2 L. W. Öholm. (LVIII 



dif fusion must be greater in the said salt solution than in 

 water. In a LiC/-solution the reverse must take place. Sugar, 

 for instance, must diffuse more slowly in a LtC/-water 

 solution than in pure water. 



With a view to ascertaining experimentally the circum- 

 stances here indicated, the following investigation was 

 carried out, the object of which thus was to establish the 

 influence of electrolytes upon the diffusion of non-electro- 

 lytes. For the experiments some of the substances were 

 chosen which were used at the investigation before mentio- 

 ned, in which the electrolytes KCl and LiCl were allowed 

 to diffuse in water solutions containing sevéral non-electro- 

 lytes, as urea, cane sugar, glycerine, starch, etc. At the follo- 

 wing experiments therefore the solvents consisted of KCl- 

 and LiC/-solutions, the diffusing substances, again, were 

 sugar and glycerine. 



In order to get a clear idea of the dependence of the 

 diffusion on the substancial c|uantities, not only of the diffu- 

 sing bodies but also of the electrolytes added, experiments 

 were carried out using several solutions, which were very 

 differently concentrated. All solutions are expressed in 

 normalities. The concentrations of KCl and LiCl were 

 2-, 1- and 0,25-/7. These, as already mentioned, not only 

 formed the diffusion liquids but also the solvents for sugar 

 and glycerine. In order to obtain the bottom-layer in the 

 apparate of diffusion, of the said non-electrolytes so much 

 was added to the electrolyte solutions, that these with 

 regard to them resulted in the following normalities: for 

 sugar 2-, 1-, 0,25-, and 0,i-, and for glycerine 5-, 2-, 1-, 0,25- 

 and 0,1-/7. 



The experiments were so combined that when, for 

 instance, 2-/7 KCl formed the solvent, 2-, 1-, 0,25 and 0,i-n 

 sugar were used as diffusing substance. Thereafter the same 

 experiment was carried out using 1-/7 KCl as solvent, and 

 so forth. Each trial, for the sake of control, was made with 

 two apparatus simultaneously. Using much diluted solutions 

 (0,25-, 0,1-/7), the results proved uncertain. The diffusion 

 coefficients considerably varied, and a great many mea- 



