A calculated petrogenetic grid for ultramafic rocks in the system CaO-FeO-MgO-Al2O3-SiO2-CO2-H2O at low pressures

Contrib Mineral Petrol (1990) 105:347-358

C o n t r i b u t i o n s to

Mineralogy and Petrology 9 Springer-Verlag1990

A calculated petrogenetic grid for ultramafic rocks in the s y s t e m CaO - FeO - MgO - A1203 -SiO2 - CO2 - H20 at low pressures Thomas M. Will j, Roger Powell 1, and Tim J.B. Holland 2 1 Departmentof Geology,Universityof Melbourne, Parkville,Victoria 3052 Australia z Departmentof Earth Sciences,Universityof Cambridge,DowningStreet, Cambridge CB2 3EQ, England,UK Received November 21, 1989 / Accepted April 12, 1990

Abstract. Calculated phase equilibria among the minerals amphibole, chlorite, clinopyroxene, orthopyroxene, olivine, dolomite, magnesite, serpentine, brucite, calcite, quartz and fluid are presented for the system CaO - FeO - MgO - A 1 2 0 3 - S i O 2 - C O 2 - H 2 0 (CaFMASCH), with chlorite and H 2 0 - C O z fluid in excess and for a temperature range of 440 ~ C-600 ~ C and low pressures. The minerals chosen in CaFMASCH represent the great majority of phases encountered in metamorphosed ultramafic rocks. The changes in mineral compositions in terms of FeMg 1 and (Mg, Fe)SiAI_ 1AI_ 1 are related to variations in the intensive parameters. For example, equilibria at high Xco2 in the presence of chlorite involve minerals which are relatively aluminous compared with those at low Xco2. The calculated invariant, univariant and divariant equilibria are compared with naturally-occurring greenschist and amphibolite facies ultramafic mineral assemblages. The correspondence of sequences of mineral assemblages and the compositions of the minerals in the assemblages is very good.

Introduction Alpine-type ultramafic rocks are present world-wide in a wide variety of tectonic settings, for example withinophiolite and greenstone belts. Particularly in the latter, pressures of metamorphism are usually low; metamorphism of ultramafie assemblages at low pressures is the focus of this paper. Although the phases encountered in ultramafic rocks at low pressures are restricted to only a few minerals, their phase relationships are complex. Commonly, phase petrological investigations of ultramafic rocks either based on experimental data or on field observations restrict themselves to the system Offprint requests to: T.M. Will

MgO - SiO2 - C O 2 - H20 (e.g. Winkler 1979; Gole et al. 1987) or, sometimes, to the system C a O - M g O - S i O 2 - H z O - C O 2 (Evans 1977). There have been, however, no detailed studies of ultramafic assemblages in the system C a O - F e O - M g O - A I 2 0 3 - S i O 2 - C O 2 - H z O (CaFMASCH). This system is much closer to natural assemblages because Tschermak's and M g - F e substitutions in the minerals can be incorporated. Previous studies that included calculations of mineral equilibria were restricted to minerals involving single end-members because the datasets used did not include a sufficiently wide range of mineral end-members (e.g. Evans and Guggenhelm 1988, using Berman 1988). As shown here, the results of such calculations do not show the complexities of systems involving solid solutions. Using the new, expanded, internally-consistent dataset of Holland and Powell (1990) with the computer program, THERMOCALC (Powell and Holland 1988), complex mineral equilibria including solid solutions can be calculated. This contribution focuses on theoretical phase relationships in the system C a O - F e O - M g O - A 1 2 0 3 - SiOz - CO2 - HzO, with chlorite and CO2 - H20 fluid in excess, and compares the results with naturally-occurring mineral assemblages. This system comes close to natural ultramafic assemblages, even though the effects of TiO2, Fe203 and Cr203 have to be neglected. With respect to CaFMASCH, TiO2 can be considered to occur in excess, stabilising either rutile, ilmenite or clinohumite, but the incorporation of ferric iron may not only stabilise, for example, magnetite, but will also affect the relative stabilities of amphibole, chlorite and - to a much lesser extent - those of pyroxene and talc. Calculations in CaFMASCHO await the generation of reliable thermodynamic data for ferric end-members of phases. Nevertheless, the system CaFMASCH will cover most of the phases and the phase relationships normally encountered in ultramafic rocks. The minerals included, and their structural formulae, together with abbreviations for mineral end-members used in this text, are given in Table 1. The activity-composition relationships used, with

348 Table 1. Mineral end-members, their abbreviations ~as used in text and ligures amph Amphibole tr Tremolite fir Ferro-tremolite hb Hornblende fhb Ferro-hornblende anth Anthophyllite

compositions

and

Ca2Mg3Mg2Si4Sir

Ca2Fe3Fe2Si4.Si4.022(OI][)2 Ca2Mg3 [MgA1] [Si3A11] Si4022(OH)2 Caye3 [FeAI] [SiaAlz] Si4.Oz2(OH)2 MgzMg3 [Mg2] Si4Si4[Si4.]O22(OH)2

Chlorite

chl clin daph ames fame

Clinochlore Daphnite Amesite Ferro-amcsite

ta ta fta tats fret

Talc Talc Mg2MgSi2 [Si2] O1o(OH)2 FezFeSiz[Siz]Olo(OH)z Ferro-talc Tschermak's talc MgzAlSi=[SiA1]O~o(OH)2 Ferro-Tschermak's talcFe2AlSi2[SiA1]Olo(OH)2

cpx di hed cats

Clinopyroxene Diopside Hedenbergite Ca-Tschermak's

opx en fs mgts

Orthopyroxene Enstatite Ferrosilite Mg-Tschermak's

MgMgSi206 FeFeSi206 MgAI[SiA1]O6

ol fo fa

Olivine Forsterite Fayalite

MgeSiO4Fe2SiO4-

Mgg[MgA1][SiA1]SiaOlo(OH)s Fe4-[FeA1][SiA1]Si20 lo(OH)s Mg4[AI2] [Ale] Si20 lo(OH)s Fe4-[A12][A12]SieOIo(OH)s

Table 2. Activity-composition relationships for the phases which are solid solutions; in amphiboles, since no cations are allowed to enter the A-site and no cations other than Ca the M4-site, Xv,~,A= 1 and XCa, M4 = t

