Chinese Acad Sci, Inst Geol & Geophys, Beijing 100029, Peoples R China
; Nanjing Univ, Dept Earth Sci, Nanjing 210093, Peoples R China
; Chinese Acad Sci, Inst High Energy Phys, Beijing 100039, Peoples R China
; Tianjin Inst Geol & Mineral Resources, Tianjin 300170, Peoples R China
It has been well accepted that basaltic magmas are derived from the mantle and most granites are generated by fractional crystallization of basaltic magmas. The evidence for the fractional crystallization of basaltic magmas and even andesitic magmas is there is corresponding cumulate, however, cumulate is rarely associated with Si-rich granitic magmas, therefore, it is unlikely that fractional crystallization happened in granitic magmas. This is probably because (1) granitic magmas have higher viscosity such that the crystallization of minerals (i.e., plagioclase) have been obstructed to form euhedral crystals and high density minerals (i.e., hornblende) have been blocked to settle down; and (2) primary minerals (e.g., plagioclase) have similar density to that of granitic magmas. We argue that plagioclase is unlikely fractionally crystallized from granitic magmas although the process was reported in numerous literature on the basis of the Harker diagrams. We argue that the Harker diagrams are more suitable for basalts rather than granites because granites are totally different from basalts in magma source and composition. We also argue that the Bowen reaction series, the principal of the Harker diagrams, is unlikely happened to form a continuous sequence from basalt through andesite and dacite to rhyolite. The unit and super-unit mapping method of granites is not suitable for regional geological mapping. Mesozoic granites in East China and Tanncherfi granites in Morocco are taken as two examples to indicate the fractional crystallization of granitic magmas is impossible. Variable compositions of granites mainly result from the composition of magma source, temperature, pressure, volatile composition, degrees of partial melting, mixing and differentiation process. Among them the composition of magma sources is the most crucial to the diverse compositions of granite whereas fractional crystallization of granitic magmas has rarely effect on the compositional variation of granites. Therefore, the role of fractional crystallization in the evolution of granite should be appreciated in a limited scale rather than a large scale.