of the So-called Storm-glass. 



103 



Lower 

 thermometer. 



Time. 



Upper 

 thermometer. 



Difference. 



Boil- 214 



ing 156 



103 



66 



62 



55 



h m 



9 50 a.m. 

 10 

 11 5 „ 



11 55 „ 



12 7 „ 

 12 45 „ 



Boil- 212 



ing 170 



126 



80 



72 



59 



2° 

 14 

 23 

 14 

 10 



4 



It is perfectly well known that in an experiment of this kind 

 the cooling takes place chiefly by means of currents set in mo- 

 tion in the external air, which generate a similar current in the 

 liquid, the warmer particles ascending by the axis of the tube, 

 and the colder ones descending by the sides ; so that while this 

 process goes on, the upper thermometer will mark a higher tem- 

 perature than the lower. In a saline solution several additional 

 phenomena are to be observed, many of which do not belong to 

 this inquiry ; but it may be noticed that even in salts which do 

 not form supersaturated solutions, such as sal-ammoniac, the 

 solution remains fluid some time after the lamp has been removed, 

 while crystallization endeavours to establish itself at the surface 

 where the air and the liquid meet. Although it is quite true 

 that the upper thermometer marks a considerably higher tempe- 

 rature than the lower, yet the coldest portion of the solution 

 before crystallization sets in is that very thin film which is in 

 actual contact with the air. I have seen in a saturated solution 

 of sal-ammoniac at the boiling-point, directly after the lamp has 

 been removed, a crystalline film form and disappear many times 

 before it could establish itself on the surface. A similar struggle 

 goes on in the case of supersaturated solutions, although in such 

 case the film may not be visible. It is dissolved, so to speak, in 

 the very act of forming ; but its existence is made quite evident 

 by the saccharine-like streams which descend from the sur- 

 face during the cooling, and are identical in character with those 

 which accompany actual crystallization mother parts of the tube. 

 There are many other points connected with the cooling and 

 crystallization of solutions of hydrated and anhydrous salts, which 

 I may perhaps be allowed to describe on another occasion. In 

 conclusion, I give a few details respecting a storm-glass bottle 

 so arranged as to note the upper and lower temperatures. 



A white glass cylindrical bottle, 5^ inches high exclusive of 

 the short neck, and 3 inches in diameter, of the capacity of 20 

 fluid-ounces, with an opening at the top and another at the side 

 near the bottom, was filled about five- sixths full with a storm- 

 glass mixture, and a thermometer, passed through a cork, was 

 placed in each tubulure ; a third thermometer was suspended on 