Amphiboles 3 2 4 XMg,Ml 3 XMg,M2 XSi,T1 3 2 4-

Tremolitc Ferro-tremolite Hornblende

a(ftr)=Xve,m3 XFe,M2Xsi,sl

Ferro-Hornblende

a(fhb)= 37.93

Chlorites Clinochlore

a(clin) = 16 x~g,m XMg,M2 XA1,MZ

a(tr) =

a(hb)=37.93

3 XMg,M13 XMg,M2 XAI,M2 3 Xsi,T1 XAI,T1 3 XFe,M13 XFe,M2 XAI,M2 3 XSi, T1 XAI,T1

XSi, 1"1 XAI,T1

dol Dolomite dol Dolomite film Ferro-dolomite mag Magnesite mag Magnesite sid Siderite antg Antigorite br Brueite cc Calcite

q

Quartz

CaMgSi206 CaFeSi206 CaAI[A1Si]O6

Daphnite

a(daph) = 16 x~,m XV~,M2XAI,M2

Amesite Ferro-Amesite

42 2 a(ames) = XMg, MI XA1, M2 XA1,T 1 a(fame)=x~e,m X2LM2 XAI,T1 Z

XSi,T1 XAI,T1

Talc Talc Ferro-Talc Tschermack's talc

2 a(ta)=x~g.m XMg.M2Xsi.Ti 2 2 a(fta)=Xe~,m Xv..r~2XSi,Xl 2 a(tats)=4 XMg,m XA1.M2XSI,Ti

Fe-Tschermak's talc

a(ftat)=4 xr2~,m XAI,MZXSJ.T1

XAI,T1 NA1,T1

Clinopyroxene Diopside Hedenbergite Ca-Tschermak's

a(di)=XMg,U2 XSI,T 2 a(hed)= xvo,M2 Xsi, 2T a(cats)=4 XA1,M2 Xaz,xXsi,T

Orthopyroxene Enstatite Fcrrosilite Mg-Tschermak's

a(en) = XMg,mXMg,M2 Xsi,'r : a(fs)=xvr XFe,M2Xgi,T a(mgts)=4 Xug,m XAI,M2XA1,T XSi, T

CaMg(CO3)2

CaFe(CO3)2 MgCO 3 FeCO~ Mg3Si2Os(OH)4Mg(OH)2 CaCO3 SiO2

As used in Holland and Powell (1990)

the exception of antigorite (which is discussed below), are given in Table 2. Using the dataset of Holland and Powell (1990), it is now feasible to calculate the P - T - x location of invariant points, univariant reactions, and divariant fields on petrogenetic grids, using the computer p r o g r a m T H E R M O C A L C (Powell and Holland 1988). The results of such a study are presented here for the C a F M A S C H system. P - T - x pseudosections (Hensen 1971) are used to consider the influence of bulk-rock compositions on mineral assemblages, as well as to generate internal and external buffering paths on isobaric T - - X c o ~ diagrams. Furthermore, the extent of (Mg, Fe)SiA1_ 1A1-1 substitutions in the minerals are related to changes in intensive parameters.

Olivine Forsteritc Fayalite Magnesite Magnesite Siderite Dolomite Dolomite Fe-dolomitc

a (Jo) = x~, a (fa) = xw2 a (mag) = x~g a(sid) = xvo

a (dol) = XMg a ( f dol) = xF,

Antigorite: activity of chrysotile Although antigorite is the c o m m o n form of serpentine in m e t a m o r p h o s e d ultramafic rocks, it is not a simple stoichiometric phase, and thermodynamic data exist in the Holland and Powell (1990) dataset only for chrysotile. However, we m a y resort to a useful artifice in simulating the thermodynamic behaviour of antigorite close to its breakdown temperatures by assuming that the activity of chrysotile in antigorite is close to 0.5. As discussed by Mellini et al. (1987), antigorite is not a p o l y m o r p h of chrysotile, but involves polysomatic behaviour with chrysotile-like structural units whose wavelength varies as a function of grade, leading to compositions extending

349

I

I

I

I

I

I

\~

I

I

ZZ i.)

eq

rJ3

<

O'~

II

c5

+

0~

.x ',,4-t ,4-a

c5

O

\

-4

o

m.

O

CD

fi

0

e~

l

l

I

I o ['-. tJ~

I

I 0 it3 Lib

I

I

O

I

0 03 ttb

o

Fig. 1. T--Xco~ petrogenic grid for the system C a O - F e O -MgO-AI2Oa-SiO2-COz-H20 at 2 k b a r projected from chlorite and C O z - H z O fluid. Bold lines are C a F M A S C H reactions; light lines are C a M A S C H reactions. Open circles surrounding CaMASCH invariant points indicate that the location of the CaF-

MASCH invariant point is identical to the CaMASCH position. Bold stars indicate C a F M A S C H singularities; open stars show the location of CaMASCH singularities. For abbreviations of the phases see Table 1 and for the labelling of the invariant points, Table 3

350 f r o m Mg2.s7Si2Os(OH)3.74 at very l o w t e m p e r a t u r e s to Mg:z.79Si2Os(OH)3.57 at its b r e a k d o w n t e m p e r a t u r e n e a r 550 ~ C. A n ideal m i x i n g - o n - s i t e s a p p r o a c h using the a b ove two c o m p o s i t i o n s yields activities of c h r y s o t i l e of 0.67-0.51. A s e c o n d m e t h o d of d e t e r m i n i n g a s u i t a b l e c h r y s o t i l e a c t i v i t y in a n t i g o r i t e is via the e x p e r i m e n t a l d a t a of E v a n s et al. (1976) o n the r e a c t i o n a n t = f o + t a + H 2 0 ; the t h e r m o d y n a m i c d a t a for c h r y s o t i l e c a n rep r o d u c e the e x p e r i m e n t a l P - - T d a t a o f E v a n s et al. (1976) o n l y if a n a c t i v i t y in the r a n g e 0.4 0.5 is used. A s Mellini et al. (1987) p o i n t e d out, a l t h o u g h the a n t i g o rite p o l y s o m a t i s m d i s p l a y e d in n a t u r a l a s s e m b l a g e s is r a r e l y in a n e q u i l i b r i u m state, the h i g h e r t e m p e r a t u r e o c c u r r e n c e s in well-crystallised m a t e r i a l u s u a l l y a p p r o a c h c o m p o s i t i o n s close to Mg2.82Si2Os(OH)3.65 ( e q u i v a l e n t to a c h r y s o t i l e a c t i v i t y of 0.58). U n t i l f u r t h e r i n f o r m a t i o n b e c o m e s a v a i l a b l e it seems r e a s o n a b l e to a c c e p t a n i n t e r m e d i a t e value o f 0.5 as the a p p r o p r i a t e c h r y s o t i l e a c t i v i t y to use in c a l c u l a t i o n s i n v o l v i n g a n t i g o rite. Thus, in Fig. 1, a n t i g o r i t e - b e a r i n g r e a c t i o n s a r e o n l y a p p r o x i m a t e l y p l a c e d ; c o r r e c t l y - p l a c e d e q u i l i b r i u m reactions s h o u l d i n v o l v e a n t i g o r i t e w h o s e c o m p o s i t i o n varies a l o n g the r e a c t i o n lines. H o w e v e r , the a p p r o a c h t a k e n d o e s a l l o w the t o p o l o g y of a n t i g o r i t e - i n v o l v i n g equilib r i a to be assessed.

