OP ARTS AND SCIENCES. 289 



acid and allowed to evaporate spontaneously, when after purification 

 it yielded 2 gr. of the triethylether of trinitrophloroglucine in- 

 stead of 0.12 gr., the usual yield when no special precautions were 

 taken to exclude atmospheric moisture. It is evident, therefore, 

 that the trinitrophloroglucine triethylether is formed at first, and 

 afterward saponified wholly or in part by sodic hydrate produced by 

 the action of moisture on the sodic ethylate. That most of this 

 sodic hydrate is formed during the spontaneous evaporation of the 

 alkaline solution was shown by two experiments, in which the tri- 

 ethylether of trinitrophloroglucine was treated with a solution of 

 sodic ethylate, with all the precautions described above, except that 

 in one of these experiments the alkaline liquid was allowed to evap- 

 orate spontaneously, when all the triethylether was saponified, the 

 whole product being soluble in water, whereas in the other, which 

 was acidified before evaporating, 0.65 gr. of unaltered triethylether 

 was recovered, 0.7 gr. having been the amount used. 



Lobry de Bruyn* has encountered similar difficulties in studying 

 the action of sodic etlrylate upon the unsymmetrical trinitrobenzol 

 (1, 2, 4), as in this case considerable quantities of dinitrophenol 

 were obtained, whereas with sodic methylate the dinitroanisol was 

 the only organic product. 



Experiments with Acetic Ester as the Solvent. 



After we had found that the presence of benzol diminished the 

 tendency of sodic ethylate to replace nitro groups in the tribrom- 

 trinitrobenzol, we tried some experiments on the action of other 

 solvents. The choice here however was limited, because there are 

 so few liquids belonging to a different class from benzol which 

 have any considerable solvent action on the tribromtrinitrobenzol ; 

 in fact, acetone and acetic ester were the only ones easily obtained 

 which fulfilled this condition, and acetone had to be rejected, as 

 it seemed to take part in the reaction, giving a dark brownish red 

 solution, very different from the orange-red obtained in all other 

 cases. This result was not unexpected after the work of Freer, f 

 published only in the preliminary paper at the time we tried this 

 experiment, and since we did this work we have seen the paper of 

 Janovsky, J describing the colors which he obtained by the action of 



* Rec. Trav. Chim. des Pays-Bas, IX. 191. 



t Am. Chem. Journ., XII. 355; XIII. 308, Freer and Higley ; XIII. 322. 



| Ber. d. ch. Ges., 1891, p. 971. 



VOL. XXVII. (ft. S. XIX.) 19 



