stokes.] TR1METAPHOSI HIMIC ACID. 1<>7 
Unsatisfactory as these figures are, they appear to indicate the exist- 
ence of: 
(1) A yellow salt, usually amorphous, occasionally crystalline, approxi- 
mating in composition to P 3 N 3 7 H2Ag(;. As this is only formed in the 
cold from a trimetaphosphimate to which alkali has been added, it may 
be a salt of the supposed open-chain acid. 
(2) A red crystalline salt of the formula P 3 N 3 6 Ag 6 . 
(3) A white amorphous salt of nearly the same composition. 
The orange-red salt, invariably containing too little silver, and vary- 
ing in depth of color, may be a mixture or an intermediate form. 
As the white salt is formed only in the presence of silver ammonia 
compounds, and is quite soluble in ammonium nitrate, it is clearly dif- 
ferent in nature from the others. Probably it contains silver united to 
oxygen only, while in the red salt one-half the metal is associated with 
nitrogen. We may regard them as salts of the two forms of trimetaphos- 
phimic acid : 
P(OAg), 
/ \ 
N N 
II I 
(AgO),P P(OAg), 
PO.OAg 
AgN' 
NAg 
AgO.PO 
PO.OAg 
N 
Ag 
Hexa-silver trimetaphosphimate 
Hexa-silver triphosphonitrilate. 
The action of silver nitrate in preventing the transformation may be 
ascribed to its opposing dissociation. No deep decomposition occurs 
luring the transformation, as the ammoniacal solution is not precipi- 
tated by magnesia mixture, and sodium trimetaphosphimate is regen- 
erated by treatment with sodium chloride and acetic acid. 
Ferric trimetaphosphimates. — Gladstone 1 states that when pyrophos- 
ohodiamic acid is added to ferric sulphate, acidified with sulphuric 
tcid, a clear solution results, which, on boiling, deposits white ferric 
)yrophosphamate, which is insoluble in acids bnt soluble in ammonia. 
The ferric trimetaphosphimates seem to vary considerably in composi- 
ion and properties and present no characteristic features. The follow- 
ug points were noted: On adding ferric chloride to a moderately 
trong solution of sodium trimetaphosphimate an amorphous white 
>recipitate forms which redissolves in an excess of the sodium salt, 
lobably with formation of a ferric-sodium salt, but which becomes 
•ermanent on adding more ferric chloride. This is scarcely soluble in 
ilute hydrochloric acid, but readily soluble in ammonia with faint red 
olor. Sodium hydroxide regenerates sodium salt. If, however, the 
olution of the sodium salt be dilute, no precipitate forms in the cold, 
j ven with an excess of ferric chloride, but on boiling, alter acidifying, 
'Quart Jour. Chem. Soc. London. Vol. Ill, p. 1^5; Jour. Cheni. Soc, London, [2], Vol. II, p. 229; 
], Vol. IV, p. 290; [2], Vol. VI, p. 67. 
