LAUSR

Abstract There is a large compositional variation of I-type granites. Although partial melting and fractional crystallization are the two primary mechanisms for the variation, it is intriguing to know how they jointly work. This issue is addressed by a combined study of geochemistry and phase equilibrium modelling for a Triassic pluton from the Qinling orogen. The pluton consists of high-K, calc-alkaline I-type granites. Geochemical results suggest that the granites are likely sourced from Triassic mafic-intermediate igneous rocks. Compared with experimental melts, it is inferred that the natural granites are not the crystallized product from pure magmatic melts. This is confirmed by phase equilibrium calculations that have utilized the Triassic mafic-intermediate igneous rocks as the source, yielding modelled melts that contain too low FeOt, MgO and Mg# to match the target granites. The addition of restitic mineral assemblages to melts can relieve the discrepancy, but the modelled magmas have much higher CaO than the granites. Replacing equilibrium melting by disequilibrium melting in terms of the plagioclase behavior can lead to a satisfactory match. Furthermore, by taking the average of the target granites as a parental magma, the modelled crystalline products can well reproduce the variation trends in the high-K, calc-alkaline I-type granites from the Qinling orogen. In this regard, the crustal anatexis would produce the parental magma at first, which experienced fractional crystallization to form the granitic pluton with the compositional variation as observed. Therefore, these two processes would jointly lead to the compositional variation of I-type granites in nature.