fact that a solution of freshly precipitated magnesium pyrophosphate in excess of cold acetic acid is quantitatively precipitated on boiling, in a permanently insoluble form, while magnesium orthophosphate remains dissolved. The above weak acetic solution is mixed with magnesium acetate in excess, the precipitate dissolved by adding about one-fifth volume of strong acetic acid, and boiled a few minutes. The precipitated granular magnesium pyrophosphate is dissolved in nitric acid, in the manner described for the separation of the decomposition products of trimetaphosphimic acid, converted into silver salt and then into sodium salt. This, after several precipitations by alcohol, is recrystallized from water. A careful determination of the yield of pyrophosphoric acid in two experiments gave 13.5 and 16 per cent of the theoretical. For analysis the sodium salt was converted into silver pyrophosphate, which gave:

DECOMPOSITION OF SODIUM TRIMETAPHOSPHIMATE BY ACETIC ACID.

As before pointed out, this salt is not decomposed by acetic acid in the cold, or on short heating. If, however, its solution is heated for two or three hours with 30 per cent acetic acid and magnesium acetate, under the conditions mentioned under the magnesium salt, a granular precipitate is obtained, consisting of a mixture or double salt of magnesium pyrophosphate and imidodiphosphate, which does not perceptibly redissolve on cooling, while the liquid contains much orthophosphoric acid. The sodium pyrophosphate, separated from this by the above method, showed the characteristic form. For analysis it was converted into the silver salt, and gave:

3. ON TETRAMETAPHOSPHIMIC ACID.

Trimetaphosphimic acid and its decomposition products have been described in the preceding pages. Some account will now be given of the next higher member of the series, tetrametaphosphimic acid, P,N,O,H, which results from the decomposition by water of tetraphosphonitrilic chloride, P,N,Cl. The molecular weight of the chloride having been established with certainty, that of the acid follows directly. Although the free acid crystallizes with 2 molecules of water, the silver salts and

the acid potassium and ammonium salts are anhydrous and conform to the above formula.

49

Notwithstanding the proximity of these two acids in the series, and the general similarity of the chloronitrides from which they are derived, they differ exceedingly, trimetaphosphimic acid being extremely soluble, uncrystallizable, undergoing rapid spontaneous decomposition, and yielding but two series of salts; while tetrametaphosphimic acid is very difficultly soluble, highly crystalline, permanent when dry, offers marked resistance to the action of acids, and forms three series of salts, PN,OHM2, P ̧Ñ‚О,H,M1⁄4, and P,N,O,M's. Those metals which form insoluble salts are precipitated from solutions of their chlorides or nitrates even by the free acid, but a very slight excess of either being required to make the precipitation complete. As is the case with trimetaphosphimic acid, the 8-atom silver salt exists in a white and a yellow modification, the former, however, being very unstable; these may be regarded as corresponding to the tautomeric forms (PN(OH)2), and (PO.NH.OH). Under the action of stronger acids it is ultimately converted into orthophosphoric acid and ammonia, but its great stability, and the instability of the intermediate products, have rendered the isolation of the latter impossible.

Experimental data for a discussion of the constitution of the acid are therefore almost wholly wanting. Presumably the phosphorus atoms in the nucleus P,N, are united by nitrogen atoms, but whether they constitute a ring of 8 atoms can not be decided definitely in the absence of intermediate decomposition products. It is quite likely that the reason of the greater stability of the acid may be found in steric considerations, and in this connection it is interesting to observe that while PN,OH is much more stable than P3N3O6H6, the reverse is true of the chloronitrides, for while P,N.Cl, also resists the action of boiling water and can be distilled with steam, it is perceptibly easier to saponify than P,N,C16. If we assume the formulas

for the acids, it is obvious that the nitrile configuration is more stable in a ring of 6 than in one of 8 atoms, while for the group-NH -PO(OH)-NH-the reverse is true.

EXPERIMENTAL PART.

Preparation of tetrametaphosphimic acid.-Tetraphosphonitrilic chloride can be saponified by alcoholic alkalies, but the yield of tetrametaphosphimic acid is poor, owing to the formation of ethers, which, however, can be decomposed by prolonged treatment. Satisfactory results are obtained by using an ethereal solution, which is agitated, in the manner described under trimetaphosphimic acid, with either water, ammonia, or a solution of ammonium acetate.

Decomposition by water.-As tetrametaphosphimic acid is not perceptibly attacked by cold hydrochloric acid, water may be used when it is desired to obtain the free acid directly. One part P,N,Cl, is dissolved in 15 volumes alcohol-free ether and gently agitated with 5 volumes water. Within half an hour needles of the free acid begin to separate from the aqueous portion, and after many hours the latter has become converted into a thick pasty mass of needles suspended in dilute hydrochloric acid. Chlorhydrines are formed as intermediate products, and these remain dissolved in the ether. It is not necessary to continue the agitation until they are completely decomposed; as soon as a few drops of the ether leave a residue wholly soluble in water, the ether is decanted and the aqueous portion filtered, the crystals being washed with 10 per cent hydrochloric acid or alcohol. The ethereal portion, containing the chlorhydrines, is evaporated cautiously. The chlorhydrines which remain as an oil from which crystals separate after complete removal of the ether, are dissolved in a little cold water and warmed, with addition of dilute hydrochloric acid, in which tetrametaphosphimic acid is much less soluble than in pure water. The separation of the acid is nearly complete after a few minutes. This portion is better crystallized than that obtained directly, but is otherwise the same. The acid may be recrystallized from hot water, but this is quite unnecessary. Yield, about theoretical.

