Thorne's method for obtaining transport coefficients in a binary rigid-sphere mixture is reexamined. First, a close look is taken at the way in which the point where the Enskog functionsx ij are evaluated is introduced. Second, the calculation of the fluxes in the system and the transport coefficients is given. Thorne's results are found to be correct and independent of the choice of the point where the transfer plane is located. This does not hold true for the diffusion flux. It is shown that a different diffusion force is obtained for each selection and that only those diffusional effects which are of first order in the density are consistent with irreversible thermodynamics.