144 
CONTRIBUTIONS TO CHEMISTRY AND MINERALOGY. 
[BULL. 107. 
Calculated, for 
PfiN 6 0,oH 6 .Mg 2 H+5H a O. 
Found. 
1. 
2. 
p 
29.26 
9.16 
28.82 
9.52 
29. 05 
10.10 
Mer 
1. P:Mg = 5:2.11. 
2. P: Mg = 5: 2.22. 
A considerable portion of the acid remains in solution, however, even 
in the presence of a large excess of the precipitant, in combination 
with less than 2 atoms of magnesium, and can be precipitated by alco- 
hol as a salt soluble in water. If this insoluble magnesium salt be 
dissolved in dilute nitric acid and ammonia added to incipient precipi- 
tation, the solution, after filtering, contains essentially the primary 
salt (PsNsOioHgVJMg; from this solution silver nitrate throws down an 
amorphous magnesium silver salt with a varying amount of silver. 
The primary salt is easily soluble in water and is remarkably stable, 
giving no precipitate with ammonia, sodium hydroxide or carbonate, 
even on boiling. The magnesium can be removed only by adding to 
its solution ammonia alid an alkaline phosphate. On boiling in neutral 
or acetic acid solution, however, a precipitate of the 2-atom salt at 
once forms. Even the latter dissolves in boiling sodium carbonate 
solution. 
Indications of an intermediate salt were observed, but it could not 
be isolated in pure condition. 
Silver salts. — The silver salts of the phosphorus nitrogen acids are 
invariably free from water, and it is upon these, therefore, that the 
formulas of the acids themselves are based. There is no difficulty in 
obtaining normal silver tri- and tetrametaphosphimates of theoretical 
composition, and their crystalline nature affords a guaranty of their 
homogeneity. The same difficulty is encountered with the silver penta- 
metaphosphimates, however, as with the sodium salts; they are amor- 
phous tlocculent precipitates, the composition of which corresponds to 
a definite formula only when they are formed under special conditions. 
In this case, as in others in this paper, the actual percentage composi- 
tion expresses very little, if compared with the calculated composition 
of a definite salt. It is therefore better to express the results of the 
analysis in a molecular formula based on the atomic ratios of phos- 
phorus, nitrogen, and silver, as actually determined; a comparison off 
the percentage composition found, with that calculated for salts of the 
lactam and open chain acids containing phosphorus and silver in the 
same ratio, then shows at once to which of these acids the salt is to be, 
referred. 
The composition of the precipitates depends altogether on the rela- 
tive amounts of the reacting bodies, and even the free acid can be almost 
