The calculation of the mass transfer attending lateritic weathering was carried out by Brimhall et al. in 1991 by an approach that assumes immobility of one or another element, especially Zr. This calculation is flawed because: (1) Zr is mobile (as evidenced by etch pits in weathered zircons and by E h -pH diagrams), as are Fe, Al, Ti, and others; (2) homogeneity of parent rocks, also required by that approach, is difficult to prove, and unlikely because of self-organization; (3) lateral fluxes, which must be neglected for the approach to work, are in fact inherent in weathering because of the infiltration instability; (4) the dilations calculated with this approach are inconsistent with long-known petrographie and field evidence of constant-bulk-volume evolution; (5) that evidence has been mischaracterized as assumption; and (6) the driving force hypothesized to cause the calculated expansion of some saprolitic laterites is contrary to reason. Better calculations of mass transfer can be carried out on the basis of constant bulk volume framework which is effected in every tropical saprolitic profile by partial pseudomorphic replacement of parent mineral grains by oxides, oxyhydroxides, or kaolinite. Petrographie evidence of widespread replacement in laterites is conclusive. The partial replacement of each parent grain is driven by congruent dissolution of parts of other parent grains slightly above. Bulk-volume-preserving pseudomorphic replacement is produced by the grain/grain stress generated by crystal growth in a rigid rock. This stress requires the growth rate of the replacing grain and the dissolution rate of the replaced grain to be always equal to each other. Replacement (also widespread in diagenesis, metamorphism, and metasomatism) has significant consequences and implications for the modelling of mass transfer in weathering.