Geochemical constraints on the composition of Archaean lower continental crust: Partial melting in the Lewisian granulites

New geochemical data for the Lewisian granulites of NW Scotland show that the protolith to the neoArchaean Scourie granulites was a tonalite containing a mafic hydrous phase, most probably hornblende. This observation when combined with recent thermobarometric calculations and new partial melting studies on the Lewisian makes it very probable that the Lewisian granulites experienced fluid-absent melting. Partial melting calculations based upon a proxy for the unmelted granulite protolith demonstrate that Scourie granulite protolith was depleted in both LILE (Rb, Th, U) and HFSE (Ta) relative to unmelted amphibolite facies Lewisian gneisses of the same age. Both granulite facies and amphibolite facies gneisses are members of the tonalite-trondhjemite-granidiorite (TTG) magmatic suite. Nb/Ta ratios in the TTG granulite protolith are suprachondritic and can be attributed to the partial melting of a basaltic source with a rutile eclogite residue, whereas the amphibolite facies TTG gneisses possess Nb/Ta ratios which are subchondritic and can be attributed to the partial melting of a basaltic source with a garnet amphibolite residue. However, these differences do not account for the differences in LILE between the two suites implying that in addition the two basaltic precursors were different. It is proposed that the depleted character of the tonalitic protolith to the Scourie granulites was inherited from ultra-depleted basalt with very low concentrations of Rb, Th, U and Ta. Felsic melts formed during partial melting are no longer present in the granulite terrane and were probably removed to the upper crust, now removed by erosion. Thus partial melting of the Lewisian granulites contributed to the process of crustal fractionation. The process of fluid-absent melting in the Lewisian took place ca 200 Ma after crust formation. This long time interval indicates that the stabilisation and differentiation of the crust was probably a separate event from that of crust formation.