82 CONTRIBUTIONS TO CHEMISTRY AND MINERALOGY. [bull. 167, 
compound is formed than is actually obtained, which, because of its 
less volatility, fails to distill over before it is attacked by the ammo- 
nium chloride, and I have also mentioned the simultaneous formation 
of large quantities of chlorides, unacted on by cold, but decomposed 
by boiling water. Still, it seems that there is some cause leading to 
the predominant formation and greater stability of this body, perhaps 
analogous to that acting in the case of the aromatic hydrocarbons. The 
proof that P 3 N 3 01 6 is a cyclic compound has not been found, but the 
occurrence of the number 6 in each case (0 6 and P 3 N 3 ) is possibly more 
than a mere coincidence. Tetraphosphonitrilic chloride i s also possessed 
of great stability, but I have observed, in various connections, that its 
chlorine is less firmly held than in the triple compound. The reverse 
appears to be true of the derived acids, P 4 N 4 8 H 8 being more stable 
than P 3 N 3 6 H 6 . 
Analytical methods. — The analysis of these bodies presents no diffi- 
culty. The chlorides, which volatilize on heating, are gently warmed 
with dilute alcoholic soda or potash. In this solution chlorine is 
directly determined as silver chloride; phosphorus is determined after 
evaporating off the alcohol, either by fusing the residue with sodium 
carbonate or by boiling for one or two hours with strong sulphuric acid, 
either of which converts it wholly into phosphoric acid. Nitrogen is 
determined as ammonia after acidifying the saponification product with 
hydrochloric acid, evaporating the alcohol, and boiling the residue with 
strong sulphuric acid. Experiment showed that no ammonia is evolved 
during saponification. In the case of the acids and other nonvolatile 
derivatives, the alcoholic saponification is omitted and the substance 
decomposed either by boiling with concentrated sulphuric acid, or 
fusing with sodium carbonate, the precaution being taken in the latter 
case first to moisten the mixture in the crucible. Owing to the strong 
reducing action of the imido group at high temperatures, these sub- 
stances can not be fused alone in platinum without risk to the crucible. 
This is true even of the salts, which should give pyro- or metaphos- 
phates on ignition. 
The chlorides. — As triphosphonitrilic chloride has been the object of 
previous investigation, and as I have begun several lines of investiga- 
tion with the object of explaining its nature and relations, I state here 
only a few incidental observations. 
This chloride, as was early observed, has great crystallizing power. 
The rhombic crystals have been measured by Miller 1 and by Groth. 2 
When pure it tends to form large, thick prisms. I have obtained these 
(from benzene) as much as 8 centimeters long and 2 centimeters wide, 
their further growth being hindered only by the size of the flask. When 
impure, it tends to form rhombic or 6-sided plates. It fuses at 114° (corr.), 
while the quadruple chloride, P 4 N 4 01 8 , fuses at 123.5° (corr.), but a 
1 Jour. Chera. Soc. London [2], Vol. II, p. 227. 
2 Ber. Deutsoh chem. Geaell., Berlin, Vol. Ill, p. 166. 
