This review presents a summary of most current beliefs about the origins of granitic magmas, together with calculations of some important properties of granitic melts and magmas. Turning attention to magma ascent mechanisms, diapirism and dyking represent two competing end-member models. The differences in processes and products between the two are fundamental, and the implications of the differences are profound for the evolution of the continental crust.
The weight of evidence favours dyke ascent for the majority of granitic magmas, and modelling shows that dyke ascent is both rapid and thermally effcient. Furthermore, the structural and petrological features expected in crust through which granitic diapirs have passed have not been discovered in nature. Granites in the upper crust commonly have forms and structures indicating that diapirism could not have been the emplacement mechanism. In contrast, granitic dykes are common, suggesting that it is reasonable to consider the ascent of granitic magmas through fracture systems. Many granitic batholiths are tabular and some have been fed by large dykes. The rate-limit in granitic plutonism is not magma ascent, but either the rate of melting and segregation (near the magma source) or the rate of deformation (at the emplacement site)