Dr Daubeny on separating Lime from Magnesia. 
Acid, 
- 
00 
/ 1 
Base, 
- 
16 
Water-, 
- 
52 
100 
Consequently, in the former instance, 16 parts of magnesia were 
contained in 800,000 parts of water, or 1 of the earth in 50,000 
of the fluid ; whilst in the latter, the respective proportions were 
as 1 to 24,000. 
The same insensibility to the presence of magnesia, I find to 
be possessed by the phosphate of ammonia, which Dr Murray 
has proposed to substitute for phosphate of soda * ; but I am 
not aware that this test, which is somewhat less easily procured, 
possesses any positive advantages over the latter, except where 
we wish to analyse the residual liquor after the separation of the 
magnesia, as in the examination of a mineral water. 
In such cases, it is better no doubt to employ a salt having a 
base of ammonia, which, in all its combinations, admits of being 
readily disengaged, than to make use of a fixed salt, where the 
alkali might by its presence occasion some perplexity in our sub- 
sequent operations. 
In order to ascertain the comparative accuracy of these two 
tests, I made use of given quantities of sulphate of magnesia, a 
salt, the composition of which, in its crystallised state, appears 
to be pretty accurately determined. 
From 50 gr. of this substance dissolved in water, I obtained, 
by means of phosphate of soda, a triple salt, which, when dried 
under an exhausted receiver, in connection with recently fused 
muriate of lime, amounted to 45.8 gr., and was reduced by a red 
heat to 20.3 ; whilst, from the same quantity of sulphate of mag- 
nesia, treated with phosphate of ammonia, I obtained 44.5 gr. 
of triple phosphate, which, after exposure to the same tempera- 
ture with the former precipitate, weighed 19.7.. 
But as it is diflicult, by the method employed above, to 
deprive the amriioniaco-magnesian phosphate of the whole of 
its water -[•, and as a very inconsiderable heat decomposes 
* See Dr Murray’s “ General Formula for the Analysis of Mineral Waters,” 
in the Edinburgh Philosophical Transactions. 
•j* Unless this be admitted, we must suppose a larger quantity of ammonia to 
enter into the composition of the triple phosphate than Dr Marcet imagines, ac- 
