Erratum to: Melts of garnet lherzolite: experiments, models and comparison to melts of pyroxenite and carbonated lherzolite

Abstract

An unfortunate coding error led to erroneous predicted abundances of Na[subscript 2]O in the melting models presented in Table 5 of Grove et al. (2013). The source of this error has been discovered and corrected; the cause of the discrepancy was an error in assignment of the partition coefficient expressions used for orthopyroxene and clinopyroxene. The values presented in a corrected Table 5 are the ones obtained using the appropriate partition coefficients for Na[subscript 2]O. Our models of garnet lherzolite melts now have Na[subscript 2]O abundances that are similar to the melts predicted by Longhi (2002) and the experimentally produced melt of Walter (1998).