This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial Share Alike License (http://creativecommons.org/licenses/by-nc-sa/3.0), which permits unrestricted non-commercial use, distribution and reproduction in any medium provided that the original work is properly cited and all further distributions of the work or adaptation are subject to the same Creative Commons License terms.

Abstract

Magmatic epidote and zoisite commonly occur in Cretaceous tonalite of the Hazu area in the Ryoke belt, central Japan. The tonalite is mainly composed of amphibole, biotite, plagioclase, quartz, and epidote/zoisite with minor ilmenite, magnetite, pyrite, zircon, and apatite. Small amounts of K-feldspar occur as an interstitial phase between other felsic phases or perthitic patches in plagioclase. Epidote occurs as inclusions in plagioclase, as interstitial phase in the matrix, and as secondary phase in chlorite pseudomorphs after biotite, and in saussuritized plagioclase. The XFe [= Fe3+/(Al + Fe3+)] value of the secondary epidote ranges from 0.27 to 0.39. Epidote inclusions in plagioclase and interstitial grains contain less Fe3+ (XFe = 0.08–0.29), Fe3+-poor epidote with XFe < 0.18 occurs only as inclusion. Zoisite with XFe value of 0.01–0.07 occurs only as inclusions in plagioclase, and usually has thin lamella-like layers of Fe3+-poor epidote with XFe = 0.09–0.14. The Fe3+-poor epidote with XFe < 0.20 and zoisite included in plagioclase occasionally form aggregates with K-feldspar and quartz. A thin sodic plagioclase zone develops at the boundary between the Fe3+-poor epidote and zoisite inclusions and their host plagioclase. Such a reaction texture is not observed at the boundary between Fe3+-richer epidote inclusions with XFe > 0.20 and their host plagioclase. Epidote grains with XFe > 0.20 in plagioclase and the matrix are a magmatic phase that crystallized directly from the tonalite magma. The Fe3+-poor epidote (XFe < 0.20) and zoisite were probably formed by a local reaction between the trapped melt and its host plagioclase, and these are considered not to have been in equilibrium with the tonalite magma. Compositions of amphibole-plagioclase assemblages allowed for temperature estimates in the range of 730–770 °C and minimum pressures of 0.47–0.57 GPa for the epidote/zoisite-bearing tonalites of the Hazu area. Epidote/zoisite-free tonalites occur in other areas of the Ryoke belt. There may be several tonalite bodies that record different intrusion processes and solidification depths in the Ryoke belt.

1. Introduction

The petrological significance of magmatic epidote was experimentally and petrographically demonstrated in the 1980s (Schmidt & Poli, 2004). Naney (1983) showed that epidote can coexist with a melt phase at 0.8 GPa in granitic and granodioritic systems, and suggested that the presence of magmatic epidote is almost-certain evidence for high-pressure crystallization of silicate magma. Zen & Hammarstrom (1984) identified epidote as an important magmatic constituent of tonalite and granodiorite within the mobile belt extending from northern California to southern Alaska, and suggested that epidote indicates a minimum intrusive pressure of about 0.5–0.6 GPa. Evans & Vance (1987) reported elongate phenocrysts of epidote, which texturally ensures the magmatic origin, from rhyodacitic dikes in Colorado, and considered that the epidote-bearing dike magma was fed from a magma chamber at a depth corresponding to 0.8–1.3 GPa. Subsequently, magmatic epidote was also described from monzogranite and diorite, as reviewed by Schmidt & Poli (2004); Schmidt & Thompson (1996) experimentally showed that the epidote-in curve is positioned on the slightly higher temperature side for tonalite compositions than for granodiorite compositions, and it shifts towards the lower pressure (P)/temperature (T) side with increasing oxygen fugacity. Furthermore, zoisite was reported from high-P migmatites and pegmatites derived from eclogites (Nicollet et al., 1979; Franz & Smelik, 1995). These petrographical and experimental studies clearly suggest that epidote-group minerals are an index phase that implies relatively high-P solidification of a silicic melt, and the stability conditions for epidote-group minerals depend slightly on the melt composition and increase towards the low P/T side with an increase in their XFe [= Fe3+/(Al + Fe3+)] values.

