LAUSR

Abstract The complementary roles of granites and rocks of the granulite facies have long been a key issue in models of the evolution of the continental crust. “Dehydration melting”, or fluid-absent melting of a lower crust containing H2O only in the small amounts present in biotite and amphibole, has raised problems of excessively high temperatures and restricted amounts of granite production, factors seemingly incapable of explaining voluminous bodies of granite like the Archean Closepet Granite of South India. The existence of incipient granulite-facies metamorphism (charnockite formation) and closely associated migmatization (melting) in 2.5 Ga-old gneisses in a quarry exposure in southern India and elsewhere, with structural, chemical and mineral-inclusion evidence of fluid action, has encouraged a wetter approach, in consideration of aqueous fluids for rock melting which maintain sufficiently low H2O activity for granulite-facies metamorphism. Existing experimental data at elevated T and P are sufficient to demonstrate that, at mid-crust pressures of 0.5–0.6 GPa and metamorphic temperatures above 700 °C, ascending immiscible CO2-rich and concentrated alkali chloride aqueous fluids in equilibrium with charnockitic (orthopyroxene-bearing) gneiss will inevitably begin to melt granitic rocks. The experimental data show that H2O activity is much higher (0.5–0.6) than previously portrayed for beginning granulite facies metamorphism (0.15–0.3). Possibilities for metasomatism of the deep crust are greatly enhanced over the ultra-dry models traditionally espoused. Streaming of ultrasaline fluids through continental crust could be a mechanism for the generation of the discrete mid-crust layer of migmatites suggested to characterize younger tectonometamorphic regions. The action of CO2-rich and hypersaline fluids in Late Archean metamorphism and magmatism could record the beginning of large-scale subduction of volatile-rich surficial materials.