108 CONTRIBUTIONS TO CHEMISTRY AND MINERALOGY. Ibull.167. 
a white amorphous precipitate is produced, which, like the above, is 
nearly insoluble in dilute acid and easily soluble in ammonia and con- 
sists of trimetaphosphimate mixed with decomposition products. 
The solubility of these salts in ammonia, which they show in common 
with other ferric amidophosphates and pyrophosphates, maybe ascribed 
to the formation of complex ions containing iron; in other words, of 
ferri-triruetaphosphimic acids, analogous to ferri- oxalic acid 1 and simi- 
lar bodies. 
Free trimetaphosphimic acid. — Silver trimetaphosphimate, decom- 
posed under cold water by hydrogen sulphide, gives a solution from 
which, when fresh, the characteristic salts may be prepared, and from 
which much alcohol gradually throws out an amorphous sticky mass. 
The solutiou, on evaporation in vacuo, leaves a transparent, gummy 
residue from which alkalies evolve much ammonia, and which, after 
redissolving, gives a precipitate with magnesium mixture, indicating 
partial decomposition. The acid seems therefore to be incapable of 
crystallization. 
In marked distinction from metaphosphoric acid, trimetaphosphimic 
acid does not coagulate albumen. 
Decomposition Products of Trimetaphosphimic Acid. 
Orthophosphoric acid and ammonia are the ultimate products of the 
action of acids on trimetaphosphimic acid, as well as of the decompo- 
sition of the free acid alone in aqueous solution when heated or kept 
for a long time. I have, however, been able to isolate three well-marked 
intermediate bodies, viz : 
Diknidotriphosphoric acid, P 3 N 2 0«H 7 ; 
Imidodiphosphoric acid, P^NOoH, ; 
Pyropliosphoric acid, P2O7H4. 
If sodium trimetaphosphimate solution (1-15) 'be boiled with 2-3 
molecular weights nitric acid, the transformation into ammonia and 
orthophosphoric acid is complete in about ten minutes, but on inter- 
rupting the boiling after three to four minutes and cooling, the presence 
of these products can be detected; the same change occurs slowly in the 
cold, many days being required for complete decomposition. From 
this solution silver nitrate throws down first the silver salt of un- 
changed trimetaphosphimic acid, then the others in the order named, 
and finally silver phosphate, which is most easily soluble in nitric acid. 
Although these products were first detected in this way, many experi- 
ments have convinced me that it is not a practical method of separation. 
The separation by fractional precipitation is not sharp; the properties 
of trimetaphosphimic and diimidotriphosphoric acids are so similar and 
the stability of the latter so slight that the results are very unsatisfac- 
tory. In the absence of trimetaphosphimic acid, however, diimidotri- 
phosphoric acid is readily separated from the later decomposition 
'Sec Rosenheim, Zeitschr. anorg, CLu-m., Vol, XI, p, 214, etc. 