Magmatic zoisite and clinozoisite-epidote (referred to as epidote hereafter) commonly occur in a tonalitic pluton, a member of the Ryoke belt, from the Hazu area, central Japan. The purpose of this paper is to document the mineral chemistry and petrological characteristics of the zoisite/epidote-bearing samples and to interpret their depths of crystallization based on the stabilities of zoisite/epidote and pressure estimations using the hornblende geobarometer.

2. Outline of Geology

The Ryoke belt, which consists of low-P/T metamorphic rocks originating from the Mesozoic accretionary complexes and Cretaceous granitoid plutons, stretches throughout the Inner Zone (the Japan Sea side) of southwest Japan over a length of roughly 600 km (Fig. 1). The south of the Ryoke belt is bounded by the Sanbagawa belt, which is a high-P/T subduction metamorphic belt of Cretaceous age. These two belts form perhaps the best-known example of paired metamorphic belts (Miyashiro, 1961). The boundary between the Ryoke and Sanbagawa belts is a major strike-slip fault, the Median Tectonic Line (MTL). Members of the Ryoke belt are distributed widely in central Japan, in which the Hazu area studied in this paper is located.

The geological configuration of the Ryoke belt in central Japan is summarized in Suzuki & Adachi (1998), including the distributions of the Ryoke metamorphic rocks and a series of granitoid plutons. The Ryoke metamorphic rocks are mainly composed of pelitic, psammitic, and siliceous lithologies, with minor amounts of metabasite. The metamorphic grade generally increases from the chlorite–biotite zone in the northwest, through the K-feldspar–cordierite zone, to the sillimanite–K-feldspar/garnet–cordierite zones in the southeast, and then locally decreases towards the MTL (Ikeda, 1998; Miyazaki, 2010). The low-grade part of the Ryoke metamorphic rocks passes into the unmetamorphosed Jurassic accretionary complex of the Mino terrane. Chemical U-Th Total-Pb Isochron Method (CHIME) monazite ages from 98.0 ± 3.2 to 100.7 ± 3.2 Ma were reported as peak metamorphic ages of the Ryoke metamorphism (Suzuki et al., 1996a, 1996; Suzuki & Adachi, 1998). Granitoids in the Ryoke belt of central Japan are categorized into fifteen plutons from the oldest Kamihara tonalite (94.5 ± 3.1 – 94.9 ± 4.9 Ma CHIME monazite age: Nakai & Suzuki, 1996) to the youngest Naegi granite (67.2 ± 3.2 – 68.3 ± 1.8 Ma: Suzuki et al., 1994b; Suzuki & Adachi, 1998), based on their intrusive relations (Ryoke Research Group, 1972) and radiometric ages (Suzuki & Adachi, 1998). The Kamihara tonalites defined by the Ryoke Research Group (1972) occur as three major bodies in the Hazu, Shimoyama, and Tenryu areas (e.g., Suzuki & Adachi, 1998).

The Hazu area, from which the tonalite samples were collected, is situated at the southwestern margin of the Ryoke belt in central Japan (Fig. 1a). In the northern part of the Hazu area, the Ryoke metamorphic rocks are distributed widely, whereas tonalite occurs along the southern coastline (Fig. 1b). The Ryoke metamorphic rocks in the Hazu area are divided into sillimanite and sillimanite–K-feldspar zones in the northern and southern parts, respectively (Asami, 1977). The typical mineral assemblage of metapelite in the sillimanite zone is biotite + muscovite + sillimanite + plagioclase + quartz. Andalusite occurs in the lower-grade part of the sillimanite zone. Staurolite is reported to occur sporadically in this zone (Asami, 1977). Metapelite in the sillimanite–K-feldspar zone is mainly composed of biotite, muscovite, sillimanite, K-feldspar, plagioclase, and quartz. The modal amount of muscovite in the sillimanite–K-feldspar zone is lower than that in the sillimanite zone and tends to decrease with distance from the sillimanite–K-feldspar isograd, and thus, in the higher-grade part of the sillimanite–K-feldspar zone, metapelite sometimes lacks primary muscovite (Asami, 1977; Asami et al., 1982). Garnet occurs sporadically throughout the two mineral zones. Asami (1977) and Asami et al. (1982) considered that the metamorphic pressure condition of the Hazu area was slightly higher than that of the rest of the Ryoke area, based on the occurrence of staurolite in the Hazu area and the difference in temperature at which aluminum silicate + K-feldspar becomes stable. The tonalite in the Hazu area intrudes into metasedimentary rocks of the sillimanite–K-feldspar zone.

