Introduction

The 1959 translation of Korzhinskii's book Physicochemical Basis of the Analysis of the Paragenesis of Minerals introduced me to the concept of inert and perfectly mobile components in open systems. At that time, O.F. Tuttle and I were studying phase relationships in the systems CaO–CO2–H2O and MgO–CO2–H2O (Wyllie & Tuttle, 1960, Walter et al., 1962, Wyllie, 1962), with the volatile components contained securely inside closed gold capsules and therefore thermodynamically inert. The results were to be applied to carbonatites, igneous rocks through which there is no doubt that volatile components have flowed influentially. In Korzhinskii's book I discovered how to represent the volatile components CO2 and H2O in chemical potential diagrams, applicable to both closed and open systems. The method was also applied to many other systems, including granitic rocks with mineralogy controlled by the chemical potentials of sodium and potassium. Korzhinskii's work has provided the basis for quantitative treatment of metasomatism.

Metasomatic processes, originally studied in connection with crustal rocks, are now believed to be important in the mantle, as well. There is evidence that peridotite nodules brought to the surface in kimberlites or alkali basalts were metasomatized within the mantle before being transported by their igneous hosts (e.g. Boettcher & O'Neill, 1980, Dawson, 1980, pp. 183–5; Harte, 1983), and mantle metasomatism is commonly assumed to explain the observation that many basaltic magmas have trace element and isotope geochemistry that is difficult to explain in terms of partial melting of upper mantle rocks with compositions considered to be normal (Walker, 1983).