rit that regions of very low and very high Xco2 values are considerably expanded. In order to portray the phase relationships among the individual phases a projection from chlorite was chosen since chlorite is the most ubiquitous Al-phase in ultramafic rocks under the conditions chosen (e.g. Evans 1977, 1982; Winkler 1979; Gole et al. 1987; Laird 1988). As pointed out by Thompson (1979) and Powell and Sandiford (1988), there is no problem in projecting phases from a phase with a variable composition, in this case chlorite, as long as the projection plane lies between the chlorite and any variable composition phases to be projected. In C a O - M g O --A1203--8iO2, such a projection plane should be between the range of possible chlorite compositions and the range of compositions of talc, orthopyroxene and the other variable composition phases. For simplicity of perception, instead of choosing such a plane, the phases are projected from chlorite onto C a O - M g O - S i O z ; it was established that no difficulty was introduced by this. As an example, a calculated projection from chlorite onto C a O - MgO-- SiOz of the assemblage amphibole-clinopyroxenetalc-quartz-calcite(-chlorite-fluid) at 500~ and Xco:=0.5 is presented in Fig. 2. Obviously, due to the incorporation of A120 3, amphibole, clinopyroxene and talc do not each plot as a single point, as, for example, in diagrams presented by Winkler (1979), but each as a trivariant field. Note that the composition of the projecting chlorite varies across diagrams such as Fig. 2. SiO2

+ chlorite tr

'C

Ultramafic mineral assemblages at low pressures: CaMASCH + chlorite + fluid The method of calculating the mineral equilibria is identical to the one described in Will et al. (1990). Briefly, to determine the T - X c o ~ location of an equilibrium, 0 = AGO+ R T In K is written for each reaction in a set of independent reactions between the end-members of the phases involved in the equilibrium, forming a set of non-linear simultaneous equations. The set of simultaneous equations can then be solved for the unknowns in the equations. For an isobaric univariant reaction at fixed Xco~, for example, the simultaneous equations are solved for T and the compositions of the phases involved in the equilibrium. in contrast to Will et al. (1990), H20 and CO2 mixing is treated in a manner slightly modified from that of Powell and Holland (1985). We retain the sub-regular type of model fitted to the Modified Redlich Kwong activities of Kerrick and Jacobs (1981), bllt the parameters have the following P - T dependence:

cc

mag CaO dol MgO Fig. 2. Calculated projection from chlorite and fluid onto CaO -MgO-SiQ of the mineral assemblages at T=500~ and Xco2 =0.50. Trivariant fields are filled; divariant fields are ,shaded. Note that clinopyroxene plots outside of the C a O - M g O - S i O 2 diagram (see Fig. 3) rnl

SiO2

Wco~(kJ tool- 1) = (13.2 - 0.29 I/T) po.25

Wmo (kJ tool- ~)= ( 7 . 0 - 0 . 1 5 1 ~ ) po.z5 where T is in K and P in kbar. This form was chosen to provide smooth behaviour over a wide range of P - T conditions between 200-2000 ~ C and 0 5 0 kbar. There is no guarantee that the MRK activities on which our approach is based are correct, hut they have been shown to yield reasonable results (Powell and Holland 1988) in the 1-6 kbar range. We have found the equations above to be superior to those used by Holland and Powelt (1985) in terms of agreement with the phase-equilibrium data used in extracting our thermodynamic dataset. This change in mixing parameterisation is responsible for the very small shift in the T - X e o 2 location of the invariant point m2 which is stable in both the metabasic (Will et al. 1990) and in the ultramafic systems (this study-). A 2 kbar isobaric T - X c o 2 section has been calculated in the system CaMASCH for the temperature range of 440 600 ~ C with chlorite and HzO--CO2 fluid in excess (Fig. 1). Note that the horizontal scale is logarithmic, symmetrical about Xco 2= 0.5; this has the me-

o.o

/

\

\

cc e" ~_ mag CaO MgO Fig. 3. Projection from chlorite and fluid onto CaO - M g O - SiO2 of co-existing clinopyroxene-amphibole pairs along the Idol] reaction emanating from m 1. The numbers next to the projected dinopyroxenes correspond to the Xco 2 values at which the projection was calculated. Clinopyroxene is co-linear with quartz and calcite at an Xco 2 of 0.3, hence creating a CaMASCH singularity

351 cpx

+ chlorite + quaaz + fluid

I MgO

CaO ] amph

cpx amph cc q C X

cc T (cO) [ _ _

x(CO 2)

amph

cpx

amph

cpx

CaO --

.~

cpx

cc

amph/

cc

T (~ [ _ _

t

CC

cc

m2

' /

,

~

--

dol

+ quartz

ta

r

,

,

/',

!

+ fluid

~'INL'

,

m32"% I: ,

I

I

I

+ chlori~e

MgO ~

Fig. 4. Sketch showing the T-Xco~ relationships around the CaMASCH singularity (open slar) of Fig. 3. The singularity reaction is q + cc = cpx( + chl + fluid) (top). Qualitative T Ca diagrams projected from chlorite, Ca+Mg quartz and fluid show the effects of the varying clinopyroxene compositions (bottom)

m3 /

I

r~

i P

dok.mp ta\ --"

I

amph

" _o

,-

c!ccdol ta

cpx cc

dol

:'\m; -o

ta

cpx cc

dol

ta

cpx

dol

h

mag

cpx

dol

mag

Fig. 5. Phase relationships among the phases involved in the high Xco 2 CaMASCH singularities on Fig. 1 chlorite, quartz and fluid in excess (bottom)

