The famous theory of the thermal diffusion column derived by Furry, Jones, and Onsager is based on the implicit premise that the mass of a component in a binary mixture is given by the product between the mass of the whole mixture and the molar concentration of that component. The premise is strictly correct only if m(2)/m(1) = 1, where m, and m(2) are the molecular masses of the two components. Here a corrected theory is derived, which is valid for any value of the ratio m(2)/m(1). Comparison of the original and corrected theory shows that for the case of a column operated at total reflux, the two theories give identical results. However, for productive operations the original Furry-Jones-Onsager theory can produce deviated data. The magnitude of the deviation depends on the ratio m(2)/m(1). It is: a) null for m(2)/m(1) = 1, b) small for heavy-isotopes (up to 0.86% for a (UF6)-U-235-(UF6)-U-238 mixture), c) significant for light isotopes (up to 33% for a He-3-He-4 mixture), and d) very high for m(2)/m(1) >> 1 (absurd results are obtained for m(2)/m(1) = 10).