288 
During the evaporation, a portion of the chloride of 
magnesium was decomposed, and water being poured on, the 
amount of magnesia left undissolved was determined. 
To the solution, oxalate of ammonia was added in 
excess, by which the lime was obtained ; and, this being 
separated, the remainder of the magnesia was insulated in 
the usual manner, by boiling with carbonate of potash, eva- 
porating to dryness, &c. 
To check the determination of the magnesia, the follow- 
ing method was taken. A known weight of the water was 
boiled with excess of lime, by which the magnesium and 
manganese were separated and replaced by calcium. The 
amount of this latter metal (the free lime being, of course, 
first removed by carbonic acid, boiling, &c.) was then deter- 
mined by oxalate of ammonia, and subtracting from it the 
calcium of the water as previously determined, the remainder 
was the calcium which replaced the magnesium and man- 
ganese. This remainder (the manganese being present in 
very minute proportion) was found to be almost exactly what 
it should be, supposing it to replace nothing but magnesium, 
or it was to the magnesium directly determined in the ratio 
of the atomic weights of the two metals. 
The solution, deprived of magnesia, lime, and manga- 
nese, was now evaporated to dryness, and ignited so as to 
expel the ammoniacal salts, and the residue gave the mixed 
chlorides of potassium and sodium, with a trace of sulphate 
of potash. The quantity of chloride of potassium in this 
mixture was determined in the usual way, by the bichloride 
of platinum, and the difference gave the chloride of sodium. 
The numbers thus obtained, some spattering having occurred 
during the expulsion of the ammoniacal salts, were looked 
upon as only relatively true, and were corrected in the fol- 
lowing manner. 
The chlorine and bromine having being determined, as 
shall be presently described, as also the different metals, 
