116 Mr GRAHAM on the Influence of the Air in determining 



In all successful cases, crystallization commenced in the upper 

 part of the receiver around the bubble of air, but pervaded the 

 whole solution in a very few seconds. A light glass bead was 

 thrown up into a solution without disturbing it. 



It occurred to me, that, since the effect of air could not be 

 accounted for on mechanical principles, it might arise from a 

 certain 'chemical action upon the solution. Water always holds 

 in solution a certain portion of air, at the temperature of the at- 

 mosphere, which it parts with upon boiling. Cooled in a close 

 vessel after boiling, and then exposed to the atmosphere, it re- 

 absorbs its usual proportion of air with great avidity. Now, this 

 absorbed air appears to affect in a minute degree the power of 

 water to dissolve other bodies, at least a considerable part of it 

 is extricated upon the solution of salts. When a bubble of air is 

 thrown up into a solution of sulphate of soda, which has pre- 

 viously been boiled and deprived of all its air, a small quantity of 

 air will certainly be absorbed by the solution around the bubble. 

 A slight reduction in the solvent power of the menstruum will en- 

 sue at the spot where the air is dissolved. But the menstruum is 

 greatly overloaded with saline matter, and ready to deposit ; the 

 slightest diminution of its solvent power may therefore decide 

 the precipitation or crystallization of the unnatural excess of sa- 

 line matter. The absorption of air may in this way commence 

 and determine the precipitation of the excess of sulphate of soda 

 in solution. 



Here, too, we have an explanation of the fact just mentioned, 

 that solutions of sulphate of soda which have not been boiled, 

 are less affected by exposure to the air than well boiled solu- 

 tions ; for the former still retain the most of their air, and do 

 not absorb air so eagerly on exposure as solutions which have 

 been boiled. 



But the theory was most powerfully confirmed by an expe- 



