Are microstructures in plutonic rocks primary or secondary? A re-examination of the metasomatism hypothesis for the roof-sourced autoliths in the Skaergaard Intrusion

The roof-derived autoliths in the floor cumulates of the Skaergaard Intrusion have been argued to have been extensively metasomatised and recrystallised, forming the foundation of the hypothesis that microstructures in plutonic rocks are essentially metamorphic. However, the augite–plagioclase–plagioclase dihedral angles and plagioclase core composition of the autoliths match with those of the roof rocks, demonstrating that they were generally solid on arrival at the floor, with no subsequent microstructural or compositional modification. Many autoliths have mafic rinds, which were used as evidence of metasomatism: these rinds fall into two groups. The rarely developed rind rock of Irvine et al. (1998) is most likely chilled magma infiltrating along fractures in the roof rocks, either associated directly with detachment of roof material, or occurring before final detachment. Thin mafic rims are widespread in LZc and MZ, present at the tops of the more elongate autoliths, with a corresponding felsic rim at the base of the most elongate autoliths. The close correspondence of thin rim development with autolith shape, rather than composition, is argued to be evidence that they formed as a result of differential migration of immiscible conjugate interstitial liquids: the dense Fe-rich liquid flowed downwards and ponded on the tops of impermeable autoliths, whereas its buoyant Si-rich conjugate flowed upwards and was trapped underneath. Any differences in microstructure and bulk composition of the autoliths compared with the remaining exposures of the roof sequence reflect the wider range of lithologies in the now-eroded regions of the roof.