But the fact of chief interest to us here, is that the relatively small-moleculed crystalloids have immensely greater diffusive power than the relatively large-moleculed colloids. Among the crystalloids themselves there are marked differences of diffusibility; and among the colloids themselves there are parallel differences, though less marked ones. But these differences are small compared with that between the diffusibility of the crystalloids as a class, and the diffusibility of the colloids as a class. Hydrochloric acid is seven times as diffusible as sulphate of magnesia; but it is fifty times as diffusible as albumen, and a hundred times as diffusible as caramel. These differences of diffusibility manifest themselves with nearly equal distinctness, when a permeable septum is placed between the solution and the water. The result is that when a solution contains substances of different diffusibilities, the process of dialysis, as Professor Graham calls it, becomes a means of separating the mixed substances: especially when such mixed substances are partly crystalloids and partly colloids. The bearing of this fact on the interpretation of organic processes will be obvious. Still more obvious will its bearing be, on joining with it the remarkable fact that while crystalloids can diffuse themselves through colloids nearly as rapidly as through water, colloids can scarcely diffuse themselves at all through other colloids. From a mass of jelly containing salt, into an adjoining mass of jelly containing no salt, the salt spread more in eight days than it spread through water in seven days; while the spread of "caramel through the jelly appeared scarcely to have begun after eight days had elapsed." So that we must regard the colloidal compounds of which organisms are built, as having, by their physical nature, the ability to separate colloids from crystalloids, and to let the crystalloids pass through them with scarcely any resistance.