3. Petrography

Most tonalite samples in the Hazu area are composed of primarily biotite, calcic amphibole, plagioclase, and quartz, with subordinate amounts of K-feldspar, titanite, ilmenite, magnetite, pyrite, zircon, and apatite (Table 1). Biotite exhibits a dark-brownish to brownish Z-axial colour. Calcic amphibole is pale brownish green to green in colour. Plagioclase usually occurs as subhedral forms and exhibits albite and pericline twins. Epidote occurs as inclusions in plagioclase (Fig. 2a–c), interstitial phases in the matrix, as secondary phases after biotite (Fig. 2d), and in saussuritized plagioclase. Zoisite is observed only as inclusions in plagioclase, and usually shows twin textures with narrow epidote layers (Fig. 2a and b), similar to that reported for a high-P pegmatite in the Münchberg Massif, Germany (Franz & Smelik, 1995). Zoisite usually forms an aggregate with quartz, K-feldspar, and epidote (Fig. 2a and b). The host plagioclase around the zoisite inclusions is locally modified in composition, and a thin zone with less calcic plagioclase characteristically develops between the inclusions and the host phase (Fig. 2a). Epidote in plagioclase also occurs as a single grain. A less calcic plagioclase zone is usually observed also around epidote inclusions with XFe< 0.2. On the other hand, there is no obvious compositional modification of plagioclase around most of the epidote inclusions with XFe> 0.2 (Fig. 2c). Although epidote/zoisite is not observed as inclusions in biotite and amphibole, these three phases coexist as inclusions in a plagioclase grain. K-feldspar occurs interstitially in the matrix or as exsolution lamellae in plagioclase. Secondary chlorite and titanite replace biotite.

4. Whole-rock chemical composition

The whole-rock major and trace element compositions (Table 2) were determined by X-ray fluorescence spectrometry (XRF) using a Shimadzu XRF-1800 spectrometer at Kwansei Gakuin University. The rhodium-target X-ray tube was energized at 40 kV and a current of 70 mA for the major-element analysis. Details and discussion of the XRF analytical methods used in this study can be found in Morishita & Suzuki (1993) and Nakazaki et al. (2004).

The “Kamihara tonalite” reported in the literature, which includes tonalite samples from the Hazu, Shimoyama, and Tenryu areas, shows variable SiO2 contents (57.8–69.9 wt%) and modified alkali-lime index (MALI) values (−2.79–4.55), and belongs to the calc-alkalic and calcic series of Frost et al. (2001) (Fig. 3a). With increasing SiO2 content, the aluminum-saturation index (ASI) (e.g., Chappell & White, 1974) of the “Kamihara tonalite” increases from 0.82 to 1.22, suggesting wide variations in the whole-rock composition from metaluminous to peraluminous groups (Fig. 3b).

The studied Hazu samples are less silicic (SiO2 = 53.6–60.2 wt%) and have lower MALI (−1.50–0.44) and ASI (0.89–1.06) values than most other tonalites reported in the literature. The Hazu samples, however, share similar compositional trends with the other “Kamihara tonalite” on the MALI–SiO2 and ASI–SiO2 diagrams, and mostly belong to the calc-alkalic series metaluminous group.

