OF ARTS AND SCIENCES. 283 



time, and still more by the addition of alcohol, but the color disappears 

 if the heating is long continued. Aqueous potassic carbonate gave no 

 action, but on the addition of alcohol a very dark purplish red solution 

 of the salt. Acid sodic carbonate had no action in aqueous solution, 

 and only very slight on addition of alcohol. The dark red alcoholic 

 solution of the ammonium or sodium salt, if made with an excess of 

 the ketone, is decomposed by water, giving a white precipitate of the 

 ketone. The decomposition is not complete, however, as the liquid 

 retains a pale red color. In the presence of an excess of sodic hydrate, 

 the salt is much more stable. An attempt was made to study the 

 action of the pale red aqueous solution of the ammonium salt with 

 various reagents, but no characteristic precipitates were obtained, 

 probably on account of the small amount of salt left in solution. 



We tried also to prepare and analyze the sodium salt, the method 

 being that adopted for the sodium salt of the bromdinitrophenylaceta- 

 cetic ester; but the results of the analyses came much too high, 9.73 

 and 8.64 per cent of sodium, instead of the 7.07 per cent required by 

 the formula. The reason for this difference is that the sodic hydrate 

 or ethylate removed a portion of the bromine from the ketone, and the 

 benzol, dissolving the organic product, left the sodic bromide formed 

 with the salt of the ketone, as was proved by dissolving the salt in 

 water, and acidifying with nitric acid, when, after removing the pre- 

 cipitate by filtration, argentic nitrate gave a heavy precipitate of argen- 

 tic bromide in the filtrate. We accordingly turned our attention to 

 the anilidoketone, our work on which is described later in this paper. 

 We add here the properties of the sodium salt of bromdinitrobenzyl- 

 metltylketone. It is a purplish black amorphous substance, very easily 

 soluble in alcohol forming a dark claret-red solution, the color of which 

 is so much more purple than that of the salt of the corresponding 

 acetaoetic ester, that the two substances can be easily distinguished in 

 this way ; it is decomposed almost, but not quite, completely by water, 

 and is insoluble in benzol. 



The ketone is a decidedly reactive substance. When treated with 

 aniline, it <nves aniline bromide and the anilidodinitrobcnzvlmethyl- 

 ketone, which is described later. With phenylhydrazine it appears to 

 form a hydrazone, but at the same time the bromine was removed 

 from the benzol ring, so that the product seemed to be a pbenylhy- 

 drazidohydrazone. It was free from bromine, and exploded when 

 touched with a hot wire; but as its purification offered considerable 

 (litliculties, we did not try to investigate it thoroughly. The ketone 

 seems to react in the same way with hydio.xylamine, since in this • 



