The Kurobegawa granitic pluton is a Pliocene batholith exposed over 700–2900 m elevation in the northern Hida Mountain Range of central Japan. The pluton is vertically zoned from the lower granite (a medium-grained equigranular to porphyritic granite, 70–74 wt% SiO2) to the upper granite (a fine-grained porphyritic granite, 72–77 wt% SiO2). Field relationships of the pluton and its textural, mineralogical, and geochemical features indicate that the pluton represents a well-exposed felsic magma chamber that was zoned as a result of fractional crystallization. Densely concentrated mafic microgranular enclaves (MMEs, 0.1–2 m major axis) of basaltic to dacitic composition (54–68 wt% SiO2) occur throughout the pluton except most of the lowermost part. The MMEs generally have sharp, fine-grained, chilled margins and show no in situ mixing with the host granite except for in the northern area where the two magmas locally mixed and mingled to form hybrid rocks. Field relationships reveal that mafic magma intruded into the chamber from the bottom and ascended as a restricted feeder dike through the lower crystal mush. The mafic magma eventually fragmented to form MMEs, and the MMEs rose upward, resulting in their dense distribution throughout the pluton.
Mafic silicate phases of the MMEs are amphibole and biotite without anhydrous mafic silicates, and the association magnetite-titanite-quartz occurs extensively, indicating that the mafic magma was originally hydrous and relatively oxidized, characteristic of subduction-related backarc basalt. We infer that the MMEs underwent vapor exsolution by second boiling on cooling during their ascent, reducing their bulk density relative to that of the host felsic magma. Relatively H2O-rich MMEs, indicated by the larger modal content of biotite compared to that of amphibole, are abundant in the shallower level of the Kurobegawa pluton. To achieve the MMEs' dense distribution in the upper granite, we interpret that their flotation was sustained because of bubbles trapped in the segregated, interstitial melt by the MMEs' chilled crustal margins. Although no evidence for settling, such as load-cast, silicic pipe, and tightly packed structures, is observed, the bubbles may have eventually been lost, with the result that the MMEs became denser and eventually settled onto the then-current top of the crystal mush.