A N:o 23) Influence of electrolytes on diffusion. 3 



surements therefore had to be repeated. The total number 

 of experiments was about 130. 



The quantities of sugar and glyceriiie in the four layers 

 of the diffusing pillar were determined with the aid of a 

 P u 1 f r i c h-Z e i s s refractometer, a differential-prism being 

 employed, and the same method of calculation was followed 

 as in some earlier investigations of the diffusion of organic 

 substances in different solvents ^). For deluted solutions, 

 besides, the L ö w e-Z e i s s water-interferometer was used. 

 Assuming the analytic proceeding to be known, I give no 

 account of it here. 



The apparatus of diffusion were the same as those used 

 at previous determinations of this kind. Likewise, the way 

 of calculation and the rest of the procedure have before 

 been described in detail 2). 



The experiments were carried out in 1915 in the Labo- 

 ratory for Applied Physics of the University of Helsingfors. 

 The temperature of the room was about 19° C, and remained 

 fairly constant during the time of each experiment, which 

 lasted 1 to 2 days. Sometimes certain variations appeard 

 owung to unfavourable external conditions. These variations, 

 however, did not exceed 1° C during 2 days. 



In the following tables only the results already calculated 

 are stated. The experiments proper, which would have 

 comprised a 23 pages, have been left aut. kf is the diffusion 

 coefficient, experimentally obtained at the temperature /, 

 and A^2o° the coefficient calculated to 20° C. The coefficients 

 of temperature, necessary for this calculation, have been 

 taken from a table before published ^). 



') Meddel. fr. K. Vetenskapsakademiens Nobelinstitut Bd 2 N:o 24 och 

 N:o 26. 1913. 



'^) Zeitschr. f. phys. Cheniie 50, 309 (1904). 



') Meddel. fr. K. V^etenskapsakademiens Nobelinstitut. Bd 2 N:o 23 pag. 

 25. 1912. 



