168 



J. G. STEPHENS. 



C<r 



the space within the apparatus was 

 evacuated by connecting D to a 

 filter pump. When working the 

 volatile solvents, such as ether, it 

 was found advisable to place some 

 shredded platinum foil in the tubes 

 B and 0, to prevent the bumping 

 which would otherwise occur. 



As regards the choice of a sol- 

 vent, volatility is essential for rapid 

 determinations, but the exact state 

 of purity of the solvent does not 

 seem to be of great importance. 

 Both the solutes should be much 

 less volatile than the solvent 

 employed, and must not undergo 

 B association in solution. 



The whole apparatus was allowed 

 tostand forabout forty-eight hours, 

 after which time the tubes A and B 

 were weighed, and the ratio of the 

 weights of liquid in the tubes deter- 

 mined. This procedure was con- 

 tinued until the ratio of liquid in the 

 tubes became constant, when the 

 Fig. 1. molecular weight was calculated 



from the formula developed above. 



As an example of the method, the following case is given 

 where benzene was used as the solvent and azobenzene 

 was the substance of known molecular weight: — 

 C Weight of unknown substance 1*3586 

 bolutes (Weight of azobenzene 1*0343 

 Weight of benzene in A after 48 hours 8*0512 

 Weight of benzene in B after 48 hours 16*3907 

 Molecular weight calculated, (after 48 hours) = 117*0 