5. Mineral chemistry

Chemical analyses of major constituent minerals were carried out using an electron probe microanalyzer (EPMA) with wavelength- and energy-dispersive X-ray spectrometer systems (JXA-8900R) at Nagoya University. The accelerating voltage and beam current were kept at 15 kV and 12 nA on the Faraday cup, respectively. A beam diameter of 5 μm was used for analyses of biotite and feldspars, and 2–3 μm for analyses of all other phases. Well-characterized natural and synthetic phases, including synthetic REEP5O14, were used as standards (REE: rare-earth elements). Detection limits of La2O3, Ce2O3 and Nd2O3 were 0.02 wt% for the 1σ level. The factors calculated by Kato (2005) were employed for the matrix correction. Total iron of zoisite/epidote was treated as Fe2O3. Amphibole compositions for the mineral descriptions are given using an average of the maximum and minimum Fe3+/Fe2+ estimates (Leake et al., 1997). Total iron of the other phases is assumed to be FeO. Selected analyses of zoisite/epidote, amphibole, and biotite are listed in Tables 3, 4, and 5, respectively.

5.1. Zoisite–epidote

Most epidote grains consist of a REE-poor solid solution (total REE2O3 < 0.3 wt%), except for a slightly REE-enriched core (up to 0.79, 1.67, and 0.59 wt% of La2O3, Ce2O3, and Nd2O3, respectively) of zoned epidote included in plagioclase in sample YM2906. The total REE2O3 content of zoisite is less than 0.05 wt%. Chemical compositions of zoisite/epidote are variable, and are closely related to their modes of occurrence (Fig. 4). Zoisite grains occurring as inclusions in plagioclase have XFe values ranging from 0.01 to 0.07, whereas epidote included in plagioclase has XFe = 0.08–0.29. Matrix epidote grains tend to be richer in Fe2O3, with XFe varying from 0.18 to 0.29. Secondary epidote is distinctly enriched in Fe2O3 (0.27 < XFe ≤ 0.39).

5.2. Amphibole

Most amphibole crystals exhibit a relatively homogeneous core and a thin Al- and K-richer rim. Al-richer amphibole similar to the rim also develops along the cleavage (Fig. 5a, c, and d). They have magnesiohornblende/hornblende–tschermakite/ferrotschermakite–pargasite/ferropargasite compositions (Fig. 6). Some grains are retrogressively rimmed by Al-poor ferrohornblende/ferro-actinolite. The homogeneous core of amphibole in epidote-bearing tonalites has the following average compositions: Si = 6.40–6.58 per formula unit (pfu), Al = 1.70–1.95 pfu, Ca = 1.78–1.90 pfu, [A](K + Na) = 0.7–0.48 pfu, and mg# [= Mg/(Mg + Fe2+)] = 0.42–0.54, where [A](K + Na) indicates alkaline contents at the 10-coordinated A-site (Fig. 6a–c). The Al-rich rim has average compositions of Si = 6.17–6.40 pfu, Al = 2.01–2.35 pfu, Ca = 1.90–1.94 pfu, [A](K + Na) = 0.50–0.55 pfu, and mg# = 0.38–0.52. Amphibole inclusions in plagioclase are homogeneous and have similar compositions to those of the core of the matrix phases: Si = 6.40–6.58 pfu, Al = 1.70–2.12 pfu, Ca = 1.83–1.95 pfu, [A](K + Na) = 0.40–0.47 pfu, and mg# = 0.40–0.55.

Amphibole crystals in epidote-free tonalites have a zonal structure similar to those in the epidote-bearing tonalites, but with a less aluminous core (Fig. 6d): Si = 6.54–6.71 pfu, Al = 1.54–1.77 pfu, Ca = 1.82–1.84 pfu, [A](K + Na) = 0.38–0.49 pfu, and mg# = 0.46–0.52 in the core; and Si = 6.37 pfu, Al = 2.06 pfu, Ca = 1.91 pfu, [A](K + Na) = 0.51 pfu, and mg# = 0.48 in the rim. The composition of the inclusion amphibole in the epidote-free tonalite is Si = 6.57 pfu, Al = 1.86 pfu, Ca = 1.88 pfu, [A](K + Na) = 0.44 pfu, and mg# = 0.48.

