Literature deahng with multicomponent mass transfer presents two distinctly different relationships defining diffusion fluxes at a phase boundary in multicomponent systems. The relationships, in matrix form, are based on the film theory. The first relationship results from the solution of continuity equations according to the so-called linearized theory of multicomponent mass transfer. The second relationship, due to Krishna and Standart, results from the exact solution of the Maxwell-Stefan equations. Principal difference between these two solutions lies in the manner, in which the influence of finite mass transfer rates on the "zero flux" mass transfer coefficients is accounted for. According to the first "linearized approach" this influence depends solely on the net total mixture flux, in the second "exact" approach this influence depends explicitly on individual molar fluxes of each component. Theoretical analysis presented in this paper shows that the influence of finite mass transfer rates on diffusion fluxes can he attriuted only to the net total mixture flux N, and not to the individual fluxes NT_ The correction factor matrix obtained here is identical with that resulting from the linearized model. The difference between the relationships describing diffusion fluxes based on the linearized model and those based on the exact solution is accounted for by an additional matrix. This matrix accounts for changes in multicomponent diffusion coefficients along the diffusion path in the film.