :— 



Zinc on copper sulphate, ferrous sulphate oxidizing in air. 

 ferrous sulphate oxidized by permanganate, barium chloride 

 and sodium sulphate, ammonia and copper sulphate, sodium 

 chloride dissolving in water. 



The following actions which are attended by considerable 

 heat-changes gave effects exactly similar to, but much less 

 intense than, tho^e observed when the tray was heated by a 

 bunsen-burner : when the flame increased the leak, these 

 action- increased it ; when the flame decreased the leak, so 

 did the actions. (Class B ; : — 



The solution in water of calcium chloride, phosphorus 

 pentoxide, sulphuric acid, ammonium sulphocvanide, sodium 

 thiosulphate ; strong potash and sulphuric acid, strong soda 

 and hydrochloric acid. Action of water on lime, and plaster 

 of Paris. 



§ 5. A window was used consisting of a large sheet of 

 thin paper rendered conducting by graphite, or a dilute 

 solution of calcium chloride, the tray being placed over the 

 window. 



The most extraordinary and complicated heat effects were 

 observed in this case. Jn general an increase occurred when 

 heat was applied, and a decrease — passing through zero and 

 amounting to a leak against the field — during cooling. Bui 

 these changes are not of importance for our present purpose, 

 since no experiments were made on chemical actions with 



