Have you results for isotopic composition of Sr,Nd,Pb?They could be useful for explanation the anomalous result for Nd.

Basaltic magmas may produce some minor amounts of silicic rocks without any contribution from crustal sources (e.g., plagiogranites of ophiolitic complexes or LMIs). Commonly, such silicic rocks form small bodies closely related to various gabbros. Do you have any field description of the sampling site and its vicinity?

Is the peridotite altered?  Is there evidence for fluid flux through the mafic and ultramafic sections that might have affected the silicic rock?  The rock is a gneiss and therefore deformed.  Was the deformation coincident with that of the peridotite?

                             Steve

The high Na2O contents suggest that you may have a K2O/Na2O ratio < 1, and you may have a trondhjemite or granodiorite. They can be formed by crystal fractionation from a mantle-derived magma, or remelting of a mantle-derived basalt or gabbro. In the second scenario, if the time elapsed between the formation of the basalt/gabbro, and the remelting event was short, then the newly formed magma will still have an isotopic signature similar to that of the mantle from which the basalt/gabbro originated. But like Petr suggested, you may indeed have a plagiogranite (or 'ocean ridge granite' in some of the Pearce et al. classifications).

Thanks for your immense contributions. To answer some of the questions posed, the Sr isotopic composition of the rock : 87Sr/86Sr=0.704737. The rock was sampled at the contact between the host gness basement and the peridotite massif. The contact rocks are as deformed as the host gneiss basement (concordant foliations). No evidence of fluid activity was observed, but that cannot be ruled out in such a geodynamic setting (subduction zone).

Does the studied contact gneiss resemble other gneisses of the basement (petrography, chemistry, isotopes), or it is distinct from them? Are the peridotites compositionally uniform or intercalated with gabbros, other granitic rocks etc.?

Thanks Petr! The rock forming the contact is mineralogically composed of plagioclase, hornblende and some accessory hypersthene and it as intensely deformed as the gneiss basement. The peridotites are associated with pyroxenites (fine bands 1-5cm thick, sharp contact with the peridotites), gabbros (thick bands of 10-80m concordant to the peridotite beds) and amphibolites dykes.

Thanks Uriah. In my opinion, the crucial point is whether the studied sample is similar to the gneisses of a surely crustal origin (which compose the basement you mentioned before), or it strongly differs from any crustal rock of this area. The comparison with the average composition of the crust would hardly provide a reliable result because of the chemical variability of crustal complexes.