268 A GENERAL FORMULA FOR THE 
base, it is probably unexceptionable. It does not, however, 
altogether accord with the object of the present formula. The 
soda of the phosphate of soda serves to neutralize the muriatic 
acid of the muriate of magnesia ; a quantity of muriate of soda 
is of course formed, which remains with the muriate of soda of 
the water, and the amount of which, therefore, it is necessary 
to determine with accuracy. This may be done from the 
quantity of phosphate of magnesia obtained giving the equiva- 
lent portion of muriate of soda, either by means of the equiva- 
lents of the acids, or of the bases. But still this renders the 
method somewhat complicated ; and it may be liable to some 
error, if any excess of phosphate of soda be added, which, in 
order to precipitate the magnesia entirely, it may be difficult 
to avoid ; this excess remaining with the muriate of soda, and 
rendering the estimate of it incorrect. And independent of 
these circumstances, it would be preferable to give uniformity 
to the operation, by employing some method by which the 
product in this, as well as in the previous steps, is removed, at 
the end of the analysis, leaving only the muriate of soda. 
It seemed probable that this might be attained, by employ- 
ing phosphoric acid with the carbonate of ammonia, to form 
the triple phosphate of ammonia and magnesia, such an excess 
of ammonia being used, as should both be sufficient for the 
constitution of this compound, and for the neutralization of the 
muriatic acid of the muriate of magnesia; muriate of ammo- 
nia would thus be substituted, the same as in the preceding 
step of precipitating the lime, which at the end would be ex- 
pelled by heat, leaving muriate of soda alone. I accordingly 
found, that when this variation of the process was employed, 
the clear liquor, after the precipitation, was-not affected by the 
addition either of phosphate of soda with ammonia, or of sub- 
carbonate of soda,—a proof that the separation of the magnesia 
had 
