Mr. Graham on the Influence of the Air in Crystallization, 215 

 above four quantities respectively by r, s, S, R, we shall always 



have 1W±£±»L= + n , H^P+GH-H.) = + y 

 s o 



n and N having the signification of article 5 : we pass, however, 

 over the demonstration of this theorem, which it is not diffi- 

 cult to find, as it is not further necessary for our purpose. 



XXXVI. On the Influence of the Air in determining the Cry- 

 stallization of Saline Solutions. By Thomas Graham, Esq. 

 AM. F.R.S. E* 



T^HE phenomenon referred to has long been known, and 

 ■*' popularly exhibited in the case of Glauber's salt, without 

 any adequate explanation. A phial or flask is filled with a 

 boiling saturated solution of sulphate of soda or Glauber's 

 salt, and its mouth immediately stopped by a cork, or a piece 

 of bladder is tied tightly over it, while still hot. The solution, 

 thus protected from the atmosphere, generally cools without 

 crystallizing, although it contains a great excess of salt, and 

 continues entirely liquid for hours and even days. But upon 

 withdrawing the stopper, or puncturing the bladder, and ad- 

 mitting air to the solution, it is immediately resolved into a 

 spongy crystalline mass, with the evolution of much heat. The 

 crystallization was attributed to the pressure of the atmosphere 

 suddenly admitted, till it was shown that the same phenomenon 

 occurred, when air was admitted to a solution already subject 

 to the atmospheric pressure. Recourse was likewise had to 

 the supposed agency of solid particles floating in the air, and 

 brought by means of it into contact with the solution ; or it 

 was supposed that the contact of gaseous molecules themselves 

 might determine crystallization, as well as solid particles. But 

 although the phenomenon has been the subject of much spe- 

 culation among chemists, it is generally allowed that no satis- 

 factory explanation of it has yet been proposed. 



In experimenting upon this subject, it was found that hot 

 concentrated solutions, in phials or other receivers, might be 

 inverted over mercury in the pneumatic trough, and still re- 

 main liquid on cooling ; and thus the causes which determine 

 crystallization were more readily examined. For this pur- 

 pose, it was absolutely necessary that the mercury in the trough 

 should be previously heated to 110° or 120°; for otherwise 

 that part of the solution in contact with the mercury cooled 

 so rapidly, as to determine crystallization in the lower part of 



* From the Transactions of the Royal Society of Edinburgh ; but re- 

 vised by the Author for the Phil. Mag. and Annals. 



the 



