292 PROCEEDINGS OP THE AMERICAN ACADEMY 



before allowing it to evaporate spontaneously. The residue, after 

 washing with water, is purified by crystallization from alcohol, 

 which separates the two ethers formed by the parallel reactions. 



Quantitative Study of the Action of Sodic Ethylate on 

 Tvibromtrinitrobenzol. 



After we had proved that this action takes place in these two 

 different ways, 



C 6 Br 3 (N0 2 ) 3 + 3 NaOC 2 H 5 = C 6 (OC 2 H 5 ) 3 (N0 2 ) 3 + 3 NaBr, 

 C 6 Br 8 (N0 2 ) 8 + 2 NaOC 2 H 5 = C 6 Br 3 N0 2 (OC 2 H 5 ) 2 + 2 NaN0 2 , 



it became of interest to determine the relative extent of each of the 

 parallel reactions under varying conditions, and the easiest way to 

 do this was obviously to find the amount of sodic bromide and sodic 

 nitrite formed, as the first must have proceeded from the reaction 

 giving the trinitrophloroglucine triethylether, the second from that 

 producing the tribromnitroresorcine diethylether. In addition to 

 this we have weighed the amount of organic matter formed, and have 

 made some not very successful attempts to determine the amount 

 of its several constituents. The method adopted was the follow- 

 ing. A weighed quantity of tribromtrinitrobenzol mixed with the 

 solvent and the ethylate in the proportion of three atoms of sodium 

 to each molecule stood over night in a corked flask. The liquid was 

 then allowed to evaporate spontaneously, and the residue washed 

 thoroughly with water. The weight of the organic matter insoluble 

 in water was taken, and also that of the aqueous solution which 

 contained all the sodium salts ; this was then divided into weighed 

 parts, one of which was acidified with nitric acid, and after filtering 

 out and weighing the diethylether of trinitrophloroglucine the 

 amount of sodic bromide it contained was determined with argentic 

 nitrate. The second portion of the aqueous liquid was used for the 

 determination of the amount of sodic nitrite, but it was not easy to 

 find a method which would be applicable in this case. The com- 

 mon way of determining a nitrite with potassic permanganate was 

 inadmissible, because of the presence of organic matter. The urea 

 method in its usual form was also inapplicable because of the pres- 

 ence of sodic carbonate, but finally by modifying it in the following 

 way we succeeded in obtaining satisfactory results. The weighed 

 portion of the solution to be tested, mixed with a sufficient quantity 

 of urea, was poured into a little flask, and a small test-tube contain- 

 ing dilute sulphuric acid put into the flask supported in a vertical 



