If only one substance S passes through a membrane, the nature of this membrane is not of importance with respect to the direction of the diffusion; this is namely determined only by the O.S.A. of the two liquids.
If, however, more substances pass through a membrane, the nature of this membrane is of great importance.
If n substances diffuse through a membrane, we can distinguish 2n cases, when we take into consideration only the direction in which each of these substances passes through the membrane; if we call each of these cases a D.T. (diffusion-type), 2n D.T.'s may be conceived. Now we can deduce: one of these D.T.'s is not possible, the other 2n x 1 D.T.'s are thermodynamically admissible. The composition of the two liquids determines which of the D.T.'s is not possible; the nature of the membrane determines which of the 2n x 1 other D.T.'s will occur.