American Mineralogist, Volume 71, pages 1343-1353, 1986

Stable-isotope geochemistry of high-grade metapelites from theCentral zone of the Limpopo belt

Mlm flunnNnn, T. Kunrrs Kvsrn, EulN G. NrsnnrDepartment of Geological Sciences, University of Saskatchewan, Saskatoon, Saskatchewan S7N 0W0, Canada

Ansrnlcr

The Central zone of the Limpopo belt, near Beitbridge, Zimbabwe, contains high-grademetamorphic rocks of Archean age. d'8O values of coexisting qtartz and garnet in meta-pelites yield apparent equilibration temperatures of 621-846C, whereas quartz-biotite andquartz-feldspar pairs give nearly concordant, although lower, equilibration temperaturesof 383-643"C. The quartz-garnet isotope temperatures agree with temperatures calculatedfrom thermometers based on cation partitioning. The apparent temperature disequilibriumamong the mineral phases can be attributed to variable diffrsion rates and closure tem-peratures wherein the quartz and garnet have retained their compositions since the closuretemperature ofgarnet and the feldspar and biotite reflect prograde events.

Whole-rock 6'80 values of metapelites, metabasites, and calc-silicates, when comparedto their probable precursors, suggest a relatively closed-system redistribution of isotopesamong the various lithologic units. The isotopic compositions of waters that are calculatedto be in equilibrium with the hydrous minerals present in these samples are similar tothose of most metamorphic fluids from other regionally metamorphosed terranes but maycontain a substantial meteoric water component.

INrnolucrtoN

The widespread exposure of Archean high-grade meta-morphic terranes provides an opportunity to examine theEarth's early continental crust. Mineral assemblages invarious rock types ofthese regions often retain a recordof the pressure and temperature prevalent at the time of

study was undertaken in order to compare cation andanion geothermometry and possibly reveal informationabout the processes involved in the evolution ofportionsof the continental crust.

RrcroNlr- cEoLocY

theirformation. Furthermore, patterns of chemical mod- The Limpopo belt lies between the Zimbabwe (Rho-ifications can provide information about the processes desian)andKaapvaalcratonsinsouthernAfricaandcon-operative in the early history of the Earth. The Central tains a variety of metamorphic rocks ranglng from earlyzone of the Limpopo belt, which straddles the Zimbabwe- to late Archean in age (van Breemen and Dodson, 1972;South Africa boundary, is an Archean metamorphic ter- Barton and Ryan, 19771, Barton et al., 1977). The marginalrane in which a variety of high-grade metamorphic rocks zones (Fig. la) are bounded by the orthopyroxene isogradis preserved. Various aspects ofthe regional geology and (Robertson, 1973) and contain a variety ofgranitoid in-petrology of this region have been studied by several re- trusions and supracrustal rocks, all of which have beensearchers (Chinner and Sweatman, 1968; Light et al., 1977; metamorphosed in the granulite facies. The mineral as-LightandWatkeys, 1978; Watkeys, 1979; Broderick, l9T91, semblage of orthopyroxene + quartz is characteristic ofHorrocks, 1980; Light, 1982; Harris and Holland, 1984; large areas of the maryinal zones. In contrast, the CentralWindley et al., 1984). These studies, together with radio- zone (Figs. la, lb) is a patchwork of granulite facies andmetric age results (Manton, 1968; van Breemen and Dod- upper-amphibolite facies rocks with assemblages of bio-son, 1972; Barton and Ryan, 1977; Barton et al., 1977) tite + quartz and amphibole * qluartz. Supracrustal rockssuggest a long and varied metamorphic evolution. Esti- comprise a significant proportion of the Central zone ter-mates of the equilibration pressures and temperatures of rane relative to the marginal zones. The rocks of the Cen-samples from the Limpopo belt have been based on the tral zone have undergone at least two major metamorphicpartitioning of cations among various mineral phases. events, the first in the granulite facies and the second, aKnowledge of the apparent equilibration temperatures of retrograde event, in the upper-amphibolite facies. In ad-various phases is critical to understanding the metamor- dition, the Central zone records several periods of defor-phic evolution of this area. In conjunction with a recent mation and several lesser retrograde metamorphic events.study of the cation thermobarometry of a series of me- The oldest rocks in the Beitbridge area, the Macuvilletapelites from the Beitbridge region of the Central zone Group, are a series of migmatites and banded gneisses.(Harris and Holland, 1984), a companion stable-isotope They have been correlated by Watkeys (1979) with the

