Diffusion of Liquids, 319 



Diffusion of NaCl. 



No. of Observed Calculated 



layer. value. value. 



*} 



3-284 3-328 



3 1-527 1-482 



4 1-317 1-290 



5 1-057 1-073 



6 0-850 0-853 



7 0-640 0-648 



8 0-460 0-469 



9 0-318 0-325 



10 0-211 0-215 



11. 0-134 0-135 



12 , 0-081 0-082 



13 0-051 0-047 



14 0-028 0-026 



15 0-017 0-015 



16 0-013 0-011 



A better correspondence than the above could scarcely be 

 expected, considering the number of unavoidable sources of 

 error there are in connexion with the experimental method. 



I have thus far mentioned briefly the chief results of the 

 work done on diffusion. The several investigations may be 

 divided into two classes : — (1) Those which are concerned 

 with the physical side of the question — that is, the determina- 

 tion of the constant k for a single substance : to this class 

 belong the investigations of Fick, Simmler and Wild, Voit, 

 Hoppe-Seyler, Johannisjanz, Weber, and Stefan. (2) Those 

 which are concerned with the chemical phase of the subject — 

 that is, the comparison of many different substances as to their 

 rates of diffusion : to this class belong the investigations of 

 Graham, Beilstein, and Marignac. In regard to the first class, 

 it may be said that there a very satisfactory end has been 

 attained. The proof of Fick's law by Weber and Stefan, and 

 the determination of the influence of temperature and concen- 

 tration of solution by the former, leave little to be desired in 

 connexion with this part of the subject. The same cannot be 

 said, however, of the other — the chemical phase of the ques- 

 tion. The experiments of Beilstein are too few to establish 

 any thing with certainty ; and those of Graham and Marignac, 

 while rather numerous, and agreeing very well among them- 

 selves, seem to establish nothing as regards the dependence of 

 diffusion on the molecular weight or other physical property 



