Wall-rock alteration by a reaction mineral (Korzhinskii, 1970) has been reinvestigated considering the effect of a varying porosity on the process itself. This is because a limit value of porosity at the replacement front has made it possible to analyze the alteration. In diffusion metasomatism, the porosity decreases toward a fissure due to the "salting out" effect. Strongly dependent upon the porosity, the effective diffusion coefficient decreases toward the fissure. Consequently, a steady state of "constant" concentrations of both replacing and replaced components (∂ci/∂x=0) is almost achieved near the replacement front. The "salting-out" also occurs even in combined diffusion and reverse-infiltration metasomatism. This process may occur in nature. In both cases, the alteration by a reaction mineral causes a porosity decrease in the replaced zone, which in turn hinders the process. In other words, this type of metasomatism is a "self-controlled" process.
The formation process of vein-type wollastonite skarn around fissures in limestone, a typical wall-rock alteration, as observed at the Fujigatani mine, SW Japan, has been numerically analyzed based on the thermodynamic assumptions. These consist mainly of constant volume and local equilibrium assumptions, essential in analyzing metasomatic processes. As a result, I have concluded that the porosity is reduced to almost zero near the fissure even when the porosity at a replacement front is extremely large. Hence, the composition of a pore solution of wollastonite zone becomes almost the same as that of the replacement front. This means that the front cannot progress. It should be emphasized that the porosity change acts as an obstacle to the development of the skarn. Therefore, to analyze naturally-occurring metasomatism, we must reexamine the assumptions used since Korzhinskii (1970).
The effect of changing porosity on the metasomatism in a solid solution mineral has also been evaluated. I have proved that only a small difference in the intrinsic diffusion coefficients of both replacing and replaced components brings about a large variation of porosity, resulting in quite different concentration profiles compared to those calculated under the constant porosity assumption.