Abstract Deciphering the former presence of melt in metacarbonate rocks that experienced high dT/dP gradients is one of the most complex and debated topics in petrology. This is due to the high complexity in identifying and interpreting microstructures indicative of partial melting, as well as to the scarcity of experimental datasets in more complex carbonate-bearing compositional systems. Microstructures that are unequivocally interpreted as evidence of partial melting or melt–residue interaction in other lithologies have often been subject of conflicting interpretation when observed in metacarbonate rocks. From a literature review, it is clear that most microstructures suffering conflicting interpretation are preferentially found in granulite-facies olivine ± spinel-bearing calcite–dolomite and dolomite marbles. Here we present a collection of case studies that either attribute a suprasolidus or a subsolidus evolution of impure marbles sharing apparently indistinguishable microstructures, represented by symplectitic coronas around olivine enclosed in dolomite pockets. We have compared unambiguous anatectic microstructures with the experimental literature, which allowed us to discriminate whether dolomite or calcite melt pockets develop, according to the composition of the protolith. Additionally, through phase equilibrium thermodynamic modelling, we have tested the possibility that the symplectitic coronas around olivine, separated from calcite by the dolomite pockets, develop under subsolidus conditions. We show that this approach fails, and we demonstrate the reason by combining our results with the experimental background in the simplified CaO-MgO(-SiO2)-H2O-CO2 compositional space. Using this integrated approach, we speculate that olivine is a peritectic mineral that focuses the carbonatitic melt around itself, and that the crystallisation of the residual melt provides the H2O and CO2 necessary to stabilise amphibole and dolomite during cooling.
Maroni et al. (Mon,) studied this question.