<<

MINERALOGICAL MAGAZINE, DECEMBER I980, VOL. 43, PP. IOO5-I3

Zoisite- relations in low- to medium-grade high-pressure metamorphic rocks and their implications

MASAKI ENAMI Department of Earth Sciences, Nagoya University, Nagoya 464, Japan

AND

SHOHEI BANNO Department of Earth Sciences, Kanazawa University, Kanazawa 920, Japan

SUMMARY. Coexisting and clinozoisite in seven- electron-probe microanalysis of coexisting zoisite teen specimens from six localities in Japan have been and clinozoisite are described below, with our view studied with the electron-probe microanalyser. Zoisite on the temperature-dependence of the gap in the and clinozoisite are commonly zoned, but compositional temperature range of low- to medium-grade meta- gaps between them are systematic. Referring to the metamorphic grade of the host rocks, a temporary and morphism of high-pressure intermediate type. schematic phase-diagram for the system Ca2AIaSi3OI2- (OH)-Ca2AI2Fea+Si3012(OH) is presented. With in- Mode of occurrence of coexisting zoisite and creasing temperature, in the range of low- to medium- clinozoisite grade , the compositional gap between the two -group shifts towards higher Fe 3+ Fig. I shows specimens localities. Brief accounts compositions. of the geology and petrology of these areas and the mode of occurrence of coexisting zoisite and EPIDOTE-GROUP minerals with the general for- clinozoisite are described below. mula Ca2(A1,Fea+)aSi3012(OH) have two series Iratsu and Tonaru epidote-amphibolite masses. of solid solutions, zoisite and clinozoisite-pistacite. The Iratsu and Tonaru masses are metamorphosed The chemical compositions of coexisting zoisite and layered gabbros that occur in the epidote clinozoisite have been reported by many authors amphibolite-facies area in central Shikoku (Banno (Banno, I964; Myer, 1966; Ackermand and Raase, et al., 1976; also for general petrology, cf. Ernst I973; Hietanen, I974; Raith, i976), but their et al., I97o and Banno et al., 1978 ). This is the stability relations are insufficiently understood. highest metamorphic grade areas in the Sanbagawa There are two opposite views on which of zoisite metamorphic terrain. Main rock-types are epidote and clinozoisite is the higher-temperature phase. amphibolite, garnet-epidote amphibolite, horn- Banno 0964), Holdaway (I97z), and Raith (1976) blende eclogite, eclogite, bimineralic favoured the view that zoisite is the higher- eclogite, and zoisite rock. The coexistence of zoisite temperature phase and hence the compositional and clinozoisite is common in these masses, and gap between zoisite and clinozoisite shifts towards has been found in three rock-types; leucocratic the Fe 3 +-rich region with increasing temperature, epidote amphibolite, hornblende eclogite, and whereas Ackermand and Raase (1973) considered zoisite rock. In the following these epidote amphi- the opposite to be true. Because of the difficulty of bolites are sometimes referred to as metagabbros, the direct determination of the phase relations by but the samples we have studied do not exhibit the synthetic experiments, petrographic data have to textures and mineralogies of metamorphism other be fully used to solve this problem. We have studied than the Sanbagawa event. the compositional gap between zoisite and clino- In the leucocratic epidote amphibolite the occur- zoisite in specimens collected at six localities in rences of zoisite and clinozoisite are classified into Japan (Iratsu, Tonaru, Fujiwara, Nishisonogi, three types by texture. In the first type they occur Yoshimi, and Omi). The results mainly based on as subhedral grains measuring 5o-5oo pm. This is Copyright the Mineralogical Society 1006 M. ENAMI AND S. BANNO Bessi area

" ~ .~

Go :::::::::::::::::::::::::::::

I R : Iratsu mass TO : Tonaru mass FU : Fujiwara mass YOSHIMI HA : Higashi-akaishi mass P