0003404x/86/ I I I 2-1 343S02.00 t343

t344

Sand River gneisses, which have been dated at 3.78 Ga(Barton eI al., 1977). The Macuville Group basementgneisses are unconformably overlain by the BeitbridgeSequence, comprising the lower Diti Metamorphic Suiteand the overlying Nulli Metamorphic Suite. These twosuites contain a lithologically diverse supracrustal assem-blage of metasedimentary and metavolcanic rocks.

A minimum age for the granulite-facies metamorphicevent in the Central zone is provided by the Messinalayered intrusion, which was emplaced into the BeitbridgeSequence around 3.27 Ga (Barton et al., 1977). Field re-lations and petrographic studies suggest that the upper-amphibolite-facies metamorphic event occurred contem-poraneously with the intrusion ofthe Bulai Gneiss around2.7 Ga (van Breemen and Dodson, 1972; Mason, 1973;Light and Watkeys, 1978). Two later, less intense meta-morphic events occurred in the middle-amphibolite andthe greenschist facies, but their ages are obscure (Tankardetal., 1982). After uplift and erosion of the metamorphiccomplex, the Soutpansberg Group, a sequence of sedi-ments and volcanics, was unconformably deposited dur-ing the middle Proterozoic (Barton and Ryan, 1977).I.a;t-er, the overlying Mesozoic Karoo sediments and Karoovolcanics were deposited while Karoo age intrusives wereemplaced.

GnorrmntyoMETRy AND GEoBARoMETRy oF THECBNrrul- zoNB

The results of several studies have been published onthe thermobarometry of rocks from the Central zone.Chinner and Sweatman (1968) deduced from the mineralassemblage of enstatite + cordierite + sillimanite +quartz (originally enstatite + kyanite + quartz), in a sam-ple near Beitbridge, that the granulite-facies metamorphicevent in the Central zone occurred at T > 860'C and atP > ll kbar. The upper-amphibolite-facies mineral as-semblages, which are best recorded in the metabasites,suggest equilibrium conditions of T :600-700"C andP: 5-7 kbar (Horrocks, 1980).

Most recently, Harris and Holland (1984) used the cat-ion distributions of Mg among phases in the MgO-AlrO,-SiO, system (garnet, spinel, cordierite, hypersthene) toargue that the equilibrium temperature of the mineralassemblages present in metapelites from the Beitbridgearea ofthe Central zone was in excess of670.C and thatthe pressure was between 4 and 5 kbar. They further con-cluded that these are minimum estimates of the equili-bration temperatures and pressures because the mineralassemblages display textural disequilibrium. Their appli-cations of other thermobarometers to the same rocks andto some intercalated metabasites yielded estimates of ap-parent equilibrium conditions of T :750 + 50"C andP: 3.5-5.0 kbar (Table l). The assemblages in thesesamples, therefore, are largely representative ofthe upper-amphibolite-facies event.

Pprnocn,tpHy AND cHEMrsrRy

The samples examined in this study are from the samplecollection studied by Harris and Holland (1984), which

HUEBNER ET AL.: STABLE-ISOTOPE GEOCHEMISTRY OF METAPELITES

O 2OO km

7- "o t " ,

sequences

l f D c r a n u t i t e - 9 r a d e m a r g i n a l z o n e s

E n c r a n i t a - g r e 6 n s t o n o t o r t a n € s

Fig. 1a. Map showing relationships of the Limpopo belt with

the Zimbabwe (Rhodesian) and Kaapvaal cratons (after Du Toit

et al., 1983); RC-Zimbabwe (Rhodesian) craton, KC-Ka-

apvaal craton, SMZ-Southern marginal zone, NMZ-Northern

marginal zone, CZ-Central zone, P-Palala shear zone, T-

Tuli-Sabi shear zone, B-Beitbridge, GD-Great Dyke.

