OF ARTS AND SCIENCES. 307 



tion, although we have not succeeded in isolating the benzyl bro- 

 mide itself, or one of its derivatives. This did not proceed from an 

 impurity of benzyl bromide in the benzyl alcohol used, since after 

 boiling it with an alcoholic solution of sodic acetate no test for so- 

 dic bromide could be obtained ; it must have been formed, therefore, 

 either in the course of the reaction, or in the subsequent processes 

 of purification. It is barely possible that this latter supposition 

 may account for it in the last case, where the strongest test for 

 bromine was obtained, as this was acidified before distillation with 

 steam, and hydrobromic acid produced by the sulphuric acid and 

 sodic bromide might have converted some of the excess of benzyl 

 alcohol into benzyl bromide, although this seems highly improbable, 

 because the solution was kept throughout in a very dilute state. 

 In the two previous cases, however, this explanation cannot apply. 

 as no acid whatever had been added, so that no hj'drobromic acid 

 could have been formed; we are forced therefore to assume that the 

 benzyl bromide was formed by the direct reaction and not bj r a sec- 

 ondary one, and under these circumstances the following reaction 

 is the only one which seems to us admissible : 



3 C 7 H 7 OXa + C 6 Br s (N0 2 ) 3 = C 6 (ONa) 3 (N0 2 ) 3 + 3 C 7 H 7 Br. 



This is certainly very improbable, and we should add, that, when 

 we tried to confirm it by a repetition of the experiment, we did not 

 succeed in detecting any benzyl bromide in the product. At first 

 sight it would seem that a reaction the reverse of that given above 

 would be more probable, but we proved that this was not the case 

 by a special experiment, which showed that benzyl bromide had no 

 action on the sodium salt of trinitrophloroglucine under the condi- 

 tions used by us. We may add that amyl bromide also has no action 

 under these conditions. On the other hand, there is little doubt 

 that the benzyl bromide, if formed by this reaction, would act on 

 the sodic benzylate forming benzylether. Accordingly, we tried to 

 isolate benzylether from the oily product of the reaction by distil- 

 ling off the excess of benzyl alcohol; there was so little of the resi- 

 due which did not pass over at 207° that it was impossible to purify 

 it properly; such as it was, however, it gave results on analysis 

 agreeing with those calculated for benzyl alcohol (carbon 76.70 

 instead of 77.78). We conclude, therefore, that if the reaction given 

 above takes place, it is only to a very moderate extent, the principal 

 reaction being the formation of the trinitrophloroglucine tribenzyl- 

 ether, which was then more or less saponified by the sodic hydrate 