Considering the composition of clinopyroxene in the assemblage along the [dol] reaction emanating from m 1 in Fig. 1, clinopyroxene plots outside of the C a O - M g O - S i O 2 triangle for high values of Xco 2, Fig. 3. With decrcasing temperature and for more H20-rich fluid compositions the composition of clinopyroxene swings towards the C a O - S i O 2 join and finally, at T = 482 ~ C and Xco2 = 0.3, it passes through the quartz-calcite join, hence generating a CaMASCH singularity. Along the reaction, amphibole occurs as a product for X co2 > 0.3, and clinopyroxene plots outside of the C a O - - M g O - - S i O 2 triangle, whereas, at Xco2'S lower than 0.3, clinopyroxene occurs within the C a O - MgO--SiO2 plane, and amphibole is a reactant. Therefore amphibole changes sides in the

(top); T - x diagrams with

reaction; the reaction is a m p h + q + c c = c p x + C O / + H 2 0 (+chl) at low Xco2, whereas at high Xco 2 the reaction is q + c c = amph + cpx + CO2 + HzO(+ chl). Clearly, the reaction involved in this singularity is c h l + q + c c = c p x + C O 2 + IlzO. Figure 4 is a Ca series of qualitative T vs ~ diagrams (projected from chlot~a + • rite, quartz and fluid) showing the prograde effect of the clinopyroxene compositions moving toward the amphibole composition. Along the reactions joining m l and m 11, and m7 and m8 (Fig. 1), clinopyroxene compositions also swing through the calcite-quartz join into the C a O - M g O - S i O 2 triangle; however, since calcite and quartz are not involved in these reactions, no singularities

352 T(~ 590

11

m8~ , ~

rnll ~

O8

m9 19~

570

9

ml0 O ' ~

Comparison with natural assemblages

10

550

9 m7 (~m12

530

GI Li

510 490

m14 (~

470

m13

m15 ml 22

m3

m5 m4 9

450 I

I

I

I

I

I

I

I

of anthophyllite assemblages in CaMASCH is not possible at this stage, a qualitative attempt based on reported anthophyllite assemblages has been made and will be discussed'in the final section of this paper.

I

I

I

0.0 0.05 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 0.95 1.0 x(CO2) Fig. 6. A 2 kbar T-- Xco2projection showing the effect of the introduction of AlzO3 into the system CaMSCH. Open circles show the T-Xco2 locations of the CaMSCH invariant points, filled circles correspond to their respective CaMASCH positions. Open circles around filled circles indicate that the CaMASCH invariant points are superimposed on the CaMSCH equilibria. Note that the CaMASCH traces of the CaMSCH invariant points 17 and 18 intersect and create the CaMASCH invariant point m5. Invariant points 17 and 18 become metastable in CaMASCH and the stabilisation of m4 and m7 is caused by the presence of m5. The anthophyllite-bearing CaMSCH invariant points 19 and 20 cannot be calculated in CaMASCH due to the lack of thermodynamic data occur along them. At high Xco2, however, there are two more CaMASCH singularities present which occur along the reactions joining the invariant points m2 and m3, and ml and m4 (Fig. 1). These two singularities are caused by amphibole moving across the dolomite-quartz join in the CaO-MgO-SiO2 triangle with changing conditions, thus creating the singularity reaction dol+ q = amph(+ chl-fluid) which connects both singularities. An onCa largement of this region of Fig. 1 and the T vs ~ relationships are shown in a projection from chlorite, quartz, and fluid in Fig. 5. Figure 6 shows the T-Xco2 locations of the invariant points in the AtzO3-freeCaMSCH sub-system. Due to a negligible amount of Tschermak's substitution the CaMASCH invariant points m 12 m16 are superimposed on the CaMSCH sub-system invariant points. For the same reason, at low Xco2, the CaMASCH univariant reactions emanating from these invariant points coincide with the CaMSCH reactions. The invariant points m13-m16 govern the stability relations among antigorite-bcaring assemblages; these equilibria occur at some 50~ above corresponding metastable equilibria involving chrysotile. When A1203 is added, an inversion of topology occurs at high values of Xco2 destabilising some of the CaMSCH invariant points and stabilising some new CaMASCH invariant points. Due to the lack of thermodynamic data for ahiminous anthophyllite, anthophyllite had to bc neglected in the CaMASCH calculations. Even though a quantitative treatment

An ideal prograde sequence for contact-metamorphosed Alpine-type metaserpentinites and metaperidoties is antgfo-cpx, antg-fo-amph, fo-amph-ta, fo-amph-anth and foamph-opx with chlorite in excess and has been discussed by Evans (1977). Apparently, there is no difference between the sequence of mineral assemblages found in contact-metamorphosed and regionally metamorphosed ultramafic bodies in the Alps (Evans 1977). A very similar sequence has been described from contact-metamorphosed ultramafic rocks in the Central Cascades of Washington (Frost 1975) and also from metamorphosed komatiites in Western Australia (Gole et al. 1987). This sequence of assemblages is accounted for in the next section.

Pseudosections and buffering paths At fixed pressures, the dependence of mineral assemblage on bulk composition with changing T--Xco~ is shown best by the use of T--Xco~ pseudosections (e.g. Hensen 1971). These T--Xco2 pseudosections are diagrams drawn for one bulk composition showing just those equilibria "seen", at 2 kbar, by this particular bulk composition. Figure 7 shows a series of calculated pseudosections for different bulk compositions which will commonly be encountered in ultramafic rocks. The bulk composition of a rock is of outmost importance with respect to the mineral assemblages developed and observed in that particular rock. In contrast to politic and marie rocks, even small changes in the SiO2/(SiO2+ MgO) ratio, without any changes in the P - Tconditions, wilt lead to different mineral assemblages in ultramafic rocks in the field. This is because this ratio differs by only some 25% along the quartz-magnesite join on the compatibility diagram between the most siliceous phase, talc, and the least siliceous phase, forsterite. T - X c o 2 pseudosections are the best way to read petrogenetic grids and to understand metamorphic rocks when buffering paths are taken into account. The two limiting cases are external and internal buffering. In the case of pure external buffering (e.g. involving infiltration processes from an external reservoir), the fluid composition is completely controlled by the infiltrating fluid, and the reaction sequence will follow a vertical path on a T--Xco 2 diagram at the Xco2 of the infiltrating fluid. In the case of internal buffering, the mineral assemblage, and the reactions that occur in the assemblage with changing P - T , control the composition of the fluid (Greenwood 1975). With an increase in temperature, for internal buffering, any assemblage starting out on a univariant curve will stay on it with trivial reaction until it either arrives at an intersection or a maximum on the curve. In both cases, substantial reaction will occur only there, continuing until a phase is reacted out: at