BESSI ( IRATSU, TONARU, FUdlWARA ) NISHISONOGI

FIG. I. Localities of the samples investigated in this paper. A metamorphic zone map of the Bessi district is shown in the upper-left of the figure. the commonest among the zoisite+clinozoisite- boundary of two layers, where the zoisite+ bearing specimens we have dealt with. The second clinozoisite + hornblende + + muscovite type consists of an aggregate of fine-grained (about +quartz assemblage occurs. Paragonite and 2o-50 #m), subhedral zoisite and clinozoisite (speci- muscovite are subordinate in amount. In this men ME75o425oi). The third type is the inter- rock-type, zoisite and clinozoisite are fine-grained, growth of zoisite and clinozoisite in a single grain measuring about 4o-20o /~m. The zoisite and (specimen ST7o55a ). clinozoisite in this rock-type are not in mutual In hornblende eclogite, clinozoisite is fine- contact but their chemical relation is the same as grained (50-lOO pm) and subhedral, and occurs in those found in mutual contact in other specimens. the interstices of large subhedral grains (o.5- Fujiwara albite-epidote amphibolite mass. The I.O mm) of zoisite. Fujiwara mass is a complex of ultrabasic and basic Zoisite rock is metamorphosed anorthosite, rocks (Onuki et al., 1978; Enami, in press). It occurs originally forming layers in gabbro, and consists in the garnet zone of Higashino et al. (1977) and of zoisite-rich leucocratic bands intercalated with Banno et al. (I978) of the Sanbagawa terrain in hornblende-rich ones. Zoisite is rare in the central Shikoku, and is situated about 8 km east hornblende-rich band, whereas clinozoisite is rare of the Iratsu mass, probably belonging to the lower in the zoisite-rich one, but they coexist at the albite-epidote amphibolite facies. Yokoyama ZOISITE-CLINOZOISITE RELATIONS too7

(1976) estimated the equilibration temperature of Metagabbro of the Omi area. The Omi area is the mass to be about 35o-45o ~ on the basis of located in the southern part of the Niigata Prefec- the antigorite + brucite and epidote + actinolite ture and is a glaucophanitic metamorphic terrain assemblages. (Banno, I958). The coexisting zoisite and clino- The coexistence of zoisite and clinozoisite was zoisite are found in a metagabbro which occurs found in a thin section containing hornblende + as a tectonic block in the albite-epidote amphi- dinopyroxene + chlorite + muscovite + zoisite + bolite facies area (Maruyama, pers. comm. I978 ). clinozoisite + albite. Clinopyroxene is a relic from The specimen studied has hornblende + zoisite + pre-metamorphic assemblage. The clinozoisite is clinozoisite + albite. Prehnite occurs in veins. Sub- subhedral, is surrounded by aggregates of prismatic hedral clinozoisite is in contact with zoisite, forming zoisite, and is optically homogeneous. Most of the an aggregate of fine (2o-3o #m) subhedral grains. zoisites show sector zoning (Enami, 1977), but some optically homogeneous grains do occur and were Chemistry of coexisting zoisite and clinozoisite used in the analyses. Metasomatized metagabbro in the Nishisonogi Zoisite and clinozoisite were analysed by an area. The Nishisonogi area is a high-pressure dectron-probe microanalyser, Hitachi Model metamorphic terrain located in the western part XMA 5A, with accelerating voltage of 15 kV, of the Nagasaki Prefecture, Kyushu. Metagabbros specimen current 0.o2 pA, and beam diameter of this area mainly consist of actinolite, chlorite, 5/~m. ornphacite, zoisite, clinozoisite, muscovite, para- Chemical analyses were performed