143 
few hours the clear supernatant liquid is carefully poured 
through a niter, taking care not to disturb the precipitate. 
In the filtrate the phosphoric acid is estimated as magnesium 
pyrophosphate. The amount of the fluid employed in the 
determination is ascertained by again weighing the flask. 
On substracting the weight of the tin sulphide calculated 
from the quantity of the metal originally employed, we 
have all the data required to determine the amount of 
phosporic acid in the entire solution. 
This method of proceeding is not altogether faultless in 
principle. (1) It presupposes that from a known weight of 
tin-foil we are able to calculate the amount of tin-sulphide 
it will yield. Now the tin-foil of commerce is seldom or 
never pure, it almost invariably contains a considerable 
proportion of lead, often amounting to ‘ one-third of its 
weight,* and this of course passes into the nitric acid 
solution of the alkaline earths. (2) Since only a por- 
tion of the phosphoric acid present is actually weighed, 
the remainder being deduced by calculation, the chances of 
ultimate error are considerably increased. 
These sources of error are removed by simply filtering 
and washing the tin sulphide by means of the Bunsen 
“ water-pump,” an operation of comparative short duration. 
We thus obtain the whole quantity of phosphoric acid in 
solution, and entirely obviate the numerous weighings, 
involving, too, the very uncertain correction for the amount 
of tin sulphide present. 
In order to test the trustworthiness of the method thus 
modified, the following experiments were undertaken. A 
quantity of pure calcium phosphate was prepared by adding 
calcium chloride to an excess of sodium phosphate, and the 
precipitate washed, dried and ignited. About 0 - 5 gram, of 
this compound was weighed out into a porcelain basin, and 
dissolved in a small quantity of nitric acid; the solution 
* The tin-foil employed in my experiments contained 31*35 per cent lead. 