5.3. Biotite

Biotite in epidote-bearing tonalites is relatively homogeneous, with crystals in the matrix (Si = 2.73–2.79 pfu and mg# = 0.38–0.49) similar to those enclosed by plagioclase (Si = 2.75–2.79 pfu and mg# = 0.38–0.49). Biotite in epidote-free tonalites has similar Si contents (2.77–2.80 pfu in the matrix, 2.76 pfu in plagioclase) and mg# values (0.44–0.46 in the matrix, 0.45 in plagioclase) to those in epidote-bearing tonalites. TiO2 and BaO contents are up to 4.1 wt% and 1.2 wt% in the epidote-bearing samples and up to 4.0 wt% and 0.4 wt% in the epidote-free samples, respectively.

5.4. Feldspars

Plagioclase crystals are slightly saussuritized, and those that survived saussuritization show normal zoning, with decreasing anorthite content from the core towards the margin (Fig. 5b). The average anorthite content of the unsaussuritized calcic core is An42(2)–47(3) in epidote-bearing tonalites and An44(2)–48(4) in epidote-free tonalites (value in parenthesis is the 1σ-standard deviation). K-feldspar grains are generally homogeneous and have orthoclase contents of 92–96% and 93–94% in epidote-bearing and epidote-free tonalites, respectively. The standard deviation is less than 2% on the 1σ level. BaO contents are up to 2.5 and 0.8 wt% in epidote-bearing and epidote-free tonalites, respectively.

6. Pressure-temperature estimates

The amphibole–plagioclase thermometry of Holland & Blundy (1994) as defined by the following two equations was applied to estimate the temperature of tonalite solidification in the Hazu area:

 

edenite+quartz=tremolite+albite

and

 

edenite+albite=richterite+anorthite

Pressure conditions were calculated using the empirical Al-in hornblende igneous geobarometers calibrated by Hammarstrom & Zen (1986); Hollister et al. (1987); Johnson & Rutherford (1989); Blundy & Holland (1990); Schmidt (1992), and Anderson & Smith (1995).

As the amphibole is chemically heterogeneous, with an Al-richer zone as rim and along cracks and/or cleavages (Fig. 5), the core composition was employed for the P–T estimates of the solidification stage. The combination of the amphibole-bearing geothermobarometers gives P-T conditions of 0.47–0.57 GPa/730–770 °C for the epidote-bearing tonalites. On the other hand, amphibole in the epidote-free tonalites has a slightly Al-poorer composition (Al = 1.54–1.77 pfu) than in the epidote-bearing tonalites (1.70–1.95 pfu), implying slightly lower P conditions of 0.37–0.48 GPa (Fig. 7).

7. Discussion

Epidote-group minerals in the tonalites from the Hazu area occur as three types: as inclusions in plagioclase, as interstitial phase in the matrix, and as pseudomorph after biotite. The pseudomorph epidote characteristically has a high XFe (0.27–0.39). On the other hand, the inclusion and interstitial epidote grains are less ferric, and all zoisite and epidote grains with XFe < 0.18 occur as the inclusions only (Figs. 4 and 8). These textural and compositional characteristics suggest that zoisite and epidote included in plagioclase and most epidote grains occurring as interstitial phases represent primary products and are probably of magmatic origin, and that the oxygen fugacity and/or Fe2O3/Al2O3 value of magma have systematically increased during crystallization.

The stability of magmatic epidote is controlled by various factors, including melt composition, oxygen fugacity, and P–T conditions during crystallization (Schmidt & Poli, 2004). The epidote-group minerals are generally a solid solution of Ca2Al3Si3O12(OH) and Ca2Al2Fe3+Si3O12(OH), with a molar Al2O3/CaO value ranging from 0.5 to 0.75. Thus, magmatic epidote is probably stabilized more easily in metaluminous melts than in peraluminous melts. However, the “Kamihara tonalite” has a wide distribution of whole-rock compositions ranging from metaluminous to peraluminous groups, and some epidote-free tonalite samples have distinctly more metaluminous compositions than the epidote-bearing Hazu tonalites (Fig. 3). Thus, the occurrence of magmatic epidote in the tonalites of the Hazu area is not simply a result of the whole-rock composition of their host rocks, and is probably controlled by the oxygen fugacity and/or P–T conditions during crystallization of the melt.

