added, whereby complete solution occurs. The separation of the neutral salt soon begins in the form of obliquely terminated flat prisms, sometimes single, sometimes in twins or quadruplets; the filtrate deposits more on heating, but in imperfect forms, broader at one end than at the other. Sometimes nothing separates from the alkaline solution until it is heated. On the contrary, strong caustic soda converts the solid acid in the salt without dissolving it. The salt is rather difficultly soluble in cold, quite easily in hot, water, but much less soluble in the presence of an excess of alkali, and the latter throws it down from its saturated aqueous solution. Sodium carbonate gives the same salt.

The substance, pressed out without washing and dried in vacuo, lost nothing at 100° and gave:

Alcohol precipitates this salt in a slimy form, which filters with dif ficulty.

A solution of the tetra-sodium salt, strongly acidified with acetic acid, gave a deposit of granules, probably the acid salt. A salt with more than 4 atoms of sodium' was not observed.

Tetra-ammonium tetrametaphosphimate, P,N ̧O2H1(NH4),+4H2O.— Strong ammonia converts tetrametaphosphimic acid into this salt without dissolving it; from its solution in weaker ammonia it is precipitated by alcohol. It usually forms well-developed, flat, monoclinic prisms with pinakoid and basal planes, the base being occasionally replaced by a set of many small planes. It is easily soluble in water, but difficultly in strong ammonia. Its solution undergoes partial decomposition into acid salt on evaporating. It loses nothing in vacuo, and analysis gave:

1 Compare the behavior of trimetaphosphimic acid toward excess of alkali.

8

Di-ammonium tetrametaphosphimate, P4N4OH6 (NH4)2.-This salt forms whenever a solution of the neutral salt is strongly acidified, and was at first mistaken for the free acid, which it somewhat resembles. It forms 4- and 6-sided prisms with base (tetragonal?), closely resembling the acid potassium salt, and, like this, is very difficultly soluble even in boiling water. It separates on heating a solution of the neutral salt with acetic acid in excess (analyses 1 and 2), or with a stronger acid, in which case it may be slightly contaminated with free acid. The preparation under analysis 3 was crystallized from a large excess of hot 5 per cent nitric acid. It contains no crystal water. Analysis gave:

Barium tetrametaphosphimate, PN4OH,Bа2+ 2H2O.-A solution of the di- or tetra-ammonium salt in 500 parts cold or, better, boiling water, is precipitated by an excess of barium chloride. The salt forms a voluminous precipitate consisting of microscopic needles, branched or forked, insoluble in water, and undecomposed by acetic acid. These have approximately the composition of neutral salt, with a slight deficiency of barium.

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

Basic barium salts of indefinite properties seem to exist, but no indication of an acid salt could be obtained.

No satisfactory results were obtained with magnesium salts, a large variety of crystalline and amorphous precipitates being obtained under different conditions.

Manganese salts give with free tetrametaphosphimic acid a flocculent

precipitate of neutral manganese tetrametaphosphimate; on adding a mineral acid this is decomposed, according to the proportion of the latter, into acid salt or free acid, or a mixture of both. The acid salt forms handsome pinkish rectangular plates, often superposed at various angles, and is one of the most characteristic salts obtained.

Ferric chloride gives no precipitate with a cold dilute solution of the neutral ammonium salt; on warming a white amorphous precipitate forms, which is soluble only on warming with much hydrochloric acid, from which solution, on cooling, the free acid separates.

Tetra-silver tetrametaphosphimate, P,N,O,H,Ag.-If to a cold solution of tetrametaphosphimic acid silver nitrate be added the resulting precipitate is white, curdy, and, under the microscope, granular or amorphous; formed in the presence of some alcohol it is even flocculent. These amorphous precipitates, after drying, slowly change to needles, and this occurs at once if they be boiled under water. The salt is obtained at once in crystalline form if a slight excess of silver nitrate be added to a hot solution of tetrametaphosphimic acid in 200-300 parts water. It requires but a slight excess of silver nitrate to bring about complete precipitation, the liberated nitric acid exercising a scarcely perceptible solvent action. Neutral ammonium salt may be used, but the addition of a little nitric acid is essential, otherwise the product is yellowish from contamination with the 8-atom silver salt.