were collected from localities within 15 km of the townof Beitbridge, Zimbabwe. They include metapelites andsemipelites containing the assemblage biotite + garnet +feldspar + cordierite + qtartz + orthopyroxene + spi-nel + sillimanite. Modal compositions of the samples aregiven in Table 2. Retrograde reactions of garnet, ortho-pyroxene, and sillimanite to form cordierite are apparentfrom reaction rims in several samples [i.e., Bblld(2),Bb I 3b, Bb20, Bb2 5 cl. Excluding retrograde reactions, mostofthe minerals appear to be unaltered by secondary al-teration, although in a few samples [i.e., Bbl ld(2), Bbl3b,

Fig. lb. Simplified geology of the Beitbridge region (afterWatkeys et a1., 1983); Mll-Messina layered intrusion; starsrepresent sample localities.

O 2 o k m

- l p o s t - K a r o o i n t r u s r v e s

Z r ^ r o o s u p e r g r o u p

I D B r t " i G n " ' . "

l- ""''o''on"on(/

' g o o r . L i m p o p o R i v e r

Table 1. Results of geothermometry for Beitbridge samples

Geothermometer rfc)Vetaoelites

Fe-Mg exchange, biotite-garnetFe-Mg exchange, cordierite-garnetAl distribution, garnet-orthopyroxeneGarnet + plagioclase + AlrSiOs + quartz assemblage

MetabasitesFe-Mg exchange, garnet-clinopyroxeneAl distribution, garnet-orthopyroxeneTwo pyroxene

Note: From Harris and Holland (1984).

Bbl6c, BblTal there is variable sericitization of the feld-spars and pinitization of the cordierite. The almandine-rich garnets often have thin reaction rims and can contain5-l5o/o inclusions consisting mainly of quartz, biotite,feldspar, and sillimanite. Biotites, showing the red-browncolor normally associated with high-grade conditions, arealso unaltered with little or no chlorite present except insample Bbl7a. Some of the samples (i.e., Bbl ld, Bb25c)show myrmekitic intergrowths of quartz with feldspar andsymplectic intergrowths of quartz with cordierite indica-tive of retrograde reactions of earlier minerals with fluids.Also noted in one sample [Bblld(2)] are inclusions ofspinel and sillimanite in feldspar. The inclusions make upless than 50/o ofthe feldspar volume and are not consideredto affect the 6180 value ofthe feldspar separate.

AN,ll,r"rrc,q.L pRocEDURE

Mineral separates were prepared by use of healry liquids, mag-netic separation, etching with hydrofluoric acid, and hand pick-ing. The purity of the individual mineral separates was typicallygreater than 950/0. Oxygen was extracted from silicate mineralsand rock powders using the BrF, method of Clayton and Mayeda(1963). Yields for the oxygen extracted from garnets showed arange from 36 to 960/0, although extractions of the same samplesgiving various yields suggest that there is no correlation betweenyield and DIEO value of garnets. Many of the garnet samples werealso treated with HF to remove the outer portion of the grainsand were then ground and analyzed. These stripping experimentssuggest that the garnets are not zoned in r8O even though theyare chemically zoned (Harris and Holland, 1984). All values arereported in the familiar d notation in units of per mil relative toV-SMOW. The isotopic values of most minerals are reproducible

t345

to +o.l%u except for biotite and the whole-rock powders, wherethe values are accurate to +0.2Va. The value of NBS-28 quartzobtained by this procedure is +9.6Va. H was extracted fromhydrous minerals using the procedures outlined by Kyser andO'Neil (1984). The isotopic values for H, also reported in the dnotation relative to V-SMOW, are accurate to +3.0%oo. The valueof NBS-30 biotite obtained by this procedure is -65Vq.