on 8 speci- gonite, and albite. Jadeite (but not with quartz) mens from the Iratsu mass, 3 from the Tonaru mass, and Na-amphiboles also occur. The associated I from the Fujiwara mass, 2 from the Nishisonogi rocks are inferred to belong to the garnet zone of area, 2 from the Yoshimi area, and I from the Omi the Sanbagawa belt, on the basis of common area. Table I lists the analyses. presence of garnet in pelitic schists (Nishiyama, The coexistence of zoisite and clinozoisite was I978; Nishiyama, pers. comm. i978 ). confirmed on four specimens by X-ray powder The coexisting zoisite and clinozoisite were patterns of magnetic and less-magnetic fractions found in banded metagabbros consisting of of epidote concentrates following the criteria of amphibole-rich and zoisite-rich bands. Their Seki (1959) as well as on three samples by a parageneses are tremolite + omphacite + single- method using an automated four- chlorite + muscovite + clinozoisite and zoisite + circle X-ray diffractometer. In all these specimens clinozoisite + chlorite + muscovite + albite, respec- clinozoisite is distinctly richer in Fe 3 § than zoisite, tively. The zoisite and clinozoisite in the zoisite-rich and they may be identified by the difference of unit are in mutual contact. Grain sizes are variable interference colour that clinozoisite is yellow to ranging from lOO/~m to 2 mm. lavendar-blue whereas zoisite is grey or bluish grey. Metagabbro of the Yoshimi area. This area is Zonal structure. Many of the clinozoisites located to the east of the Kanto mountains, where examined are optically heterogeneous, but the a small isolated faulted block of metamorphosed zoisites appear to be homogeneous optically, basic and ultrabasic complexes is exposed in an although compositional variation was found by area of Tertiary sediments. Main rock-types are microanalysis. Three types of zoning patterns were garnet amphibolite, schistose amphibolite, horn- recognized in the clinozoisites. In the first type blendite, pyroxenite, two-mica gneiss, and serpen- observed in a specimen from the Tonaru mass tinite (Murai, 1968 ). Yokoyama (i976) has distin- (specimen HK9oIC), the Fe 3+ content, being guished two stages of recrystallization in this measured in terms of X~ = Fe a +/(Fe 3 + q- A1) here- complex, garnet amphibolite and greenschist after (suffixes zo and cz stand for zoisite and clino- stages. The constituent minerals formed in the zoisite respectively), decreases from the core to- garnet amphibolite stage are hornblende, garnet, wards the rim, with minor fluctuations, but it again clinopyroxene, and , and those of the increases at the very margin. In this specimen the greenschist stage are actinolite, chlorite, zoisite, zoisite composition shows minor fluctuations clinozoisite, and albite. (X~ = o.o36-o.o46), but X~ invariably increases Zoisite and clinozoisite coexist in garnet amphi- at the very margin, thereby giving nearly the same bolite. The plagioclase, once equilibrated with the X~ and X~ at the zoisite-clinozoisite interface, garnet and hornblende, now consists of albite and respectively. Fig. 2 shows the compositional profile subordinate amounts of fine-grained (about 2o- across such an interface in this specimen. 30 #m) zoisite and clinozoisite, which are in mutual In the second type, X~ increases more or less contact. The zoisite + clinozoisite assemblage was continuously from the core to the rim, and X~ stable in the second, lower-grade stage. increases from the core to the rim as well. This I008 M. ENAMI AND S. BANNO

