OF ARTS AND SCIENCES. 255 



II. 0.2197 gr. of the substance gave 31.6 c.c. of nitrogen at a 

 temperature of 25° and a pressure of 755.4 mm. 



The substance gave no test for bromine with a copper wire, or 

 even when treated according to the method of Carius. As the 

 amount of impurity is so small that we thought it could have no 

 effect, we have used this preparation for the determination of the 

 molecular weight of the yellow modification, which, with that of the 

 orange prisms, was made by the method of Eaoult, using benzol as 

 the solvent, since preliminary experiments had shown that these 

 compounds were not sufficiently soluble in glacial acetic acid. 



Yellow Plates. 

 Substance 0.2290 gr. Benzol 10.299 gr. Depression 0°. 27. 



Orange Prisms. 

 Substance 0.2561 gr. Benzol 10.468 gr. Depression 0°.29. 



From these results the following molecular weights are ob- 

 tained: — 



Molecular Weight. 



Yellow Plates 404 



Orange Prisms 413 



Calculated for C 6 H(C 6 H 5 NH) 3 (N0 2 ) 2 441 



There can be no doubt, therefore, that the substances are not 

 polymeric. The benzol solutions obtained in the determinations 

 were mixed with a little alcohol, and allowed to evaporate sponta- 

 neously, (the substance is deposited as a varnish from benzol alone,) 

 when each yielded as the principal product the modification which 

 had been originally dissolved in it, although in each case this was 

 mixed with an insignificant amount of the other form. 



From the observations given above we should infer that these 

 two modifications of the trianilidodinitrobenzol are not true chem- 

 ical isomeres, but physical isomeres, or perhaps rather that the 

 substance is dimorphous. 



We have made a great many experiments to get other isomeric 

 forms of the trianilidodinitrobenzol, both by varying the method 