The stability relations of magmatic epidote in the tonalite composition (Schmidt & Thompson, 1996) show that the minimum P stability of the epidote + melt assemblage (1) is 0.5 GPa and 0.3 GPa at 680 °C for NNO- and QFM-buffered oxygen fugacities, respectively, and (2) increases with increasing temperature. Schmidt (1993) conducted systematic experiments on the relationship between the composition and stability of epidote in the tonalite composition and showed that the lower-P stability limit of the epidote + melt assemblage increases with decreasing XFe value of epidote at constant T, and exceeds 1.0 GPa for Fe3+-poor epidote with XFe < 0.12–0.13 (Fig. 8a). The fields of possible coexistence of zoisite + quartz + melt in the K2O–CaO–Al2O3–SiO2–H2O and K2O–CaO–MgO–Al2O3–SiO2–H2O systems were explored experimentally by Schliestedt & Johannes (1984) and Hoschek (1990), respectively, showing that this assemblage is more stable at higher P than at 0.7 GPa.

The P conditions of the Hazu tonalites, 0.47–0.57 GPa for the epidote-bearing samples and 0.37–0.48 GPa for the epidote-free samples, were estimated using Al-in hornblende geobarometer. This geobarometer was suggested to use for granitoid intrusions containing a limiting assemblage of quartz, K-feldspar, plagioclase, biotite, and amphibole (Hammarstrom & Zen, 1986; Hollister et al., 1987), and thus, Al content of amphibole is probably controlled by the following reactions:

 

phlogopite+anorthite+quartz=tschermakite+K-feldspar
( Hollister et al., 1987)

and

 

tremolite+phlogopite+anorthite+albite=pargasite+K-feldspar+quartz
( Mäder & Berman, 1992).

The Hazu tonalites, however, contain K-feldspar only as minor and interstitial phases in the matrix and as exsolved phase in plagioclase. The estimated T conditions (730–770 °C) are far higher than the K-feldspar–in temperature of tonalitic composition (about 680 °C at 0.5 GPa: Schmidt & Thompson, 1996). Thus, the K-feldspar was probably a late-stage product during solidification of the Hazu tonalites, and was not stable during the major stage of crystallization of amphibole and plagioclase. The Al-control reactions described above have gentle and positive ΔPT slopes, and the K-feldspar–bearing right-hand sides are stable at higher P. Therefore, the estimated P conditions for the epidote-bearing tonalites (0.47–0.57 GPa) probably indicate a lower P limit. The amphibole and biotite employed for the P–T estimations coexist with epidote as inclusions in plagioclase, and thus epidote grains were crystallized at a pressure ≥ 0.47–0.57 GPa.

The magmatic epidote-group minerals in tonalite from the Hazu area show a wide compositional range from zoisite to epidote with an XFe value of 0.29. However, most epidote grains have Fe3+-rich compositions with XFe > 0.24. The occurrence of such Fe3+-rich magmatic epidote is not inconsistent with the lower P limit of magmatic epidote-bearing tonalite formation (0.47–0.57 GPa). These Fe3+-rich magmatic epidote grains occur both as interstitial phases in the matrix and as inclusions in plagioclase, and the host plagioclase around the inclusions shows no textural evidence suggesting a reaction relation between the two phases (Fig. 2c). Therefore, these Fe3+-rich epidote grains were probably in equilibrium with plagioclase and crystallized directly from tonalitic magma. On the other hand, the presence of Fe3+-poor epidote with XFe< 0.20 and zoisite implies an unusual high-P condition above 0.7 GPa, if they were in equilibrium with the tonalitic melt. However, the Ryoke metamorphic country rocks of the tonalite are considered to have recrystallized under sillimanite-stable conditions of 0.21–0.76 GPa at 700 °C (Fig. 7), and thus, such high-P estimates are not expected for the melt solidification. These Fe3+-poor epidote and zoisite grains all occur as inclusions in plagioclase, unlike the Fe3+-rich epidote. Furthermore, the zoisite and Fe3+-poor epidote included in plagioclase sometimes form aggregates with K-feldspar and other phases, and a thin zone of sodic plagioclase (An15–19) characteristically develops between the Fe3+-poor epidote and zoisite inclusions and their host plagioclase (Fig. 2a). Zoisite usually forms a lamellar texture with thin layers of Fe3+-poor epidote (Fig. 2b), implying an exsolution phenomenon during the cooling stage due to the miscibility gap between the orthorhombic and monoclinic phases (Enami & Banno, 1980; Prunier & Hewitt, 1985; Brunsmann et al., 2002). These textural characteristics most likely indicate that local re-equilibration between the inclusion and host phases and that modifications of chemical compositions and species of minerals occurred after the formation of the inclusion–host textural relation. The most probable interpretation of these textural characteristics, thus, is that (1) the inclusion aggregate was formed by a local reaction between a melt inclusion and its host plagioclase, (2) the Fe3+-poor epidote and zoisite were not in equilibrium with the tonalitic melt, and (3) their formation was controlled by local characteristics of the chemical system.