o o c-. m o , N , I g g ~: ~ 4 , 4 , , ,,,: 0 ~ 0 -- 0 ,I ~ ,, , 0 0

......

4'~ ,g ~ g ', 4 ', ', ~ ,g

B ~, ~ ~: ~ ~: ,~ ~ : 4 ', ', ,~ ......

...... ~o-ooooooo m m ~ ~ NNOOOO~OOOO o ~~s~;s~ ~222222G222

o ~ N ~ g M g g g I : I :lgg

i

~~

o~oNoo~oooo

~ ~ ~ ~ ~ ~ ~ ~ , ~ , , ~ ~

o .~ N ~ o N o o ~o ~ ~ o ~ o o ~ o o ,~ o ...... o ~0 o ~ o o c,, o o o o ~ m N 0", ~ ~l O O O O ~ O o O O

c,- ~o o ~ o o i-- ...... r

,~m N c; c; ~ ~ 4 ; 4 ; I ,~ o ~0 o - o , c,, , , o o

,~ .o ~ ~ ~ ~~ ', ,.- I ', ~ "" ~ ':" ...... ~" ~

m o m ~ o N ,~ o o m o ~ o m o N ~0 o o m o

o 0 ,c

~ o N ,~ o m ~, o o ~ o ~o o m o m ~ o o .~ o ~ '~

i

,~ ~ ~.~ ~ --E~0,~ ~o 0~0 0. o ~ + ~ ZOISITE-CLINOZOISITE RELATIONS xoo9

interfaces were equilibrated at the same stage of 015 l 2o" - reerystallization, even if the temperature or exact r~ ,~' time of equilibration may differ from place to place to some extent, giving rise to slight variation of Xps~176 ~. _o._... ~..._?-'%...... X~ at the interfaces. In this specimen the composi- 0o5 ,...~~ ....~ . tional variation of the zoisite at the interface is .7oF: o.o32 to o.o46, and is much smaller than that of rim [ core [ rim the clinozoisite. 0 zoisite clinozoisite The variation of Xp~ at the interfaces in other specimens differs in different specimens, but is FIG. 2. Compositional profile of a zoisite-clinozoisite smaller than that shown in the above example (cf. interface of sample HK9oIC from the Tonaru mass of the Bessi district. Table I). In the specimens of the Iratsu and Tonaru masses of the Bassi area and those of the Omi area, the zoning type is common in the specimens from the rims of clinozoisites that are not in contact with Iratsu mass and is not uncommon in those from zoisite have the same composition as those in the Nishisonogi area, but is absent in others. contact with zoisite. In the specimens from the The third type of zoning also shows the con- Yoshimi area, however, the former type of clino- tinuous variations of Xps, but X~ decreases from zoisite has a Fe3+-richer composition than the the core to the rim. In the associated zoisite, X~~ other type as shown in Table I. This shows that decreases from the core to the rim as well. This chemical equilibrium was satisfied only in small type occurs in the Nishisonogi area. domains in the specimens from the Yoshimi area. Choice of equilibrium pair. Fig. 3 shows the To avoid possible mix-up of various equilibrium frequency distribution of X~z at the core and rim domains, only the compositions at the interface of of the type I clinozoisite. The X~ at the core has coexisting zoisite and clinozoisite will be treated a wide variation ranging from o.o79 to o.I2o even in the following discussion, even for the specimens in a single thin section, but the rim in contact with from the Bessi and Omi areas. zoisite has fairly constant Xp~, ranging from o.t Io In the zoisite-clinozoisite composite grain from to oA3o. The precision of microanalysis is tr = the Iratsu mass (specimen ST7o55a), the composi- o.ooz X~ in count statistics, and hence the tional gap at the interface is quite different from observed variation of Xp~ at the interface exceeds those at the interfaces of independent zoisite and the analytical errors. The presence of such a clinozoisite grains. In this particular composite variation may be due partly to the difference of grain X~ = o.o4o___o.oo7 and is fairly constant, equilibration temperatures from one interface to but X~s ranges from o.x45 to o.I98. In other another, but may also be due to the difficulty of specimens from the same mass the coexisting zoisite determining X~ at the rim of zoned clinozoisite. and clinozoisite have X~ = o.o5I+O.OO4 and As is seen in fig. 2, the change of X~z for o.oI takes Xg = o.14o +o.oo6 at the interfaces. The fact that place over a short distance of IO ~m. At all the Xp~ varies between different faces within the interfaces of coexisting pairs examined in this composite grain as well as differing significantly specimen, the pattern of the zoning, so far as it from values at the faces of isolated grains suggest exists, is the same. Therefore, we consider that the that the interfaces of the composite grain were not in chemical equilibrium. N Based upon the above discussion, we proceed assuming that the interfaces of independent zoisite I I core /o and clinozoisite were in chemical equilibrium, even ~'////~ rim if this was not attained at exactly the same tempera- ture in each specimen. The deviation from equili- brium may be measured in terms of the standard deviation of X~, which is smaller than o.oi Xg in each specimen. The relationships between Xb~ and X;~ at the

