818 PKOCEEDINGS OF THE AMERICAN ACADEMY 



recently made by Garzino * by the action of nitric acid on the pro- 

 pionic ester of metabromphenol. This coincidence is, however, 

 purely accidental, as in Garzino's phenol the bromine atoms stand 

 in the ortho and para, in ours in the meta positions, to the hy- 

 droxyl. As was to be expected, the two phenols gave different 

 salts, the potassium salt made by Garzino, containing one half of 

 a molecule of water, while ours is anhydrous, his barium salt con- 

 taining three, ours two molecules of water. Mixed with the 

 dibromdinitrophenol is an oily phenol, which we have not suc- 

 ceeded in purifying in spite of many attempts. 



The two substances, therefore, act with water and sodic carbonate 

 in the same general manner as with sodic ethylate, the trinitro 

 compound in both cases showing two parallel reactions, in one of 

 which nitro groups, in the other bromine atoms, are removed, while 

 the dinitro compound in both cases loses part of its bromine. It is 

 to be observed, however, that while the alcoholate removes two 

 nitro groups from the trinitro and two bromine atoms from the 

 dinitro compound, water in each of these cases removes only one; 

 although, like the alcoholates, it removes all three of the bromine 

 atoms from the tribromtrinitrobenzol. 



The constitution of the new tribromdinitrophenol can be onl} r 

 C 6 OHBi\N"0 2 BrN0 2 Br, as the substance from which it is derived is 

 symmetrical. The dibromdinitrophenol, on the other hand, can be 

 either C 6 OHN0 2 BrN0 2 BrH or C 6 OHN0 2 BrHBrN0 2 , and we have 

 no experimental data for determining which of these two formulas 

 is correct. 



Action of Water and Sodic Carbonate on Tribromtrinitrobenzol. 



As this action takes place very slowly it is better to carry it on 

 as follows. Four or more flasks provided with return condensers 

 were charged each with about 1 gr. of tribromtrinitrobenzol and a 

 moderate quantity of a dilute solution of sodic carbonate, and al- 

 lowed to boil for twelve or more hours. After a short time the 

 solution turned yellow, and at the end of the boiling had become 

 deep red. The unaltered tribromtrinitrobenzol was filtered out, and 

 boiled again with a fresh solution of sodic carbonate. The deep red 

 filtrate showed the presence of sodic bromide and sodic nitrite when 

 the proper tests were applied. When acidified with dilute sulphu- 



* Att. R. Ace. Sc. Torino, XXV. 263. Ber. d. ch. Ges., 1892, R. 119. 



