Diffusion of Liquids. 



325 



solution, as well as with those of Graham on NaCI and other 

 solutions. Other experiments on the influence of concentra- 

 tion were also tried. Three KC1 solutions, ft, 2n, and 3 ft, 

 were prepared and diffused one after the other from the same 



tube. I titrated with j^ AgN0 3 as usual, with the following 

 results : — 



40 



n i 



solution. 



2n solution. 



3ft solution. 







Cub. cent. 



Cub. cent. 



Cub. cent. 



Time. 





a A ^ NO = 



Time. ^ A g N ° 3 



Time. £AgN0 3 







per hour. 



per hour. 



per hour. 



12-3 





1-80 



5- 8 A.M. 3-60 



6-12 A.M. 5-83 



3-5 





1-75 



8-10 3*65 



12- 5 5-70 



5-8 



A.M 



1-66 



10- 6 3-51 



5- 9 a.m. 5-60 



8-6 





1-93 



6- 9 a.m. 3-53 





6-9 



A.M 



1-73 



9- 1 3-62 





9-2 





1-90 



1- 3 3-55 





The results obtained from the n and 2ft solutions agree 

 very well with those obtained from the two NaCl solutions. 

 In both cases the amount diffused from the 2ft solution is 

 slightly less than twice the amount from the n solution. But 

 the results obtained from the 3 ft solution are somewhat ano- 

 malous and require an explanation. This will lead me to 

 speak of the chief sources of error in connexion with the 

 method. 



In some of the earlier experiments of the investigation, 

 solutions of the greatest possible concentration were used, in 

 order to make the determinations of the diffused salt easy 

 because of its greater quantity. But it was soon noticed that 

 in such cases this amount was much larger than was expected. 

 Five solutions of NaCl for instance, n, 2 ft, 3 ft, 4 ft, and 5 ft, 

 diffused in the same time quantities of salt proportional to 1, 

 2-1, 3*8, 6-4, and 8. 



Something similar was observed in the case of strong solu- 

 tions of CaCl 2 and SrCl 2 . These unexpected results seemed 

 to point to the correctness of the opinion expressed by Beil- 

 stein*, that from different solutions of the same substance the 

 amounts diffused increase at a greater rate than the concen- 

 trations. About this time, however, in the hope of clearing 

 up the matter, another little experiment was tried, the results 

 of which were so important for the understanding of the whole 

 method that I will describe it fully here. To the water in the 

 supply-bottle a small amount of indigo solution was added. 

 The blue liquid flowing down through the tube a r s t displaced 



* Loc. cit. 