I interfaces of all the studied specimens are shown o. 1o O.l~ in fig. 4. The cross at the upper right of the figure shows the average variation of Xps of zoisite and Xps clinozoisite within a single thin section. There is a FIG. 3. Frequency distribution of the values of X~] for distinct correlation between X;~ and X;~ over the sample HK9oIC. specimens studied, although the variation of the I010 M. ENAMI AND S. BANNO trend exceeds slightly that in an individual thin temperature, or zoisite solid-solution is the higher section. temperature phase. In the binary system the compositions of the Epidote-group minerals contain Mn 3+, Fe 2+, coexisting phases plotted in a diagram like fig. 4 and Mg 2 +, but the contents of MnO (total Mn as should define a line determined by the transitional MnO, but probably mostly in the trivalent state) loop in the phase diagram. and MgO determined in this study are so low that they may hardly affect the phase relations in regard Discussion to Fe 3 +-A1 substitution. Some Fe 2 + may be con- tained in zoisite. The Fe2+/Fe 3+ ratio in some Temperature-dependence of the compositional zoisites analysed by wet chemistry approaches I, gap. Among the six localities studied in detail, including a zoisite from the Tonaru mass described obvious difference of metamorphic grade is recog- by Banno (I964). The possible variation of the nized between the Iratsu and Tonaru masses of the Fe2+/Fe 3+ ratio of zoisite may partly be respon- Bessi area on the one hand, and the Nishisonogi sible for the scatter of points in fig. 4. area on the other; the former belonging to the However, the general tendency for the composi- epidote amphibolite facies and the latter to the tions of the coexisting zoisite and clinozoisite to epidote-glaucophane schist facies. Both of them become FeS+-richer with increasing temperature belong to the high-pressure intermediate facies is strong enough to ignore, at least in the studied series. The compositional gaps observed on the specimens, the effect of Fe 2 +. The effect of pressure specimens from the Iratsu and Tonaru masses among the Japanese specimens, except on the scatter rather widely in fig. 4, but if we combine Yoshimi pair, which are from the high-pressure them into three subgroups based upon the mode intermediate type metamorphism may be very of occurrence, the scatter in each of them becomes small, in view of the fact that the unit cell volumes significantly reduced to be more or less comparable of zoisite and clinozoisite (I/2 unit cell volume for with that of the Nishisonogi specimens. The three zoisite) extrapolated to the same composition are groups are the central and marginal parts of the quite similar as shown by Seki 0959) and later Iratsu mass and the Tonaru mass. As the thermal confirmed by Myer 0966) and Dollase (I968). history of these masses is complex, it is possible Accordingly, we conclude that the compositional that the final temperature of equilibration could gap between zoisite and clinozoisite shifts towards have been slightly different between these masses Fe 3 +-richer side as the temperature increases. In and even in different parts of the masses. All three the following, the phase relations between zoisite subgroups of metagabbros of the Bessi area belong and clinozoisite in other areas will be examined in to the epidote amphibolite facies and they were order to test the validity of our conclusion. equilibrated at higher temperatures than the In the epidote amphibolite of the Omi area the Nishisonogi specimens. Thus it follows from fig. 4 compositional gap is between o.o43 ___o.oo4 and that the compositions of coexisting zoisite and o.I 16 + o.oo2 Xps, which is comparable to those of clinozoisite become FeS+-richer with increasing the specimens from the Iratsu and Tonaru masses. The Fujiwara mass is situated in the lowest-grade Equilibrium pair area of the albite-epidote amphibolite facies; that oe central part (6) ...... -q is, its metamorphic grade is intermediate between marginal port (I)----o J IRATSU TONARU ~t those of the Bessi and Nishisonogi areas. The Xp, + FU41WARA ( t ) OMI (t) values of zoisite and clinozoisite in the specimen NISHISONOGI (2) I~ YOSHIMI (2) from this mass are also between those of the Bessi Disequilibrium pair 20" and Nishisonogi areas. The coexisting zoisite and -e-lntergrowth pair (I) ..... IRATSU clinozoisite of the specimens from the Yoshimi area 0.05 ~ :": a" ~ n probably represent the lowest temperature among the studied areas. This conclusion is not positively x '~ps :2':"..... F. disequilibrium pa '71 proved by independent mineralogical data, but is u B o compatible with the occurrence of albite + chlorite + o u o epidote + actinolite in some of the associated rocks. o ~ ~ ~ , I , , , t I , t , , I Therefore, the relative temperatures of equilibra- 0.I0 O.l& O.EO X ez tion of the five localities (excluding Yoshimi) being ps studied in this work, as estimated by the composi- FIG. 4. Compositions of coexisting zoisite and clinozoisite tional gaps, are not contradictory with the para- at their interfaces. Numbers in parenthesis show the genesis of associated minerals and metamorphic number of samples studied in detail. (For tile disequili- grade of the associated rocks. Fig. 5 shows the brium pairs, see the text.) inferred phase diagram of the zoisite-clinozoisite- ZOISITE-CLINOZOISITE RELATIONS IOII pistacite system at the temperatures of low- to aureole of northern Sierra Nevada. However, the medium-grade metamorphism. The temperature coexistence of three epidote-group minerals of scale cannot be calibrated quantitatively. The dia- essentially binary system, zoisite, clinozoisite, and gram is part of a transitional loop, but we cannot epidote in one thin section makes it difficult to extrapolate the observed relations to higher tem- evaluate the significance of the observed para- perature unless we know more about the mixing genesis. Tanner (I976) reported that clinozoisite properties of the epidote-group minerals. with Xg = o.I3O-O.t6o coexists with zoisite in his zones 2 and 3 of the Moine schists, in which plagioclase (An = 33-97) occurs. If zoisite with T I X~ = 0.04 coexists with clinozoisite Xp~ = oA6o, i / i it is not contradictory with our view, but if it I epidofe --11"- --IRA-- r coexists with the one with X;~ = o.I 3 in equili- ornphibolite , 7 focies -e- --OM-- --o-/ brium, our view has to be modified. Further consideration may be fruitless until more detailed zoisife ,; ," cl[nozoisife descriptions of the chemical heterogeneity of clino- --e--s --FU-- --e-I epidofe- zoisite from this area are available. Ackermand and r schisl -~ --NI-- ~' focies / / Raase (I973) showed that the compositional gap ] / i between zoisite and clinozoisite in plagioclase- greenschisl l i t facies ,o --YO-- ~ = bearing specimen from Hohe Tauern is about __ ~ I , / I , I 0.05 010 015 o.o4-o.Io X~. This compositional range corre- Xps sponds to that of the lower albite-epidote amphibolite-facies conditions similar to those of FIG. 5- Speculative temperature-composition relations for the Fujiwara mass. It is not evident whether or not the zoisite-clinozoisite series. The transition loop is the zoisite + clinozoisite assemblage coexists stably approximated by straight lines, and the relative formative with plagioclase, because Miller (I977) has shown temperatures were inferred from X~. Abbreviations. that, in the associated eclogitic rocks of Hohe OM: Omi, TO: Tonaru mass of the Bessi district, FU: Fujiwara mass, NI: Nishisonogi metagabbro, YO: Tauern, this assemblage was not in equilibrium Yoshimi metagabbro, IR: Iratsu mass of the Bessi district. with plagioclase (4- to ~ An) but postdated it. The Subscripts C, M, and A denote the central part, marginal Xps values of coexisting zoisite and clinozoisite part, and average of the Iratsu mass, respectively. from Trollheim, Norway, as described by Myer (I966) are estimated to be o.o4 and o.I2, respec- tively, that again correspond to the albite-epidote amphibolite facies. The mineral facies of the Nor- Data from other metamorphic terrains. In the wegian specimen is considered to be the middle literature, there are several groups of data which almandine amphibolite facies. are apparently contradictory with the temperature In these examples the compositional gap gives dependence proposed here. Ernst (I977) reported a lower temperature than that estimated by the optically identified clinozoisite with Xp, = o.oz 5- mineral paragenesis, if we accept the phase rela- o.o29 from eclogitic rocks of the Alpe Arami tions obtained on Japanese rocks. In this connec- mass in the assemblage garnet + clinopyroxene + tion it is possibly significant that the phase relations hornblende + plagioclase (An = 15-37) + biotite + in fig. 5 were constructed from data on high- quartz. With the courtesy of Professor Ernst, we pressure metamorphic rocks. Pressure may hardly obtained the Alpe Arami specimens described by affect the zoisite-clinozoisite relations, but it may him and their epidote-group minerals were re- affect the temperature range of mineral facies, examined by automated four-circle diffractometer. especially of the epidote amphibolite facies, bring- The epidote-group mineral in specimen F-~6fwas ing its higher temperature limit to higher tempera- found to be zoisite with possible space groups Pnma ture with increasing pressure (Miyashiro, i96I). It if centric and Pn21a if acentric. The unit cell is, therefore, possible that the