DIIMIDOTRIPHOSPHORIC ACID. Ill 
The Lri- sodium salt is not precipitated even in strong solution by 
magnesium chloride or magnesia mixture unless the latter contains 
much ammonia; in this case the resulting precipitate is easily soluble 
in ammonium salts, and hence is probably merely magnesium hydroxide. 
Barium chloride precipitates from a not too dilute solution microscopic 
spherules, which dissolve readily in a cold, strong solution of sodium 
chloride. On boiling this solution a crystalline precipitate is at once 
formed, consisting of microscopic rhombohedra (?) resembling sodium 
barium trimetaphosphimate and probably a double salt. 
Silver salts. — Two of these were obtained, one having the formula 
P 3 N20 ;! II 4 Ag3, and the other P 3 2 8 H 2 Ag 5 . The tendency to formation 
of the latter is so strong that a solution of the tertiary sodium salt gives, 
with silver nitrate, a white amorphous precipitate, which, however, is 
mixed with some crystals of the former. 
Tri-silver diimidotr (phosphate, P 3 N 2 0;jHiAg ; j. — This salt is very char- 
acteristic. It can be obtained pure only by precipitating an acidified 
solution of the tertiary sodium salt by an excess of silver nitrate. To 
a solution of the sodium salt in 30 parts water, 1 to 2 molecular weights 
nitric acid are added, and then silver nitrate gradually; as soon as a 
milkiness, due to the formation of the penta salt, is observed, more acid 
is added, and then more silver nitrate, acid being added as often as 
necessary:, the crystallization soon starts, but several times the theo- 
retical amount of nitrate is required, and under no circumstances is the 
precipitation complete. Generally spherules with many rhombic facets 
are obtained, but with slow crystallization (from more dilute solutions) 
the forms shown in figs. 5 and result. 
The colorless crystals are monoclinic, and suggest those of tertiary 
ilver trimetaphosphimate, but are shorter and thicker, and approach 
lore nearly to the orthorhombic form. No other faces are ever seen, 
lie angle a (between the edges), measured under the microscope, is 
ery nearly 98°; the acute angle between the edges of B is, roughly 
peaking, 80°, so that when the form fig. G occurs, and this face is 
ppermost, it is apt to be mistaken under superficial observation for 
:inietaphosphiiuate (angle a, 78°). 
