the Substances usually present in Water. 41 



lime and chloride of sodium ; and thus also a solution of bi- car- 

 bonate of magnesia and sulphate of lime yields carbonate of lime and 

 sulphate of magnesia. These changes seem to be determined by the 

 relative solubilities of these salts in a certain proportion of the 

 liquid. In like manner, alcohol may often precipitate from a mix- 

 ture of acids and bases, not the existing compounds, but such new 

 compounds as are insoluble in it. Temperature too has a great effect ; 

 for instance, sulphuric and muriatic acids, soda and magnesia, at the 

 freezing point of water, give rise to muriate of magnesia and sulphate 

 of soda, which crystallizes ; but at the boiling point, to chloride of 

 sodium, which crystallizes, and a solution of sulphate of magnesia. 

 It is evident then that the complicated methods of separating the 

 different salts in waters, copied by many chemical system writers 

 downwards from the time of their contriver Mr. Kirwan, are, to say 

 the least, useless ; even if the practicability and correctness of many 

 of them were not more than questionable. 



In proceeding with the analysis after the substances contained in 

 the water have been discovered by testing, the gaseous substances 

 may be first estimated. For the atmospheric air, if it be thought 

 necessary to measure it, take a small tubulated retort, cork the end 

 of the neck, fill it completely with a known quantity of the water, and 

 then stop the tubuture with a cork pierced with a small tube to con- 

 duct away the water displaced by the evolved gas. This will be 

 collected in the neck on heating the water to the boiling point- 

 Transfer it into a measure tube. Absorb any carbonic acid by lime 

 or potass, then remove the oxygen by phosphorus. Thus the bulk 

 of the oxygen and nitrogen emitted by a given weight of water, 

 may be readily known. 



The free carbonic acid is usually directed to be ascertained by 

 expelling it from the water by boiling, and collecting it over mercury. 

 This, however, gives no good or uniform result in many instances, 

 owing to the long boiling required, and the quantity of water vapo- 

 rised, before the whole of the excess of carbonic acid is driven off 

 from the bi-carbonates present in the solution. A better process 

 seems to be the following. Pour 10,000 grains of the water into a 

 stoppered flask, add to it an excess of lime water, allow the precipi- 

 tate to subside, pour off by a syphon the liquid above, wash. The 



G 