353

T(~ I

T(~ 590

59O

570

570

550

550

530

530

510

510

490

490

470

470

450

450 0.0 0.05 0.1

0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 0.95

x(CO2) T(~ 590 570

550

530

510 490

470

450 0.0 0.05 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 0.95 e x(CO2) Fig. 7a-c. T--Xeo2 pseudosections for bulk compositions A, B, C in CaMASCH

0.0 0.05 0.1

0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 0.95 x(CO2)

an intersection, the assemblage leaves along another univariant curve; at a maximum, buffering capacity is lost and the assemblage leaves to higher T at the Xco2 of the maximum. For example, consider the internally-buffered reaction path the bulk composition " A " (Fig. 7a) will experience with prograde metamorphism, focussing on the silicates, but noting that carbonates need to be present for buffering capacity to be maintained; calcite is the carbonate present originally. At low to moderate temperatures and very low Xco 2 a rock of this bulk composition will be characterised by the assemblage fo-cpxantg(+chl) (Fig. 7a). With rising temperature, this assemblage reaches the invariant point m 16, clinopyroxene reacts out and the assemblage leaving this invariant point is fo-amph-antg; then, across m15, calcite is replaced by dolomite; until, at m14, antigorite is finally consumed, and talc appears. After antigorite has reacted out, the assemblage fo-ta-amph(+ chl) will remain stable across a wide range of fluid compositions. This is because the invariant point m 12 involves dolomite being replaced by magnesite, and also because there are no maxima along the T - X c o 2 univariant line that this assemblage has to interact with. At an Xco~ value of probably not much less than 0.77 (m 10), this assemblage reaches the univariant reaction fo + ta = anth + H20 (Fig. 9), and the assemblage departing from this reaction will be fo-amphanth(+ chl). This sequence of mineral assemblages produced by internal buffering corresponds to the prograde "critical paragenesis" in contact-metamorphosed ultramafic rocks observed in the field (e.g. Trommsdorff and

354 Evans 1974; Irving and Ashley 1976; Evans 1977 and references therein). Ultramafic rocks described from the Inner Piedmont belt of South Carolina (Warner et al. 1989) show the same sequence; the reported presence of dolomite as part of the peak-metamorphic assemblage is easily accounted for by a slight shift in bulk composition (e.g. to bulk compositions "B" or "C" on Fig. 7b and c). Evans (1977) states that his zone 3 fo-taamph(+chl) - is "typically quite broad" for contactmetamorphosed rocks; this coincides extremely well with the wide range of Xco2'S (~0.06-0.75) for which this mineral assemblage is stable for the given bulk composition on our calculated grid. Evans' (1977) highest-temperature assemblage, fo-opx-amph(+ chl), is not "seen" by our chosen bulk composition; however, only a minor variation in the SiOz/(SiOz+MgO) ratio of the bulk composition can lead to the stabilisation of this assemblage, at high temperatures and high Xco2. Similar silicate sequences may be produced in rocks in which carbonate was never present or in which carbonate has reacted out at low to intermediate Xco~. Clearly, without information on the Xco~ of metamorphic fluids, distinguishing between this case and the above case involves careful evaluation of buffering capacity, in relation to the role of carbonates in the assemblages, and the actual variance of assemblages. The invariant point m2 links the grid in Fig. 1 to the metabasic grid of Will et al. (1990). In terms of bulk composition, however, invariant point m2 will only very rarely be seen in ultramafic rocks, and never in metaserpentinites, requiring a very calcic bulk composition to stabilise the assemblages ta-dol-cc, ta-dol-amph or amph-dol-cc(all+chl-fluid). The mineral assemblages around invariant points ml and m3 m7 are even less accessible and will be found in exceptional circumstances only. This is partly because fluid compositions of Xco2 > 0.85 are required and partly because of the very calcic bulk compositions necessary for most of the univariant reactions to be observed. Furthermore, bearing in mind that an internally-buffered assemblage will exhaust its buffer capacity and leave its univariant line as soon as a T--Xco~ maximum is approached, assemblages with an intermediate fluid composition buffered up temperature along the [q], [ta] or [cc] reactions from m2 will not reach the invariant points, m 1 or m 3. Instead, they will leave the reactions near their maxima and follow a constant Xco~ path up T until other univariant equilibria are intersected at higher temperatures. As a curiosity, an internal buffering path for a calcic rock with an initial fluid composition of Xco~ of about 0.40 will be almost vertical on the T--Xco ~ section and will resemble an externally buffered path because the maxima of the relevant univariant reactions occur very close to Xco 2= 0.40 over a temperature range of some 80~ C. Changes in mineral composition with prograde metamorphism

The progressive changes in Tschermak's substitution (MgSiAI_ 1A1_ 1) for amphibole, chlorite, clinopyroxene, orthopyroxcne and talc in CaMASCH are shown in

Fig. 8. As discussed above, the mineral assemblages around the invariant points m l m 7 will only be sampled in exceptional circumstances and are therefore omitted from the discussion for clarity. The compositions of the minerals at the invariant points are given in Table 3. Low-T low-Xco~ chlorites (Fig. 8a) tend to be extremely sub-aluminous. Laird (1988) states that "chlorite in low-grade ultramafic rocks (below enstatite-in) shows less Tschermak's substitution than clinochlore". With an increase in grade the chlorites become more aluminous (e.g. Pinsent and Hirst 1977) and their compositions lie between clinochlore and amesite. At the invariant point governing the first appearance of orthopyroxene in prograde ultramafics, m 10, the composition of chlorite is y=0.46; this is just slightly less aluminous than the ideal clinochlore composition of y=0.5. Our calculated transition from low-grade sub-aluminous to higher grade "aluminous" chlorites matches the behaviour of naturally-occurring chlorites in ultrabasic rocks (Laird 1988, Fig. 24). Low-grade amphiboles (Fig. 8 b) involved in ophicarbonate equilibria around our invariant points m iZPm 16 are pure tremolites. This corresponds to amphiboles from contact-metamorphosed antigorite-schists around the Bergell pluton in the Alps (Trommsdorff and Evans 1972). Amphiboles involved in the [antg] reactions emanating from m 15 and m 16 increase slightly in A1203 towards invariant point m11. A marked increase in Tschermak's substitution occurs towards the high-Xco2 orthopyroxene-bearing equilibria around m8 and m9. However, the amphiboles never become more hornblende-rich than about y = 0.2 (y meaning Tschermak's substitution). Our calculated sequence from almost pure tremolite in the greenschist and lower amphibolite facies to amphiboles with about y=0.2 towards the middle amphibolite facies coincides extremely well with natural amphibole compositions as summarised by Evans (1977). Clinopyroxenes (Fig. 8 c) involved in assemblages associated with the low-T, low-Xco2 invariant point m16 are very close to the pure diopside end-member. Only at temperatures above invariant point m 11 do the clinopyroxenes start to become more aluminous, but they never exceed y=0.2, except for the clinopyroxenes of two-pyroxene-bearing assemblages emanating from invariant point m8. This corresponds well with natural clinopyroxenes (e.g. Evans 1977 and references therein); tow-grade clinopyroxenes are always close to the diopside end-member composition whereas the higher-T, high-Xco~ clinopyroxenes are considerably more aluminous. The first appearance of orthopyroxene, in an average ultramafic bulk composition (e.g. the bulk compositions chosen on Fig. 7), occurs at temperatures above 560 ~ C and at high Xco2'S (m t0). The first orthopyroxene is still very close to the enstatite end-member composition (Fig. 8 d). With an increase in temperature the orthopyroxenes become more aluminous and in two-pyroxene-bearing assemblages the extent of the Tschermak's substitution in orthopyroxenes increases up to y=0.4 at temperatures around 600 ~ C. Talc is very Al-poor and its composition is always close to the AlaO3-free endmember (Fig. 8e).

