Experimental, theoretical and natural studies of layered mafic intrusions have revealed that processes at or near the advancing front of solidification of magmas play important roles in the evolution of layered intrusions. Here we show diapiric structures of anorthositic layers that developed near the basis in a layered intrusion, Murotomisaki Gabbro, Japan. The sill-like intrusion is approximately 220 m thick. Based on the mode of occurrence of the anorthositic layers and the whole-rock major-element compositional data of the gabbros and the anorthosites, we proposed a hypothesis that segregation and ascent of anorthositic crystal mush played an important role during magmatic differentiation. To verify this hypothesis, we studied in detail mineral chemistry and compositional zoning patterns of plagioclase. Systematics between the plagioclase zoning pattern, rock texture and rock types constrain the timing and mechanism of the formation of the anorthositic layers.
The anorthositic layers and gabbroic pegmatite pods with anorthosite roofs occur in olivine gabbro between 40 and 100m from the bottom of the sill and are aligned nearly parallel to the sill. The anorthositic layers (including the roofs of the pegmatite pods) typically have wavy structures, some of which grade into diapiric structures.
Zoning pattern of the plagioclase and mineral composition in both the anorthositic layers and the host rocks suggest: (1) Anorthositic layers were originally mixtures of 20-30 vol% calcic plagioclase crystals and 70-80 vol% melt, the anorthositic crystal mushes or crystalline magmas. These mushes are proposed to have been fluidal and mobile. (2) The calcic plagioclase crystals, now frozen as calcic cores of plagioclase in the anorthosites, may have been derived from the surrounding host crystal mushes, now solidified as olivine gabbros, when only olivine and plagioclase were liquidus phases. This inference suggests that the segregation of the anorthositic crystal mush occurred at relatively high temperatures just below the liquidus, most probably at or close to the crystallization front (i.e., the boundary layer) of the solidifying magmas. The morphology of the plagioclase crystals suggests that the melt in the anorthositic crystal mush was not in equilibrium with the calcic plagioclase and thus was probably highly fractionated and became hydrous. Such disequilibrium melts might have derived from lower horizons below the boundary layer in which crystallization differentiation is more advanced. The anorthositic mushy layers thus developed in the boundary layers may easily evolve into plumes and eventually ascend diapirically towards unfractionated magmas above.