The temperature dependence ofthe partitioning ofoxygen iso-topes among coexisting minerals in the temperature range ofmetamorphic rocks has been estimated from both laboratoryexperiments and empirical methods. Equations of the formAIEOu-r : A + (B x 106)/72 are normally used to describe therelation between temperatuie and the fractionation of oxygenisotopes between two minerals. AttO"-o is defined as 1000ln a.-o and, for relatively small fractionations, can be approx-imated by d'EO" - d'EOo (e.g., Friedman and O'Neil, I 977; Deines,1977), where a and b are two minerals whose isotopic compo-sitions have been measured, a is the fractionation factor betweentwo minerals, and,4 and B are coefrcients derived from eitherexperimental data or empirically from natural samples.

The use ofexperimentally derived fractionation factors is usu-ally preferable to the use of empirically derived curves to calculateequilibration temperatures of mineral phases because one mustassume that equilibrium was obtained in the samples used togenerate the empirical curve. For minerals that are ofinterest inthis study, however, published experimental results exist only forquartz (Matsuhisa et al., 1979), pyroxene and feldspar (Matthewset al., 1983), and muscovite, which can be used as an approxi-mation for biotite (Garlick and Epstein, 1967; Graham, l98l).The empirical isotope geothernometers of Bottinga and Javoy(197 3, 197 5)have been chosen for this study because they includea complete and internally consistent set of fractionation factorsfor all the minerals of interest.

Recent experimentally determined fractionation curves indi-cate that Bottinga and Javoy may have underestimated the amountofretrograde exchange that occurred in certain minerals used intheir empirical formulation which leads to excessively high equil-ibration temperatures. For example, apparent equilibration tem-peratnres for both qtrartz + biotite and qvartz + feldspar pairs inthe Beitbridge samples calculated using experimentally derivedfractionation factors are consistently lower than those calculatedfrom the factors ofBottinga and Javoy (1975) by about 140 deg(Table 4). Similarly, the quartz-garnet temperatures calculatedusing preliminary results from experiments by Matthews andGraham (pers. comm.) are 70 deg lower than those calculatedusing the empirical thermometers. Isotope equilibration tem-peratures of polymetamorphic rocks may not represent real tem-peratures but do indicate possible equilibrium among various

HUEBNER ET AL.: STABLE-ISOTOPE GEOCHEMISTRY OF METAPELITES

600-950<650740-820740-820

740-840700

800-975

Table 2. Modal compositions of metapelites from the Beitbridge area

Ksp opx Spn opaPlgsirGrtSample

B b 1 1 bBb1 1 d(1 )Bb1 1d(2)Bb12aBb13bBb1 3eBb1 6cBb17aBb20Bb25c

45403020< l20403040

1 51 5201 51 5

1

1 0201 5

51 01 035401 54020201 0

zc3020201 51 030251 0

< 1< 1 < 1

< 1 < l< 1< 1< 1< 1< 1

lc

255 2 5

5010 50

255 1 5

15 5050

< l< 1

< 1

t -C /r5

< l

5 1 55 2 5

Notej Compositions estimated from thin-section examination. qtz : quartz, Bio : biotite, Grtplagioclase, Ksp : potassium feldspar, Opx : orthopyroxene, Spn : spinel, Opa : opaques,determined by optical methods.

= garnet, Sil : sillimanite, Crd : cordierite, Plg :Pin : pinite, An : anorthite content ot feldsoars

1346

Table 3. d18O and 6D values of Beitbridge metapelites

Sample WR Qtz Fsp Bio 6D Gnt Crd Opx Sil

HUEBNER ET AL.: STABLE-ISOTOPE GEOCHEMISTRY OF METAPELITES

B b 1 1 bBb1 1 d(1 )Bb1 1d(2)Bb1 2aBb13bBb13eBb1 6cBb1 7aBb20Bb25c

Arotei WR : whole rock 0D values are those for biotite.

phases. As such, only the relative equilibration temperatures forvarious mineral pairs are critical, and these are similar for boththe empirical set of thermometers and those derived from ex-perimentally determined fractionation factors despite marked dif-ferences in their absolute temperatures.

