Chap, xxii.] DIFFUSION OF LIQUIDS. 249 



The phenomena and laws of diffusion were studied 

 at great length and with much care by Graham. 

 Graham employed in his experiments what he terms 

 a diffusion cell. It consisted of a 4-ounce phial, the 

 mouth and bottom of which were ground flat. This 

 phial was filled up to the base of the neck with the 

 solution for diffusion. The bottle was then placed in 

 a cylindrical jar with a flat bottom ; and when in the 

 jar it was filled up to the mouth with distilled water 

 in such a way as to prevent mixing of the water and 

 the solution by movement. Distilled water was then 

 placed in the jar till it stood one inch above the 

 mouth of the phial. By this means the saline solution 

 communicated freely with the distilled water. After 

 the phial had been allowed to stand undisturbed in 

 the jar for a varying time, its mouth was closed with 

 a plate of glass ; it was then lifted out of the jar, and 

 tests were employed to find how much of the salt had 

 found its way out of the phial into the surrounding 

 distilled water. 



As the result of many experiments, Graham found 

 that the rate of diffusion, the speed, that is, with 

 which the different fluids mixed, varied with the 

 degree of concentration of the solution. If the solu- 

 tion were very concentrated it proceeded fast, if less, 

 more slowly ; and the rate was in direct proportion 

 to the concentration. At first, therefore, the diffusion 

 from the cell would proceed with a certain degree of 

 rapidity. But as the salt diffused, the concentration 

 would be diminished, and would be, besides, no longer 

 into distilled water but into water plus the quantity 

 of salt already diffused. As a result the rate would 

 constantly decrease, and when the liquid outside of 

 the diffusion cell had gained so much salt as to be 

 nearly of the same density as the liquid in the cell, 

 the rate would be very slow indeed, though the 



/ 



diffusion would not cease till both solutions were of 



