Diffusion of Liquids.—The diffusion of dissolved salts may obviously be considered under this heading. The phenomenon may be conveniently studied by introducing a strong solution of some highly-coloured salt, such as bichromate of potash, into the bottom of a tall glass cylinder nearly filled with water. The rate of diffusion varies with the nature of the liquids. Graham was the first to investigate the subject carefully. He filled a number of similar glass vessels with solutions of different salts. The mouths of these vessels were carefully ground so that they could be closed by means of glass plates. The different vessels were then placed in equal glass jars, and covered with water to a definite extent. Next the glass covers were cautiously withdrawn, and the diffusion was allowed to go on for a certain time. The rate at which each liquid diffused was thus obtained. Graham found that, for any one solution, the rate is proportional to the gradient of concentration—i.e. to the rate at which the quantity of salt dissolved per unit-volume varies per unit-length. Thus the law regulating diffusion of liquids is analogous to that which regulates the conduction of heat in a homogeneous solid. Hence the equations obtained by Fourier in his Théorie de la Chaleur apply to the problem under consideration.
Graham found also that rise of temperature greatly increases the effect. He divided substances into two classes, Colloids and Crystalloids, the members of the first class diffusing very much more slowly than those of the second. His investigations have been much extended by more recent observers employing various methods of observation.
If two miscible liquids be separated by a membrane of bladder or of parchment paper, &c., diffusion takes place through the septum at rates which are usually very different for different liquids. This phenomenon is known as Osmose (q.v.). It was first shown by Nollet that, if a vessel filled with alcohol be closed by a piece of bladder and placed in water, the diffusion of the water is so much more rapid than that of the alcohol that the bladder is burst because of the increase of the contents of the vessel which it closes. By this means the various constituents of a mixture of colloid and crystalloid substances may be separated to any desired extent. The rate at which liquids diffuse into each other through a septum depends greatly upon the molecular action between them and the septum.