355

T(~

59o cm~

,n~

570 55O

I

5v0 ~ ' Q

, , - ~

\

amphiboles

_

m

400 L/

530 510 490 470 45O 0.1

0.2

0.3

0.4

0.5

0.6

T(~

0.7

I

590

0.8

0.9 x AI.M2

0.1

0.2

/'.~'

~0~/, clinopyroxenes

550

55o

530

530

51o

510

490

490

470

470 Ira5

-

I 0.1

/I 0.2

I

I

I

1

0.3

0.4

0.5

0.6

,--,, 0.7

450

I

I

0.8

0.9 XA1,M2

0.5

0.6

/\'

'

8

0.7

0.8

0.9 x AI,M2

'

'

'

orthopyroxenes

~9

570

,450

0.4

h

I

570

0.3

F

i

i

I

I

i

0.1

0.2

0.3

0.4

0.5

m>.. 0.6

0.7

I

I

0.8

0.9 x AI,M2

d

T(~

I

59o

I

I

I

II ....

I

I

I

I

talc 57o /

mlO

550 53o

Fig. 8a-e. T-xa~ diagrams showing the changes in mineral

51o 490 m6

470 m15 450 0.1

0.2

0.3

0.4

0.5

I

I

I

I

0.6

0.7

0.8

0.9 x A1,M2

composition along the univariant reactions in Fig. 1. a Chlorites, b amphiboles, e clinopyroxenes, d orthopyroxenes, e talc. In a, an xA~,~12 of 0.50 corresponds to clinochlore and 1.0 to amesite, in b e , an XM,MZ of 0 corresponds to tremolite, diopside, enstatite and talc, respectively, and 1.0 to hornblende, Ca-Tschermak's, MgTschermak's and Tschermak's talc. Note that the equilibria around ml-m7 will only occur under exceptional circumstances (see text)

356 Table 3. P-T--Xco2 locations and mineral compositions of T -Xco2 invariant points ml (ta, fo, antg, mag, opx) T(~ x(CO2) y(amph) y(chl) y(cpx) 472 0.871 0,513 0.889 0.460 m2 (cpx, op• fo, mag, antg) T(~ x(COz) y(amph) y(ta) y(chl) 453 0.52 0,106 0.069 0.524 m3 (cpx, opx, fo, antg, cc) T(~ x(COz) y(amph) y(ta) y(chl) 468 0.927 0.412 0.302 0.851 m4 (opx, fo, ta, antg, cc) T(~ x(CO2) y(amph) y(chl) y(cpx) 454 0,966 0.724 0.943 0.665 m5 (fo, ta, dol, antg, cc) r(*c) x(COz) y(amph) y(chl) y(cpx) y(opx) 458 0.972 0.763 0 . 9 5 0 0.709 0.635 m6 (cpx, fo, amph, dol, antg, cc) r(~ x(COz) y(ta) 473 0.964 0.507

y(chl) y(opx) 0.923 0.506

m7 fro, ta, q, antg, cc) T(~ x ( C O z ) y(amph) y(chl) y(cpx) y(opx) 549 0.902 0.369 0.819 0,348 0.291 m8 (ta, mag, q,

T(~ 580

antg, cc) x(COz) y(amph) y(chl) y(cpx) y(opx) 0.833 0.231 0.707 0.223 0.183

m9 (cpx, ta, q, antg, cc) T(~ x(C02) 573 0.83

y(amph) y(chl) y(opx) 0.203 0.675 0,158

ml0 (cpx, amph, dol, q, antg, cc) T(~ x(COz) y(ta) 561 0.77 0,069

In general, our calculated changes in mineral compositions correspond very well with compositional changes described from natural assemblages (e.g. Evans 1977; Laird 1988; and references in both).

Comparison with ultramafic rocks other than serpentinites and peridotites Ophicarbonate rocks O.e. rocks bearing serpentine and carbonate minerals) are often developed in ophiolite belts (Trommsdorff and Evans 1977). The mineral equilibria around the invariant points m 13-m 16 (Fig. 1) control the mineral assemblages in these rocks. The topology of our grid is identical to that of Trommsdorff and Evans (1977) except that their invariant point I-talc, forsterite] is not present in our grid, and our equilibria occur at slightly lower T and at lower Xco~'S. These differences are due to the use of different thermodynamic data especially those related to the non-ideal mixing of the H 2 0 - CO2 fluid. The assemblage, dolomite + diopside, which is not stable in Fig. 1, would appear if the Xco2 of m16 was greater than the Xco 2 of m15; this is within error of being correct because, for m15, the calculated uncertainty (at 2a) on Tis +_ 14 ~ C and on Xco2 is 0.008, and, for m I6, on T is _+12~ and on Xco~ is 0.006, and the positions of m 15 and m 16 overlap within error. Sagvandites

(i.e. carbonate-orthopyroxenites) as described, for example, by Schreyer et al. (1972), Ohnmacht (1974) and Evans and Trommsdorff (1974) are stabilised at Xco~'S>0.77 and T > 5 6 0 ~ C at 2 kbar. When anthophyllite-bearing assemblages are considered, sagvandites will probably form at even higher temperatures and more CO2-rich fluid compositions (cf. Fig. 9).