epidote amphibolite dimensions are ao = I6.2o2(I), bo = 5.554(0, and facies in high-pressure terrains represents a similar Co = I o.o33(~) A, respectively, in which the number temperature to the middle of the almandine amphi- in the parenthesis is the standard error in the last bolite facies of medium-pressure type of meta- digit. These values agree well with those given by morphism. Dollase 0968). If the Alpe Arami epidote is actually Frey (I978) has reported that biotite schists of zoisite, it is consistent with our proposed phase the Helvetic zone of the Alps have zoisite with relations. Hietanen (i974) described clinozoisite X~'~ = 0.03 (Fe203 = 1.7 wt~o) coexisting with with Xp~ = 0.044-o.094 coexisting with zoisite and zoned with clinozoisite rims with Xps = epidote in the amphibolite-facies area of the contact o.o9. The grade is lower than the staurolite iso- IOI2 M. ENAMI AND S. BANNO grad and associated plagioclase is oligoclase or 6C, we have to consider that the rock reached the andesine. The associated rocks commonly contain albite-epidote amphibolite facies before the final margarite + quartz. The paragenesis of calc-alumino equilibration. The latter view is hardly consistent silicates of this area indicates lower temperatures with the general petrology of the Sanbagawa schists than the metagabbros of the Bessi area (Tonaru in Shikoku. and Iratsu masses) that contain kyanite + zoisite + quartz, in harmony with what is expected from zoisite-clinozoisite relations. Alternative model of temperature dependence. Our /~ foregoing discussion leading to fig. 5 is essentially N core based upon the assumption that the equilibrium rim temperature of interface pairs of the metagabbros, 5 at Bessi, Nishisonogi, and Yoshimi areas, decreases in this order. If we accept the opposite temperature dependence that clinozoisite is the higher- temperature modification, we must consider that the temperature increases in that order. As the ~_j parageneses of the host and associated rocks of 0 )/0 015 02O these pairs show that the temperature decreases in Xps that order, we must consider that the Bessi inter- g faces were produced by retrograde metamorphism in which material exchange and growth of zoisite / s n continued to much lower temperatures than in the epidofe other areas. In this model the core of clinozoisite amphibolite zoisit~// /linozoisite from the Bessi metagabbro (Xp~ =0.08-O.I2) facies represents the epidote amphibolite-facies. This pos- nm sibility cannot be fully rejected as we know little greensehist / ,/ about the rate of cooling, kinetic effects on recrys- facies t t tallization, etc. that affect the temperature of freez- , ,, core ing the local equilibration. To examine the alternative model further, we consider epidote in I I-. J 0 or o2o basic schists of the Sanbagawa belt. At low-grade, Xps of epidote generally decreases with the grade, Xps as pointed out by Miyashiro and Seki (I958), but C at grades higher than the breakdown of pumpelly- T c~ Iratsu mass % 0 r~m- ite, Xr~ is largely determined by the bulk rock r////~ core chemistry (Nakajima et al., 1977). The frequency epidote ~ rim~ AS-78B distribution of X~z of clinozoisite in specimen amphibolite AS-78B from the Asemi-river route in central facies \ inozoisite Shikoku, which was donated by Mr T. Nakajima, zoisite0 is shown in fig. 6A. The host rock contains alb- ite + clinozoisite + chlorite + aluminous actinolite ,~176176 greenschist i (AlzO s = 2 wt %), but not zoisite. The grade of the facies rock is slightly higher than the disappearance of I pumpellyite, and since associated pelitie schists contain garnet with MnO = 5-7 wt % (Kurata and J I , J Banno, ~974; Higashino, I975; Nakajima, pers. 0 OJO 02O comm.). Thus it is slightly lower than the epidote Xps amphibolite facies. The clinozoisite is zoned with an Fe 3 +-poor core (X;z = o. 12) and Fe 3 +-rich rim FIG. 6A. Frequency distribution of X~ in a greenschist (X;x = o.I7). Nakajima (pers. comm.) interprets of central Shikoku (sample AS-78B). The rock is from this zoning by the selective decomposition of the the garnet zone of pelitic schists, and of lower grade than the first appearance of biotite in pelitic schists. B. Pro- Al-end member of epidote solid-solution with pro- posed temperature variation of the sample AS-78B. gressive metamorphism. This view as shown in fig. Legend is shown in Fig. 6C. C. Alternative model of 6B is consistent with our proposed temperature temperature variation of the same greenschist. In this dependence. On the other hand, if we accept the model the schist had to be in the epidote amphibolite opposite temperature dependence, as shown in fig. facies when the core of the clinozoisite was formed. ZOISITE CLINOZOISITE RELATIONS 1o~3

