ACTION OF PHOSPHIDE OF SODIUM ON HALOID ETHERS, yal 
ticular appeared to have occurred, the contents of the tube consisting of a brown 
viscous mass. But on boiling this with water, and cooling the solution, chloride 
of tetrabenzyl-phosphonium crystallised out, and was proved to be pure by a 
determination of chlorine. 
0°656 required 13°5 cc. decinormal AgNO,=7°3 per cent. Cl. 
(C,H,),PC1,2H,O requires : : Paro Ate . 
This experiment would have definitely proved the resin to consist of tribenzyl- 
phosphine, had the phosphonium chloride been produced in large quantity ; but 
such was not the case, for about 20 grms. of resin only gave about 2 grms. of 
the chloride ; still it shows that the resin contained the phosphine. 
As Hormann (loc. cit.) has found that dibenzyl-phosphine does not combine 
with acids, we could scarcely expect to obtain salts of the tertiary-phosphine. 
We, however, heated the resin with fuming hydrochloric acid, but, as we 
expected, obtained no salt. We have also tried to obtain the well characterised 
oxide of tribenzyl-phosphine, by treating the resin with various oxidising 
agents, but without success. Nor could we obtain any definite compounds on 
treating the resin with bromine, chloride of platinum, or iodide of zinc. We 
therefore abandoned its further investigation, 
Attempts to prepare Tribenzyl-Phosphine. 
So far as we are aware, no method has been discovered for converting 
the oxide of a tertiary-phosphine or a salt of a compound phosphonium into a 
tertiary-phosphine itself. 
Considering the remarkable stability of the former class of bodies, and the 
tenacity with which the oxygen adheres to the phosphorus, we scarcely expected 
to effect the reduction of the oxide of tribenzyl-phosphine. We, however, sub- 
jected it to the action of potassium, of sodium, and of zinc dust, but, as we 
expected, it either remained unchanged, or suffered complete decomposition. 
We hoped, however, to meet with better success in attempting to obtain 
tribenzyl-phosphine from chloride of tetrabenzyl-phosphonium. Two methods 
suggested themselves to us, the first being to act on the chloride with sodium, 
which we anticipated would give chloride of sodium, dibenzyl, and the 
phosphine, 
2[ (C,H,),PCl] + Na,=2NaCl+C,,H,,+2(C,H,)sP. 
A preliminary experiment showed that when chloride of tetrabenzyl- 
phosphonium is boiled for some hours with xylol and sodium, chloride of 
’ sodium is produced, 
We therefore made a carefully conducted experiment as follows :— 
24 orms. of the pure chloride were carefully dried and introduced into a 