y(chl) y(opx) 0.455 0.067

Rodingites (i.e. metasomatised calc-silicate-rich rocks)

mll (opx, ta, mag, q, antg) T(~ x(CO2) y(amph) y(chl) y(cpx) 578 0.31 0.022 0.162 0.021

have been described from all the major alpine ultramafic complexes (Evans 1977) but are, because of their domi nant calc-silicate assemblages, outside the scope of this paper. The interested reader is referred to a detailed discussion by Rice (1983).

m12 (cpx, opx, q, antg, cc) T(~ x{COz) y(amph) y(ta) 533 0.291 0 0

y(chI)

m13 (cpx, opx, amph, dol, q, cc) {for a(chr)=0.5} T(~ x(CO)2 y(chl) 483 0.071 m14 (cpx, opx, mag, q, cc) {for a(chr) =0.5} T(~ x(CO2) y(amph) y(ta) y(chl) 484 0.064 0 0 m15 (cpx, opx, ta, mag, q) {for a(chr)=0.5} T(~ x(CO2) y(amph) y(chl) 474 0.021 0 m16 (opx, ta, mag, dol, q) {for a(chr)=0.5} T(~ x(CO2) y(amph) y(cpx) y(chl) 471 0.0138 0 0 y is XA~,~2in amph, chl, cpx, opx, and ta. The numbers ml-ml6 correspond to the invariant points given in the figures. The list of phases following the invariant point label are the phases not involved at the invariant point

Discussion

Anthophyllite is a c o m m o n phase in the middle and upper amphibolite facies of ultramafic rocks. However, due to the lack of thermodynamic data for aluminous anthophyllite it had to be excluded from our calculations for CaMASCH. Based on the C a M S C H locations of invariant points involving anthophyllite (Fig. 6) and reported natural assemblages, a qualitative petrogenetic grid has been constructed to account for anthophyllilebearing assemblages (Fig. 9). The grid has been generaled based on the following premises: (a) the reaction opx+ ta = a n t h ( + chl + fluid) has a negative d T/dXco 2 slope (due to the presence of chlorite and H20), intersects the mag + ta = o p x ( + chl + fluid) and ta = opx + q ( + chl + fluid) reactions and, therefore, establishes the invariant points m17 and m18 (Fig. 9); (b) the reaction fo+ t a = a n t h ( + c h l + f l u i d ) has a negative slope; (c)talc reacts out before amphibole. For reference, Fig. 9 shows

357 0

metastabR; equilibria

opx

anthcpx

allap anth'~ m26

J

aml'h

mag

fo

Oml0

m20 m7

~.--m12

T(~

m17 ~m5 ma~

t~ x(C02)

~6

m5

Lm6

Fig. 9. Qualitative T - X c o 2 diagram including aluminous anthophyllite in CaMASCH at high temperatures and high Xco2. The open circles indicate the locations of CaMASCH invariant points that become metastablc when anthophyllite is included as an additional phase

the locations of the invariant points m 8 - m l 0 rendered metastable by the addition of anthophyllite. The low-grade assemblage fo-ta-amph( + chl) is internally buffered up temperature along the reaction mag+ ta = fo(+ chl) until it reaches the reaction fo + ta = a n t h ( + c h l ) at m21 (for bulk compositions " A " and " B " on Fig. 7a, b) and follows this reaction until talc has been consumed. The assemblage leaving this reaction is the two-amphibole-bearing assemblage anth-amph-fo. This is a common high-temperature assemblage in regional- and contact-metamorphosed ultramafic rocks (e.g. Springer 1974; Evans 1977; Matthes and Knauer 1981; Warner et al. 1989). In the case of bulk composition " C " (see Fig. 7c), the assemblage buffered up temperature along the reaction amph + mag = fo + dol(+ chl) is amph-fo-dol(+chl). At invariant point m23 (Fig. 9) the tie-line between amphibole and forsterite breaks down and the divariant assemblage leaving this invariant point will be anth-dol-fo(+chl). The presence of dolomite as part of the peak-metamorphic assemblage has been described by Warner et al. (1989). The highest temperature assemblage of Matthes and Knauer (1981) (opxta-anth-chl) is not found on our grid because it cannot be represented chemographically on the CMS projection.

Even though the incorporation of anthophyllite into CaMASCH cannot be quantified at the present stage we feel that Fig. 9 is a useful tool in understanding and documenting phase relationships involving anthophyllite. In Fig. 1, a full extension into the system CaFMASCH was avoided because of the limited MgFe_ 1 substitutions in the phases involved under the given conditions. However the C a F M A S C H traces of the CaMASCH invariant points have been calculated to show the effect of ferrous iron and are shown by the arrowheaded lines emanating from the C a M A S C H invariant points (Fig. 1). These C a F M A S C H reactions have been drawn so that the length of the arrows is proportional to the amount of ferrous iron incorported in the projecting chlorites, the arrow ends corresponding approximately to Xvo,chl=0.2. In general, the introduction of ferrous iron shifts equilibria to slightly lower temperatures but invariant points such as m5, m12 and all the equilibria involving antigorite do not change their T -Xco2 locations with respect to their sub-system positions. One complication is that the C a F M A S C H trace of invariant point m 10 is concavely shaped towards low Xco2'S and has a singularity occurring at an Xco2 of 0.85. The C a F M A S C H reaction emanating from this singularity is ta + mag = opx(+ chl + fluid). This reaction is sub-paralM to the ta + mag = opx(+ chl + fluid) CaMASCH sub-system reaction but occurs at slightly higher values of Xco2. This reaction links the high-temperture singularity along the C a F M A S C H reaction coming out of m 10 with another singularity occurring along the C a F M A S C H trace of the invariant point m6 at lower temperatures and higher Xco2'S (Fig. 1). The agreement between the calculated grid and natural mineral assemblages provides support for the approach followed in this paper. Further work, especially with respect to obtaining reliable thermodynamic data on aluminous ortho-amphiboles, is necessary in order to increase the applicability of the presented petrogenetic grid to middle-upper amphibolite facies anthophyllitebearing assemblages. Acknowledgements. We would like to thank Martin Engi and Ber-

nard Evans for helpful reviews. Thomas Will would like to acknowledge support from a Melbourne University Postgraduate Fellowship and the "Studienstiftung des Deutschen Volkes".