The efect of pressure on the fractionation of oxygen isotopeshas to be considered for high-pressure metamorphic rocks. Ex-perimentally determined fractionation factors in the systemCaCOr-HrO indicate that oxygen-isotope partitioning is not af-fected to any significant extent by pressures up to at least 20 kbar(Clayton et al., I 975). This has been confirmed for silicate-bearingsystems by the experimental work of Matsuhisa et al. (1979) andMatthews et al. (1983). Therefore, high pressures shouldnotaffectthe oxygen-isotope fractionation between coexisting mineralphases in metamorphic rocks of upper amphibolite or granulitefacies.

Rnsur,rs AND DrscussloN

Whole-rock values

The range of whole-rock 6'8O values for the Beitbridgemetapelites is + 8.8 to +9.9 (Table 3). These compositionsare similar to those of other Archean metasedimentaryrocks (Longstaffe and Schwarcz, 1977;Fourcade and Ja-voy, 1973). The protoliths for metapelites are clastic sed-iments composed of detrital qvartz, feldspars, clays, androck fragments derived from a variety of igneous, meta-morphic, and sedimentary terranes. Mixed in with thesecomponents are authigenic and diagenetic minerals. Theisotopic compositions of the protoliths are a function ofthe isotopic compositions of the endmembers, with D'8Obeing * I 5 to + 30 for clay minerals (Savin and Epstein,I 970) and + 5 to + l0 for primary igneous rocks (Taylor,1968). Based on the chemical and mineralogic composi-tions, the original protolith for the metapelites probablyhad a 6180 value of +10 to +15.

Interlayered with the metapelites are calc-silicates(6'80: +6.7 to +8.9) and metabasites (D'8O: +7.5 to+8.7). Carbonates from the Archean of southern Africainclude those of the Fig Tree Group (>2.9 Ga), whichhave Dt8O isotopic compositions of + l2.l to * 16.6, andthose of the Shamavian Group (>2.5 Ga), which havevalues of +20.6 to +28.0 as reported by Perry and Tan(1972). The d'8O compositions of modern limestones areon the whole more positive with a range of +20 to +28(Keith and Weber, 1964;Yeizer and Hoefs, 1976). TheFig Tree Group carbonates have been metamorphosed to

Table 4. Apparent equilibration temperatures ("C) ofquartz-mineral pairs

Otz-Fsp Qtz-Fsp Qtz-Grt Qtz-Bio Qtz-BioSample ( 1 ) (2) (1) ( 1 ) (3)

B b 1 1 bBbl 1 d(1 )Bb1 1 d(2)Bb12aBbl3bBbl3eBb1 6cBbl 7aBb20Bb25c

/Votei Calculated using empirically and experimentally determined frac-tionation factors. (1) Bottinga and Javoy (1975); (2) Matsuhisa et al. (1979);(3) From approximation ot muscovite for biotite (Garlick and Epstein, 1967)and experimental quartz-muscovite curve of O'Neil and Taylor (1969).

lower-greenschist facies and together with the ShamvianGroup carbonates have undergone oxygen-isotope ex-change with either infiltrating metamorphic or meteoricwaters thereby lowering their d'8O values. Modern basaltshave a rather restricted range of dr8O values from *5.5to +7 .2 (Kyser et al., 1982). Comparison of the Beitbridgewhole-rock values with the assumed values for the pro-toliths (Fig. 2) implies that a redistribution of oxygenisotopes has occurred among the various lithologic units.Oxygen-isotope redistribution of this type suggests ex-change with eitherpervasive metamorphic water in a nearlyopen system or a closed-system redistribution ofoxygenvia fluid communication among the various lithologicunrts.

