Diffusion of Liquids. 421 



found that the two days allowed, in which the diffusion should 

 reach a condition of constancy, was not a period long enough 

 for these salts. It is also possible that this is the normal 

 action of these substances. 



But before speaking any further of the probable cause of 

 the above discrepancies, it would be well to compare my 

 results with those of previous experimenters. Of all the 

 results obtained by others only those of Graham can be readily 

 compared. Beilstein neglected to give the dimensions of his 

 diffusion-cell with sufficient accuracy to allow a fair compa- 

 rison to be made; and Marignac's experiments were conducted 

 in such a way that a direct comparison is impossible. His 

 results can be considered only with reference to each other; 

 and, besides, he has shown that the relative rates of two salts, 

 when simultaneously diffused, are often quite different from 

 their relative rates when diffused alone. 



But, fortunately, Graham's work was so performed and his 

 results given in such a form that their reduction and compa- 

 rison with my own can be more readily made. Graham's dif- 

 fusion-cells, mentioned at the beginning of this paper, con- 

 sisted of small phials, 9*64 centims. high, holding 134*8 cubic 

 centims. to the base of the neck. The neck itself was 12*7 

 millims. high and slightly conical, being on the upper surface 

 31*49 millims., and on the lower 30*48 millims. in diameter. 

 These phials were filled to the base of the neck with the solu- 

 tion to be examined ; and the neck was filled with pure water 

 in such a manner as not to mix the two liquids. The phials 

 were placed in large vessels containing 567 cubic centims. 

 water, and here allowed to remain a number of days. At the 

 end of this time the water was drawn off and the amount of 

 diffused substance analytically determined. The results of 

 these determinations are given in tables, of which the follow- 

 ing may serve as a sample. It is from Graham's second paper 

 on the subject*: — 



Diffusion of Chloride of Barium in 8*57 days at 63°: two 

 cells. 



grs. Katio. 



From 1-per-cent. solution . . 6*32 1*047 



„ 2 „ „ . . 12*07 2*000 



„ 4 „ „ . . 23*96 3*970 



„ 8 „ „ . . 45*92 7*608 



The temperature is given in degrees Fahr. " The quanti- 

 ties 1, 2, 4, and 8 per cent, indicate the parts of salt in a con- 



* Phil. Trans. 1850, p. 817. 