Sial et al. (1999) particularly investigated modes of occurrence of magmatic epidote in granitoids from five Neoproterozoic tectonostratigraphic terranes in northeastern Brazil, Early Palaeozoic calc-alkalic granitoids in northwestern Argentina, and from three batholiths in Chile. They documented that the magmatic epidote occurs only as inclusions in biotite and K-feldspar, and concluded that the inclusion phase survived, whereas the matrix crystals were probably dissolved by the host melt. In the Hazu tonalites, magmatic epidote grains mostly occur as inclusions in plagioclase, but not in biotite and amphibole. However, individual plagioclase grains in many cases include biotite and amphibole, suggesting that these two inclusion phases coexisted with epidote, plagioclase, and melt in equilibrium. Thus, a combination of the inclusion and matrix phases probably suggests that crystallization of the magmatic epidote has started at a relatively early stage of solidification and continued until near-solidus stage. Brasilino et al. (2011), studying magmatic epidote-bearing calc-alkalic granodiorites to monzogranites from northeastern Brazil, concluded that oxidation of host magma stabilized and preserved magmatic epidote. The systematic increase of oxidation state of the host magma during crystallization, which is revealed by systematic increase of XFe in epidote from inclusion to matrix crystals, probably stabilized the magmatic epidote in the matrix of the Hazu tonalites.

Amphibole-bearing tonalites occur as intrusive bodies in the Ryoke metamorphic rocks of the Shimoyama and Tenryu areas (e.g., Ryoke Research Group, 1972; Suzuki & Adachi, 1998). The Shimoyama and Tenryu tonalites show wide ranges of whole-rock compositions, from metalumious to peraluminous, with some of them having similar compositions to those of the Hazu tonalites. Magmatic epidote, however, has not even been reported as either inclusions in plagioclase or in the matrix of tonalite in the Shimoyama and Tenryu areas (e.g., Tsuboi & Asahara, 2009). We cannot completely deny the possibility that epidote in the Shimoyama and Tenryu tonalites was crystallized only as an interstitial phase, before being dissolved by the host magma with decreasing P during emplacement. However, the Al contents of amphibole in the Shimoyama and Tenryu tonalites (usually Al <1.4 pfu: Enami et al., unpublished data) are distinctly lower than those in the Hazu tonalites. Thus, the Shimoyama and Tenryu tonalites possibly solidified under epidote-unstable P-T conditions, which are probably lower P and/or P/T conditions than for the Hazu tonalites. In other words, the tonalite magma in the Shimoyama and Tenryu areas was probably intruded into the Ryoke metamorphic rocks and solidified at a shallower level than in the Hazu area. The tonalites in the Hazu, Shimoyama, and Tenryu areas have been considered to share the same parent magma and intrusive and solidification processes, and therefore are collectively called “Kamihara tonalite” (e.g., Ryoke Research Group, 1972). The occurrence of epidote-group minerals in “Kamihara tonalite” possibly indicates that tonalites in the Hazu area underwent intrusion processes that were distinct from those in the Shimoyama and Tenryu areas.

Acknowledgements

The authors are deeply indebted to A.N. Sial, R. Gieré and an anonymous reviewer for their careful reading and constructive suggestions, which led to significant improvements in this manuscript. This research was partially supported by grants from the Japan Society for Promotion of Science Nos. 14540448, 18340172, 19654080, and 25400511 (ME), and 19540510 (MT).