Dewatering of fluids from a sedimentary pile duringburial results in a decrease in porosity and permeability.Below a certain depth, the strength ofthe rock is no longerable to hold fractures open, and the fluid pressure canexceed the lithostatic pressure (Walther and Wood, 1984).Hydrofracturing is a proposed mechanism for dewateringthat occurs when the fluid pressure exceeds that of thelithostatic pressure and leads to the development of frac-tures in the rock. These fractures tend to cut across lith-ologic boundaries and provide pathways for unidirection-al fluid escape (Norris and Henley, 1 9 76; Fyfe et al., I 9 7 8).Rumble and Spear (1983) argued that it is via these chan-nelways that isotopic exchange can occur between lith-ologic units and suggested that the degree to which thelithologic uirits have reached equilibrium may be indi-cated by the isotopic composition of a common mineralsuch as quartz. d18O values of quartz from the metapelitesrange from 9.9 to 11.6, although the majority are neilr11.2, and are indicative of a reasonable approach to iso-topic equilibrium among the metapelite units (Table 3).In contrast, the D'8O values of quartz from all lithologicunits vary substantially from 8.5 to ll.6 (Huebner, un-pub.), which suggests that isotopic equilibrium was notobtained among all the lithologic units in the metamor-phic pile.

Walther and Orville (1982) discussed two endmembercases for fluid flow that have a bearing on the limits of

4.86.26.0o.z6 .1

6.16.26.4

8.89.29.2

9.9

9.6

9.68.9

9 9 8 01 1 3 9 . 211.4 9.311.2 9.21 1 . 6 9 . 21 1 . 5 9 11 0 . 6 I 711.4 8.91 1 . 1 9 . 210.2

8.9

7.9 8.1

-58-49

-61-64

- c 4

- a I

- 58- o J

o.o8.38.28.s8.67.98.38.88.87 0

562 398 683 532 392463 329 707 s39 392463 329 661 511 368460 325 750 539 400464 334 707 s05 360384 271 621538 381 846 561 427437 31s 779 518 375541 384 846 528 408

707 643 534

C a n a d i a n m e t a s e d i m e n t s - A t c h e a n-

A l g e r i a n m e t a s e d i m e n t s - 3 . 3 G a-

B e i t b r i d g e m e t a p e l i t e s-

t g n e o u s r o c k s-

C l a y m i n e r a l s

B e i t b r i d g e c a t c - s i l i c a t e sI

F i g T r e e G r o u p c a r b o n a t e s - ) 2 . 9 G a-

S h a m v i a n G r o u p c a r b o n a t e s - > 2 . 5 G a

M o d e r n c a r b o n a t e s

B e i l b r i d g e m e t a b a s i t e sI

M o d e r n b a s a l t s-

Fig. 2. Range in whole-rock drEO values of Beitbridge meta-morphic rocks, Canadian metasediments (Longstaffe andSchwarcz, 1977), Algerian metasediments (Fourcade and Javoy,1973), igneous rocks (Taylor, 1968), clay minerals (Savin andEpstein, 1 970), Fig Tree Group and Shamvian Group carbonates(Perry and Tan, I 972), recent carbonates (Keith and Weber, I 964;Veizer and Hoefs, I 976), and recent basalts (Kyser et al., 1 982).

fluid-rock interaction. In the first case, the fracture spacingis on the order of grain size; therefore, all rocks abovewhere the fluids originate will come in contact with theescaping volatiles and have a chance to react. The secondcase has the bulk ofthe fluids escaping in widely spacedfractures so that the bulk of the rocks overlying will notbe in contact with the escaping fluids. The episodic natureof fluid loss, the duration allowed for equilibration tooccur, the mechanism utilized for fluid esrcape and thefluid/rock ratio may preclude isotopic equilibration be-tween the lithologic units, as is recorded in the Beitbridgerocks. The progressive decrease in porosity and perme-ability with increasing depth of burial eventually mayhave restricted the intercommunication of fluids betweenthe lithologic units, leading to the development of local-ized mineral equilibrium and distinct d'8O values withineach unit.

