OF ARTS AND SCIENCES. 281 



compounds are attacked by alcoholates and other reagents with the 

 removal of nitro groups. This publication brought to our notice an 

 earlier paper* by him, in which he states that by the action of 

 alcoholic potassic cyanide on metadinitrobenzol one nitro group is 

 replaced by the ethoxy (or methoxy) radical, but at the same time 

 an atom of cyanogen is substituted for one atom of hydrogen, giv- 

 ing as the product C 6 H 3 N0 2 OC 2 H 5 C]S\ This curious result seems 

 to have more affinity with the additions of hydrocyanic acid to nitro- 

 halogen benzols studied by v. Riehter f than to simple substitu- 

 tion; but nevertheless we have here the replacement of a meta nitro 

 group by another radical. As Lobry de Bruyn's later work was 

 published almost simultaneously with ours (although a little later), 

 and had grown out of his earlier researches, we felt that the study 

 of substitutions of nitro groups in simple trinitro compounds 

 should properly belong to him, and after a pleasant correspondence 

 on the subject it has been agreed that he should follow out this 

 line of work, while we confine ourselves, so far as the removal of 

 nitro groups is concerned, to the completion of our study of tri- 

 bromtrinitrobenzol. 



At present, therefore, we know only three meta nitro compounds 

 from which nitro groups have been removed, dinitrobenzol, sym- 

 metrical trinitrobenzol, and symmetrical tribromtrinitrobenzol. 



Lobry de Bruyn's work shows that the removal of nitro groups 

 observed by us is caused by the position of these nitro groups, but 

 the presence of the three meta bromine atoms evidently also has a 

 loosening effect on the nitro groups, as we have replaced two of 

 these groups by ethoxy radicals, whereas in the trinitrobenzol only 

 one was replaced. 



In taking up the study of the subject again, we first examined 

 the action of tribromtrinitrobenzol and sodic ethylate more care- 

 fully, to see whether the bromine atoms were not also affected by 

 it, and found that in addition to the tribromnitroresorcine diethyl- 

 ether there is always formed a quantity of trinitrophloroglucine 

 triethylether, C 6 (OC 2 H 5 ) 3 (N0 2 ) s , melting at 119-120° ; as however 

 this substance is easily decomposed by sodic hydrate, the amount 

 of it obtained as such is small, unless special precautions are taken 

 to exclude atmospheric moisture. When such precautions are not 

 taken, the trinitrophloroglucine triethylether is converted into the 



* Rec. Trav. Ch., II. 205. 



f Ber. d. ch. Ges., IV. 465, VII. 1145, VIII. 1418. 