References

Berman EL (1988) Internally-consistent thermodynamic data for minerals in the system N a 2 0 - K 2 0 - C a O - M g O - F e O - - Fe20 3- - A 1 2 0 3 -- SiO2 - - TiO2- H 2 0 - CO2. J Petrol 29:445522 Evans BW (1977) Metamorphism of alpine peridotite and serpentinite. Ann Rev Earth Planet Sei 5:397--447 Evans BW (1982) Amphiboles in metamorphosed ultramafic rocks. In: Veblen DR, Ribbe PH (eds) Amphiboles: petrology and experimental phase relations. Mineral Soc Am Rev Mineral 9B:98 113 Evans BW, Guggenheim S (1988) Talc, pyrophyllite, and related minerals. In: Bailey SW (ed) Hydrous Phyllosilicates(exclusive of micas). Mineral Soc Am Rev Mineral 19:225~94

358 Evans BW, Trommsdorff V (1974) Stability of enstatite+ talc and CO2-metasomatism of metaperidotite, Val d'Efra, Lepontine Alps. Am J Sci 274:274-296 Evans BW, Johannes W, Otredoom WH, Trommsdorff V (1976) Stability of chrysotile and antigorite in the serpentine multisystern. Schweiz Mineral Petrogr Mitt 56:79-93 Frost BR (1975) Contact metamorphism of serpentinite, chloritic blackwall and rodingite at Paddy-Go-Easy Pass, Central Cascades, Washington. J Petrol 16:27~313 Gole M, Barnes S, Hill RET (1987) The role of fluids in the metamorphism of komatiites, Agnew Nickel deposit, Western Australia. Contrib Mineral Petrol 96:151-162 Greenwood HJ (1975) Thermodynamically valid projections of extensive phase relations. Am Mineral 60:1-8 Hensen BJ (1971) Theoretical phase relations involving cordierite and garnet in the system M g O - F e O - A 1 2 0 3 - S i O 2 . Contrib Mineral Petrol 31:191 214 Holland TJB, Powell R (1985) An internally consistent thermodynamic dataset with uncertainties and correlations: 2. Data and reslts. J Metamorphic Geol 3:343 370 Holland TJB, Powell R (1990) An enlarged and updated internally consistent thermodynamic dataset with uncertainties and correlations: the system KzO -- Na20 - CaO - M n O - FeO - Fe20 3 - A 1 2 0 3 - T i O 2 - S i O a - C - H z - O 2. J Metamorphic Geol 8:89 124 Irving AJ, Ashley PM (1976) Amphibole-olivine-spinel,cordieriteanthophyllite and related hornfelses associated with metamorphosed serpentinites in the Goobarragandra District, near Turout, New South Wales. J Geol Soc Aust 23:19 43 Kerrick DM, Jacobs GK (1981) A modified Redlich-Kwong equation for H20, CO2, and H 2 0 - C O 2 mixtures at elevated pressures and temperatures. Am J Sci 281:735 767 Laird J (1988) Chlorites: metamorphic petrology. In: Bailey SW (ed) Hydrous phyllosilicates (exclusive of micas). Mineral Soc Am Rev Mineral 19:405-453 Matthes S, Knauer E (1981) The phase petrology of the contact metamorphic serpentinites near Erbendorf, Oberpfalz, Bavaria. Neues Jahrb Mineral Abh 141 : 59 89 Mellini M, Trommsdorff V, Compagnoni R (1987) Antigorite polysomatism: behaviour during progressive metamorphism. Contrib Mineral Pctrol 97:147-155 Ohnmacht W (1974) Petrogenesis of carbonate-orthopyroxenites (sagvandites) and related rocks from Troms, Northern Norway. J Petrol 15:303 323 Pinsent RH, Hirst DM (1977) The metamorphism of the Blue River Ultramafic body, Cassiar, British Columbia, Canada J Petrol 18:567~594

Powell R, Holland TJB (1985) An internally consistent thermodynamic dataset with uncertainties and correlations. 1. Methods and a worked example. J Metamorphic Geol 3 :327-342 Powell R, Holland TJB (1988) An internally consistent dataset with uncertainties and correlations. 3. Applications to geobarometry, worked examples and a computer program. J Metamorphic Geol 6:173 204 Powell R, Sandiford M (1988) Sapphirine and spinel relationships in the system FeO--MgO--A1203--SiO2--TiOa--O2 in the presence of quartz and hypersthene. Contrib Mineral Petrol 98:64-71 Rice JM (1983) Metamorphism of rodingites: part I. Phase relations in a portion of the system C a O - M g O - A 1 2 0 3 - S i O 2 - C O 2 -HzO. Am J Sci 283A: 121-150 Schreyer W, Ohnmacht W, Mannchen J (1972) Carbonate-orthopyroxenites (sagvandites) from Troms, Northern Norway. Lithos 5:3458-64 Springer RK (1974) Contact metamorphosed ultramafic rocks in the Western Sierra Nevada Foothills, California. J Petrol 15:160-195 Thompson JB Jr (1979) The Tschermack substitution and reactions in pelitic schists. In: Zharikov VA, Fonarez Wl, Korikovskii SP (eds) Problems of physiochemical petrology, vol 1 (in Russian). Moscow Academy of Science, pp 146-159 Trommsdorff V, Evans BW (1972) Progressive metamorphism of antigorite schists in the Bergell Tonalite aureole (Italy). Am J Sci 272:423~437 Trommsdorff V, Evans BW (1974) Alpine metamorphism of peridotitic rocks. Schweiz Mineral Petrogr Mitt 54:333 352 Trommsdorff V, Evans BW (t977) Antigorite-Ophicarbonates : phasc relations in a portion of the system C a O - M g O - S i O 2 - H 2 0 - C O 2 . Contrib Mineral Petrol 60:39 56 Warner RD, Griffin VS, Steiner JC, Schmitt RA, Bryan JG (1989) Ultramafic chlorite-tremolite-olivine schists: three bodies from the Inner Piedmont belt, South Carolina. In: Mittwede SK, Stoddard EF (eds) Ultramafic rocks of the Appalachian Piedmont. Geol Soc Am Spec Publ 231:63-74 Will TM, Powell R, Holland TJB, Guiraud M (1990) Calculated greenschist facies mineral equilibria in the system C a O - F c O --MgO-A1203--SiO2--COz--HzO. Contrib Mineral Petrol 104:353-368 Winklcr HGE (1979) Petrogenesis of metamorphic rocks. Springer, Berlin Heidelberg New York

Editorial responsibility: V. Trommsdorff