S-type granites always contain more Al than the amounts of Na, Ca and K in the rock required to form feldspars, primarily owing to their derivation from source components that had previously been weathered. Those rocks are therefore always saturated in Al, or peraluminous. Many I-type granites are also peraluminous, despite I-type source rocks typically not being saturated in Al. It has previously been suggested that this may result from the fractional crystallisation of amphibole. However, data from compositionally zoned high-temperature plutons in the Lachlan Fold Belt show that it is difficult to generate large quantities of peraluminous melt by removal of amphibole.Most of the I-type granites (~. 95%) in the Lachlan Fold Belt formed at lower temperatures and almost half of those rocks for which bulk chemical compositions are available are peraluminous. Among these granites there are 98 separate suites for which there are chemical data for two or more samples, with 47 suites that include both metaluminous more mafic and peraluminous more felsic compositions. The origin of those peraluminous compositions is fundamental to any understanding of I-type granite petrogenesis in this region. Compositional variations caused by the assimilation or by partial melting of supracrustal rocks are very small as the isotopic variations within these rocks are dominantly between different suites, not within suites. The partial melting of more mafic source rocks, rather than the fractional crystallisation of more mafic magmas, is favoured for the origin of these rocks.Partial melting is the most likely process involved in the petrogenesis of felsic granites where broadly granodioritic-monzogranitic batholiths are associated with lesser amounts of tonalite and very minor amounts of mafic rock. Experimental studies have shown that the melts generated by the partial melting of basaltic to andesitic rocks under crustal conditions are mostly peraluminous. During the dehydrational melting of I-type granite source rocks at pressures below the garnet stability field, biotite and amphibole melt incongruently to yield pyroxenes. The excess Al is incorporated into the felsic liquid, resulting in the generation of peraluminous melts. In this instance, the excess Al in felsic I-type granites is a function of the melting process, and unrelated to the bulk composition of the source. The observed gradation from peraluminous felsic granites to metaluminous compositions in less felsic rocks in largely isotopically closed systems could happen in two ways. At higher temperatures of partial melting, Ca and other components of clinopyroxene could dissolve in the melt, with the melt eventually becoming metaluminous. Alternatively, minerals residual from partial melting, dominantly pyroxenes and plagioclase, could be incorporated in suspension in the melt, so that the resulting bulk magma is metaluminous. Examples of these two extremes, and of intermediate cases, are well developed among the granites of southeastern Australia.
- Partial melting