Decomposition by ammonia.-The ethereal solution of the chloronitride is shaken for a short time with aqueous ammonia. This does not give chloramide as is the case with PaNaCl, but produces at once neutral ammonium tetrametaphosphimate, P,N,OH,(NH),+4H2O, which is precipitated by adding alcohol and washed with dilute alcohol. The decomposition is complete in a few minutes.

Decomposition by ammonium acetate.-One part chloronitride, dissolved in 15 volumes ether, is agitated with a solution of 4 parts ammonium acetate in 8 parts water. After several hours crystals begin to separate. The final product is a mixture of acid ammonium salt with some neutral salt and some free acid, and is nearly insoluble in the strong salt solution. It is washed with alcohol, dissolved in dilute ammonia, and reprecipitated by alcohol. Yield, theoretical. Ammonium acetate is to be preferred to sodium acetate, as the sodium salt does not precipitate well with alcohol.

Chlorhydrines. None of these were isolated. The oil left on distill

ing off the ether solidifies to plates as soon as the ether is completely removed. The crystalline substance is much more soluble in ether than the chloronitride, even momentary exposure to ether-vapor causing it to liquefy. It dissolves in cold water to a clear solution, from which, in a few minutes, or sooner on warming, tetrametaphosphimic acid is deposited. On one occasion the flask containing the chlorhydrines was placed in hot water and a current of dry air passed through. Some hydrochloric acid was given off, and on treatment with water incomplete solution occurred, a portion being converted into a sandy powder, which, under the microscope, was seen to consist of short, spindle-shaped crystals. This acid was insoluble in boiling water, and warm ammonia converted it without dissolving into an acicular ammonium salt, difficultly soluble in water and insoluble in an excess of ammonia; nitric acid dissolved this salt, which was precipitated by ammonia in a bulky amorphous form like precipitated alumina, which quickly changed to needles. This ammonium salt was very slowly attacked by hot concentrated sulphuric acid, and contained 27.96 per cent phosphorus and 13.28 per cent nitrogen (P: N=4: 4.2). Lack of material prevented further study of this finely crystallized acid, of the nature of which I can form no conjecture.

Tetrametaphosphimic acid, P,N ̧OH8+2H2O.—The free acid may be prepared by the first method above described. It is not readily obtained pure by decomposing most of its soluble salts, as even in the presence of a large excess of acid more or less acid salt is deposited. Thus on dissolving neutral ammonium salt and adding a large excess of hot 5 per cent nitric acid, nearly pure acid salt was obtained. It may also be prepared by boiling the silver salt with much water, containing about the requisite amount of hydrochloric acid for its decomposition, and precipitating by further addition of the same acid to the filtrate.

The analytical methods are the same as those employed for trimetaphosphimates, but much longer heating with strong sulphuric acid is necessary because of its greater stability.

Analyses of four air-dried preparations gave:

3. P N H 4: 4.06 12.43. From chlorhydrine.

4. PN: H-4 4 12.64. First method, from aqueous portion.

The crystal water is not given off in vacuo over sulphuric acid. At 100° it loses weight rapidly, then slowly, but the total loss never

reaches the theoretical, a portion of the water being taken up in producing intramolecular changes.

Loss calculated for 2H2O

10. 22
9.37

8.91

(1) Loss after twenty-four hours (weight constant) (2) Loss after seventy-eight hours (weight constant)... If rapidly heated to 130-140° the loss is still less, and is soon replaced by a gain; after thirty hours at this temperature a net increase of 11.15 per cent, corresponding to somewhat more than 2 molecules H2O, was observed, which must have come from the atmosphere. The product consisted of unchanged acid with ammonium phosphate and apparently pyrophosphate.

Tetrametaphosphimic acid forms colorless needles, either single or in radiating groups, and visible without a lens; they appear to consist of flat rectangularly terminated prisms. One hundred parts water at 20° dissolve 0.64 part crystallized acid; in boiling water it is somewhat more soluble, but is insoluble in alcohol. From its saturated aqueous solution it is partially precipitated by adding one of the stronger acids, and more rapidly if heated; 100 parts by weight of 10 per cent nitric acid at 20° dissolve only 0.26 part of the crystallized acid, but on decomposing its salts in the cold by an excess of acid it frequently dissolves completely and separates out later. Its saturated aqueous solution does not coagulate albumen. Boiling alkaline solutions cause no evolution of ammonia. Its stability toward acids is illustrated by the following: 0.1 gram, dissolved in 200 cubic centimeters water with 8 cubic centimeters strong nitric and some hydrochloric acid, was evaporated to dryness on the water bath, whereby a large part was recovered unchanged; ten minutes' heating would have sufficed for the total decomposition of the same quantity of trimetaphosphimic acid. Nitrous acid has no perceptible action.

Di-potassium tetrametaphosphimate, P,N,OH, K.-The free acid dissolves easily in cold dilute caustic potash; on strongly acidifying with acetic acid the solution remains clear, but on warming the above salt is deposited as a heavy, sandy powder, consisting of microscopic thick rectangular (quadratic?) prisms with basal planes. It is very difficultly soluble even in boiling water.

The air-dried substance lost nothing at 1000 and gave:

The tetra-potassium salt formed large, flat, obliquely terminated plates, very soluble, and was not isolated.

Tetra-sodium tetrametaphosphimate, P,N,O,H,Na, + 2(?) H2O.—The acid is suspended in a little water and an excess of dilute caustic soda