When precipitated hot the salt sometimes consists of thick needles with tufts or brushes at each end, sometimes of long pointed plates, more rarely of rhombic plates, and in no case are they well formed or very characteristic. An acid salt could not be obtained, but the neutral salt always shows a slight deficiency of silver.

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

Octa-silver tetrametaphosphimate, P,N,O,Ag.-If an ammoniacal solution of a tetrametaphosphimate be added to an excess of silver nitrate, a yellowish flocculent precipitate forms, which is insoluble in ammonium nitrate and which does not change on boiling under its mother-liquor

or under pure water. On drying at 100° it became grayish olive colored, apparently because of the separation of some silver oxide. It gave the figures under 1, 2, and 3.

If a solution of a neutral tetrametaphosphimate be added to ammoniacal silver nitrate containing the least possible excess of ammonia, a nearly white flocculent precipitate forms in small amount, which is soluble in ammonium nitrate solution on gently warming; from this solution an orange-yellow semicrystalline body is thrown out on boiling for a moment. The latter is obtained in greater abundance by boiling the original filtrate from the silver-nitrate precipitate. It seems to be merely the crystalline modification of the above amorphous yellow salt. It gave the figures under analysis 4. The nearly white modification is obtained in greater amount by using a solution of silver oxide in ammonium nitrate solution. It shows a considerable deficiency of silver (P Ag 4: 6.7) and seems to be less stable than the corresponding white hexa-silver trimetaphosphimate, as it can not be obtained entirely free from the yellow modification. The yellow form also shows a slight deficiency of silver, and unlike the trimetaphosphimate, it can not be converted into a red form.

=

4. THE HIGHER CHLORONITRIDES.

In the first section of this paper I have shown that in addition to the phosphonitrilic chloride,' P3N3Cl, discovered by Liebig, there exists another, PN,Cl, of similar properties, which is formed at the same time, but in smaller quantity. The opinion was expressed that these bodies belong to a series of polymers, (PNC12), the existence of other members of which was indicated by the formation, in small amount, of a liquid of the same empirical composition. The yield of this secondary product, only 2 per cent of the theoretical or 1 per cent of the pentachloride used, was too small to allow of its preparation in quanti

I propose in future to use the term phosphorus chloronitride to denote any body composed of phosphorus, nitrogen, and chlorine, the name phosphonitrilic chloride being reserved for chloronitrides belonging to the series (PN Cl2).

ties large enough to admit of the isolation of its supposed constituents, but a fractional distillation of the few grams at my disposal showed that it contained crystalline substances of higher boiling points than those of the two bodies thus far known.

The method of preparation then employed consisted in distilling phosphorus pentachloride with a large excess of ammonium chloride in a retort, at atmospheric pressure; it offered but little prospect of obtaining the higher members. The total yield of phosphonitrilic chloride was but 15 per cent of the theoretical, most of the pentachloride being converted into "phospham" by the excess of ammonium chloride, while only those members could be obtained which distill unchanged at atmospheric pressure. Decreasing the amount of ammonium chloride resulted only in a loss of pentachloride by volatilization, without increasing the yield of the bodies sought after.

The following method has been found to give entirely satisfactory results; several new bodies have been obtained, and the simpler phosphonitrilic chlorides, at least, are now easily accessible substances. If equal molecular weights of phosphorus pentachloride and ammonium chloride be heated in a sealed tube, there results a mixture of chloronitrides, which is partly crystalline and soluble in gasoline, but for the greater part liquid and insoluble in this solvent, and of a high degree of complexity. This may be distilled almost without residue, the distillate being a crystalline mass, impregnated with an oil, and composed almost wholly of a mixture of members of the series (PNC12) in nearly theoretical amount, containing about 50 per cent PзN Cl, and 25 per cent P,N,Cl, the remainder consisting of the higher homologues. From this distillate the new bodies, with one exception, have been isolated.

The series, as at present known, consists of the following:

3

Boiling point.
(Corrected.)

760 mm.

There were obtained, further, a liquid residue of the same empirical composition, of a mean molecular weight corresponding nearly to