For these reasons, so far as the first-hand data rocks, northern Sierra Nevada, California. Am. we have and the general petrology of the Mineral. 59, 22-40. Sanbagawa belt are concerned, we prefer the model Higashino (T.), 1975. Biotite zone of Sanbagawa meta- shown in fig. 5. morphic terrain in the Shiragayama area, central Shikoku (in Japanese with English abstract), d. Geol. Soc. dpn. 81, 653-7 o. Acknowledgement. Our sincere thanks are due to Pro- --Hide (K.), and Banno (S.), I977. Metamorphic zone fessor W. G. Ernst, who donated us his Alpe Arami map of the Sanbagawa belt in Shikoku island and Kii specimens, and kindly reviewed the manuscript with peninsula (in Japanese with English abstract) in many comments, and Dr K. Kihara who helped us to Hide, K., ed., Sanbagawa belt, Hiroshima Univ. Press. use the X-ray goniometer. We are also due to Messrs T. Pp. 2oi-6. Nakajima, T. Nishiyama, and S. Maruyama for personal Holdaway (M. J.), 1972. Thermal stability of AI-Fe communications of the Asemi, Nishisonogi, and Omi epidote as a function of fco~ and Fe content. Contrib. areas, and Dr K. Yokoyama, Messrs T. Nakajima, T. Mineral. Petrol. 37, 3o7-4 o . Nishiyama, O. Iwata, and T. Tanaka for donating his Kurata (H.), and Banno (S.), 1974. Low-grade progressive zoisite-clinozoisite-bearing rocks. We also wish to thank metamorphism of pefitic schists of the Sazare area, here Messrs S. Kasashima, K. Nakamura, and Mrs Sanbagawa metamorphic terrain in central Shikoku, Y. Fukada for making thin sections, drafting the figures, Japan. J. Petrol. 15, 361 82. and typewriting the manuscript. Miller (C.), I977. Mineral parageneses recording the P,T history of Alpine eclogites in the Tauern Window, Austria. Neues Jahrb. Mineral. Abh. 130, 69-77. Miyashiro (A.), 1961. Evolution of metamorphic belts. REFERENCES J. Petrol. 2, 277-311. -- and Seki (Y.), I958. Enlargement of the composition Ackermand (D,) and Raase (P.), I973. Coexisting zoisite field of epidote and piemontite with rising temperature. and clinozoisite in biotite schist from the Hohe Tauern, Am. J. Sci. 256, 423-3 o. Austria. Contrib. Mineral. Petrol. 42, 333-41. Mural (T.), 1968. On the garnet amphibolites of the Banno (S.), 1958. Glaucophane schists and associated Yoshimi Hill, Saitama Prefecture, Japan. Sci. Rept. rocks in the Omi district, Niigata Prefecture, Japan. Saitama Univ. 2, 65-71. Jpn. J. Geol. Geogr. 29, Z9-44. Myer (G. H.), I966. New data on zoisite and epidote. --1964. Petrologic studies on Sanbagawa crystalline Am. J. Sci. 264, 364-85 , schists in the Bessi-lno district, central Shikoku, Japan. Nakajima (T.), Banno (S.), and Suzuki (T.), 1977. Re- J. Fac. Sci. Univ. Tokyo, sec. II, 15, 2o3-319 . actions leading to the disappearance of pumpellyite in -- Yokoyama (K.), Iwata (O.), and Terashima (S.), 1976. low-grade metamorphic rocks of the Sanbagawa meta- Genesis of epidote amphibolite mass in the Sanbagawa morphic belt in central Shikoku, Japan. J. Petrol. 18, metamorphic belt of central Shikoku (in Japaneses with 263-84. English abstract). J. Geol. Soe. Jpn. 82, 199-2IO. Nishiyama (T.), 1978. Jadeitite from the Nishisonogi -- Higashino (T.), Otsuki (M.), Itaya (T.), and Nakajima metamorphic region (in Japanese). J. Geol. Soc. Jpn. (T.), 1978. Thermal structure of the Sanbagawa meta- 84, I55 6. morphic belt in central Shikoku. J. Phys. Earth, 26, Onuki (H.), Yoshida (T.), and Suzuki (T.), 1978. The Supplement, 345-56. Fujiwara mafic- ultramafic complex in the Sanbagawa Dollase (W. A.), 1968. Refinement and comparison of the metamorphic belt of central Shikoku I. Petrochemistry structures of zoisite and clinozoisite. Am. Mineral. 53, and rock-forming mineralogy. J. Jpn. Ass. Mineral. 1882-98. Petrol. Econ. Geol. 73, 311-22. Enami (M.), 1977. Sector zoning of zoisite from a meta- Raith (M.), I976. The A1-Fe(III) epidote miscibility gap in gabbro at Fujiwara, Sanbagawa metamorphic terrain a metamorphic profile through the Pennic series of the in central Shikoku. J. Geol. Soc. Jpn. 83, 693- 7. Tauern Window, Austria. Contrib. Mineral. Petrol. 57, Ernst (W. G.), 1977. Mineralogic study of eclogitic rocks 99-117. from Alpe Arami, Lepontine Alps, Southern Switzer- Seki (Y.), I959. Relation between chemical composition land. J. Petrol. 18, 371-89. and lattice constants of epidote. Am. Mineral. 44, --Seki (Y.), Onuki (H.), and Gilbert (M. C.), 197o. 72o-3 o . Comparative study of low-grade metamorphism in the Tanner (P. W. G.), I976. Progressive regional metamor- California Coast Range and the outer metamorphic phism of thin calcareous bands from the Moinian rocks belt of Japan. Mere. 124 Geol. Soc. Am. 276 pp. of N.W. Scotland. J. Petrol. 17, Ioo-34. Frey (M.), I978. Progressive low-grade metamorphism Yokoyama (K.), 1976. Ultramafic and related rocks in of a Black Shale formation, central Swiss Alps, with the Sanbagawa metamorphic belt. Ph.D. Thesis, Univ. special reference to pyrophyllite and margarite-bearing Tokyo. assemblage. J. Petrol. 19, 95-I35. Hietannen (Anna), 1974. Amphibole pairs, epidote [Manuscript received 15 November 1979; minerals, chlorite and plagioclase in metamorphic revised 5 June I98o ]