OF ARTS AND SCIENCES. 5 



the radical CJLO, which in this case would be added instead of 

 hydroxyl, would make a very great difference in the percentage com- 

 position of the salt; but this preparation, like the others, gave on 

 analysis numbers corresponding to the formula 



C H 2 Br(NO 2 ) 2 CNa(COOC 2 H 5 ) 2 . 



Secondly, the ease with which the sodium salt is formed, and with 

 which it is decomposed by an acid giving the original substance, mili- 

 tates against the first formula, as we can hardly suppose that a ring of 

 carbon atoms, such as must be assumed in this formula, would break, 

 when treated with an alkali, and reunite under the influence of an acid 

 in dilute solutions. Finally, we may add that the meta position of the 

 bromine atoms replaced deprives the first formula of any support from 

 analogy.* The principal objection to the formula 



C G H 2 Br(NO : ,) 2 CH(COOC 2 H ;i ) 2 , 



which we adopt, is that we cannot account as yet for the formation 

 of such a substance by the reaction of sodium malonic ester on tri- 

 bromdinitrobenzol. We hope, however, that a careful study of the 

 quantitative conditions of the reaction, and of the oil which is the 

 secondary product, will throw light on this part of the subject ; but 

 owing to the departure of one of us from Cambridge, its further 

 investigation must be postponed till next autumn. 



The yield was far from satisfactory, 14 grm. of tribromdinitroben- 

 zol after treatment with 12 grm. of malonic ester and 1.7 grm. of 

 sodium, as described above, giving only 4.3 grm. of the bromdinitro- 

 phenybnalonic ester instead of the 14 grm. required, if the whole of 

 the tribromdinitrobenzol was converted into the new substance, that 

 is, 30.7 per cent of the theoretical yield. A large amount of the tri- 

 bromdinitrobenzol was recovered, however, amounting to 5.5 grm., or 

 39.3 per cent. This leaves 30 per cent unaccounted for, which prob- 

 ably remains dissolved in the oil deposited by the alcoholic mother 

 liquors ; but that the dissolved substance is bromdinitrophenylmalonic 

 ester admits of great doubt, as we have not succeeded in obtaining any 

 considerable quantity of this substance from the oil by treatment with 

 aqueous sodic hydrate. This point will be investigated more thor- 

 oughly hereafter. We have tried several modifications of the process, 

 in hopes of improving the yield, but none of them have given satisfac- 

 tory results ; if, for instance, the proportion of sodium malonic ester is 



* Compare J. Stanley Kipping, Ber. d. ch. G. 1888, p. 32. 



