246 PROCEEDINGS OF THE AMERICAN ACADEMY 



we did not succeed in finding any solvent from which it could be crys- 

 tallized. That the sodic hydrate had acted on the bromine was shown 

 by an actual experiment, which gave the following results. 



0.67 gr. of the acid yielded by treatment with sodic hydrate so as 

 to form the green salt 0.1136 gr. of argentic bromide, which corre- 

 sponds to 0.0483 gr. of bromine instead of the 0,1'757 gr. contained in 

 the amount of acid taken ; that is, 27.5 per cent of the whole had been 

 removed. 



The yellowish brown solution obtained, when a large excess of sodic 

 hydrate was used, or the mixture of sodic hydrate and metabromdi- 

 nitrophenylacetic acid was boiled for some time, gave on addition of 

 an acid a viscous red substance, which was exceedingly unmanageable. 

 That bromine was removed in quantity in this way is shown by the 

 following determination. 



0.2412 gr. of the metabromdinitrophenylacetic acid was boiled with 

 sodic hydrate for three quarters of an hour; the product, after acidify- 

 ing with nitric acid, and filtering out the precipitate, gave 0.1180 gr. 

 of argentic bromide, corresponding to 0.0502 gr. of bromine instead of 

 the 0.0632 gr. contained in the weight of acid taken ; that is, 79.4 per 

 cent of the whole. 



It is evident, therefore, that the original compound has undergone a 

 deep-seated alteration, and the new crimson substance is probably a 

 phenol. This view of its nature receives some confirmation from the 

 fact that it is decomposed with the greatest ease by even somewhat 

 dilute nitric acid. We have not yet succeeded in bringing either of 

 the products derived from sodic hydrate into a state fit for analysis, 

 although the slight solubility of the barium salt of the crimson acid 

 awakened hopes of success ; but these ended in disappointment, as the 

 salt proved quite as viscous as the free acid. 



Potassic carbonate gives with the metabromdinitrophenylacetic 

 acid results similar to those obtained with sodic hydrate, but less 

 marked. 



We hoped to be able to throw some light on this action of alkalies 

 on the metabromdinitrophenylacetic acid by the study of the correspond- 

 ing anilido compound, but have not succeeded as yet in preparing it, 

 as aniline converts the free acid direct into metanilidodinitrotoluol,* 

 C,3H2CIl3(CeH5NH)(N02)2, melting point 142°, and the saponification 

 of anilidodinitrophenylmalonic ester with sulphuric acid yielded only 

 uninviting viscous products. 



* Hepp, Ann. Cheni., ccxv. 371. 



