ON" TM H DIFFUSION* OF LIQUIDS. 3 



of concentrations at two given points. It will be seen that, of these 

 methods the optical and the electrical arc better suited for our purpose, 

 because if we use these methods, the liquid is not disturbed and we 

 can day after day study how the diffusion is progressing, and get a 

 clear insight into the phenomena. Optical methods, however, have 

 one defect in common, as was first pointed out by Stefan. ' His 

 objection was that as the liquid is diffusing, the medium is uii- 

 homogeneous and consequently a ray of light meant to be sent along 

 a given laver horizontally is curved out of that layer and tends toward 

 that side which is optically rare, so that the concentration deduced 

 from that experiment does not correspond to that of the layer in ques- 

 tion. Wiener, (2) however, availed himself of this very fact, and devised 

 an arrangement which could be used in the determination of k and 

 also of the conductivity of heat in liquids. 



The constant of diffusion k is not an absolute constant. It is 

 a function of temperature and increases with it, It is also highly 

 probable that it depends on the concentration. This was noticed by 

 Graham, Wroblewski, Weber and Schuhmeister in ZnS0 4 and NaCl 

 solutions, though their results do not agree with one another. It can 

 be easily shown that when k is variable, the differential equation for 

 », is 



lu _ i / ft lu \ (o) 



It Zx\ Ï1X ' 



where k=f{u). Wiedeburg and Wiener took k as a linear function of». 

 §2. My present communication relates to a new method of study- 

 ing diffusion optically, which is quite free from Stefan's objection and 

 is capable of giving an accurate value of k. I experimented upon Zn- 

 S0 4 and NaCl, and found that the diffusion in these cases agrees with 



(1) Stefan, Sitzungsber. d. k. Akad. d. Wissensch. zu Wien, 78 and 79- 1S78— 1870. 

 ri) Wiener, Wied. Ann. 49, 109, 1893. 



