Felsic rocks are minor in abundance but occur ubiquitously in International Ocean Discovery Program Hole U1473A, Southwest Indian Ridge. The trace element abundances of high-Ti brown amphibole, plagioclase, and zircon in veins, as well as the presence of myrmekitic texture in the studied felsic rocks support crystallization origin from highly-evolved melts, probably controlled by fractional crystallization. Based on geochemical criteria and texture of the mineral assemblage in felsic rocks and their relationship with host gabbros, they can be divided into three types: (1) Felsic rock with sharp boundaries is formed when felsic melt intrudes into fractures of host gabbros, resulting in minimal interaction between the melt and the wall minerals. (2) Replacive felsic rock, which is characterized by a pseudomorphic replacement of minerals in the host gabbro. This vein type is caused by the replacement of the host mineralogy by minerals in equilibrium with the felsic melts. (3) Felsic rock with diffused boundaries is formed either by infiltration of felsic melt into the solidifying gabbro body or crystallization of interstitial melts. Infiltration modes of felsic melts are likely controlled by the temperature condition of the cooling host gabbros.

Amphibole plays a pivotal role in mediating the flux of volatiles and partial melting that ultimately contribute to arc magmatism. The influence of amphibole from the lower crust to the upper mantle remains unclear due to limited opportunities for observation. Amphibole-rich ultramafic rock characterized by large poikilitic hornblende grains with olivine and pyroxenes occurs in the Nishidohira metamorphic rocks in the southern Abukuma Mountains of Northeast Japan (we call poikilitic hornblende pyroxenite hereafter). Amphibole exhibit zoning in color and chemical composition: the dark core has higher TiO ₂ and Al ₂ O ₃ contents than the light green rims. Dark-colored high-TiO ₂ pargasitic amphibole formed early from magmatic melts. Melt compositions calculated from the dark-colored amphibole core based on melt-mineral partitioning indicate that the poikilitic hornblende pyroxenite resulted from adakitic magmatic activity. Reactions between pre-existing ultramafic rock and adakitic melt are likely to form poikilitic hornblende pyroxenite when the melt/rock ratio is low, and hornblende gabbro when the ratio is high. The U-Pb zircon age of approximately 120 Ma for poikilitic hornblende pyroxenite and associated hornblende gabbro is interpreted as a magmatic age. In the Early Cretaceous tectonic framework of Northeast Japan, adakitic magmatism is attributed to the westward subduction of the Izanagi (or Kula) plate beneath the eastern margin of the Eurasian Plate.

Ultramafic rocks, i.e., peridotites and pyroxenites, occur in a variety of tectonic settings on Earth. Ultramafic rocks can form as accumulation of mafic minerals from basaltic to komatiitic melts and be a major component of the Earth's mantle. The origin and history of ultramafic rocks are expected to provide information on the processes of partial melting and melt migration/extraction in the mantle and on the tectonic evolution of geologic units containing ultramafic rocks. I study ultramafic rocks in metamorphic belts, ocean floor, and mantle sections of ophiolites. My career began with a study of the Horoman Peridotite Complex in the Hidaka metamorphic belt in Japan. The ultramafic rocks and associated mafic rocks in the Horoman body record a very complex evolutionary history from the mantle conditions to crustal conditions. It is difficult to constrain the tectonic setting affecting events in the Horoman Peridotite Complex. On the other hand, ultramafic rocks in the mantle section of ophiolites and abyssal peridotites directly recovered from ocean floor to study melting processes and melt-rock interactions in the mantle can be used to constrain their tectonic setting, or at least as analogs to these tectonic settings. Studies on the Oman ophiolite by Japanese groups and literature studies of other ophiolites suggest that many ophiolites are later modified by subduction-related magmatism. Several ophiolites are being studied to elucidate the maturing process by subduction-related magmatism. Simple partial melting and melt extraction is expected in the adiabatically upwelling mantle beneath the mid-ocean ridge. In fact, abyssal peridotites directly recovered from mid-ocean ridges provided a unique opportunity to elucidate these processes. Comparison of abyssal peridotites recovered from the mid-ocean ridges and arc regions (fore arc and back arc) is key to understand the differences in magmatic processes in the two regions. Ocean science with research vessels has a well-defined working hypothesis that can only be addressed by direct sampling from the seafloor. To understand a crucial issue in Earth science as to why plate tectonics occurs on Earth, it is essential to elucidate the life of the oceanic lithosphere from its birth to its subduction into the mantle. Direct sampling of oceanic lithosphere by drilling is the key to solving this issue. I would like to emphasize that members of the Japan Association of Mineralogical science can play an essential role in leading analyses of rock samples directly recovered from seafloor. Rock samples recovered from seafloor by drilling and any methodology, as well as samples from anywhere on Earth, should be published in as papers, and these data would help integrate knowledge about the history of the Earth and planet and its future.

Peridotitic, dunitic, and wehrlitic rocks drilled at Atlantis Massif (Mid-Atlantic Ridge), where mantle-derived rocks were exhumed to the sea floor via detachment faulting, were investigated employing in situ geochemical analyses. Since dunitic and wehrlitic rocks are commonly interpreted as fossil melt channels, they likely record transport process of melts generated at depths beneath the ridge axis. Although the rocks are severely serpentinized, primary chromian spinels and clinopyroxenes were found in several samples. The Cr/(Cr + Al) ratios (Cr#) of the chromian spinel show bimodal distribution, where one group records higher Cr# (0.35–0.5) and the other group is represented by lower Cr# (0.2–0.3). The extent of mantle melting is greatly correlated to the Cr# of the chromian spinel, thus the mantle-derived materials of the Atlantis Massif bear at least two end-members with regard to the extent of melting: refractory mantle material with higher Cr#, and enriched mantle material with lower Cr#. To quantitatively evaluate mantle melting and melt transport mechanisms in the suboceanic mantle at the portion of the Mid-Atlantic Ridge, numerical modelling was carried out using rare-earth element concentrations of clinopyroxene. Our modelling results demonstrate that the melting region of the mantle extends unevenly across the ridge axis, with deeper melting occurring on-axis from the garnet-stability field, and shallower melting occurring off-axis from the spinel-stability field. We propose that depleted melts generated at depths in the absence of garnet within the off-axis region were focused to the ridge axis via fracture-induced melt channels and produced wehrlite. The fracture-induced melt migration occurred dominantly due probably to a decrease of temperature when mantle exhumation was facilitated by detachment faulting.