Oxygen-isotope geothermometry

Isotopic data from Table 3 are plotted on an "isotherm"plot (Javoy et al., 1970) in Figure 3 to assess the degreeof relative isotopic equilibrium among the minerals ineach sample. The solid lines represent the theoretical equi-librium fractionation between quartz and other minerals

1347

o.o 1.0 2.O 3.0 4.O

BFig. 3. Isotherm plot for quartz-mineral pairs from the Beit-

bridge metapelites. The slope of the line represents the idealequilibration temperature of the mineral assemblage using theempirically derived fractionation factors ofBottinga and Javoy(1975). Temperatures for the same quartz-mineral pairs usingexperimentally derived fractionation factors would be lower by100-200qc.

at the temperatures of 500, 600, 800, and 1000"C usingthe empirical thermometers of Bottinga and Javoy (1975).As discussed previously, temperatures calculated usingexperimentally determined fractionation factors would besomewhat lower although the differences among the tem-peratures of each pair are nearly the same. The dashedlines represent data for individual samples of the Beit-bridge metapelites and connect several quartz-mineral pairsfor which apparent temperatures have been calculated.Equilibrium partitioning of oxygen isotopes between allmineral phases would result in a linear relation similar tothose represented by the reference isotherms. Apparentequilibration temperatures calculated from either empir-ical or experimental curves for quartz-biotite and quartz-feldspar are reasonably concordant for individual samples(Table 4), although the intersample variation in temper-atures is 220 deg. This is in contrast to the quartz-garnetpairs, which yield consistently higher apparent equilibra-tion temperatures.

The majority of apparent temperatures calculated from

HUEBNER ET AL.: STABLE-ISOTOPE GEOCHEMISTRY OF METAPELITES

EN

f

oo

20

6 t 8 0 % o

l 348

[1 / roK] x to .

Fig. 4. Diffirsion rate of rEO in various minerals plotted as afunction of temperature. I : feldspar (An*; Giletti et al., 1978),2 : feldspar (An,; Giletti et a1., 1978), 3 : feldspar (Ano; Freerand Dennis, 1982), 4 : P-qtar+Lz normal to c axis (Giletti andYund, 1984), 5: a-qrartz parallel to c axis (Giletti and Yund,1984), 6 : P,-quafiz parallel to c axis (Giletti and Yund, 1984),7 : garnet (Freer and Dennis, 1982), 8 : phlogopite (Giletti andAnderson, 1977). All rates measured using secondary ion massspectrometry except for phlogopite which was determined usingthe bulk-isotope method. On the right maryin of the diagram arethe times required for the oxygen isotopes in a l-mm-radiusspherical crystal to equilibrate with a fluid.

the quartz-garnet pair agree well with those calculatedfrom the cation geothermometers used by Harris and Hol-land (1984) and with other cation thermobarometers (Ta-ble l), whereas apparent isotope equilibration tempera-tures calculated from quartz-biotite and quartz-feldsparpairs are significantly lower. This lack of agreement be-tween some of the temperatures calculated from anion-based thermometers and those from cation geothermo-meters has been noted in previous studies of granulite-facies rocks (McNaughton and Wilson, 1980, 1983; Wil-son and Baksi, 1984).

Crystal gronth and diffusion effects onisotopic compositions

Isotopic disequilibrium may result from (l) noncon-temporaneous nucleation and growth of minerals during

HUEBNER ET AL.: STABLE-ISOTOPE GEOCHEMISTRY OF METAPELITES

Toc

o o o oN O @ 6

different metamorphic conditions or (2) contemporaneousnucleation and growth of minerals having different closuretemperatures ofisotope exchange so that some phases maycontinue to change their 180/'60 ratios during cooling andexhumation.

The timing associated with the crystallization of min-eral phases during metamorphism is potentially importantin the development of the 6'80 value of any mineral. Theparameters pressure, temperature, and fluid compositionvary continually during the metamorphic cycle, and there-fore minerals that nucleate and grow at different timesmay reflect those differences. Several stable-isotope stud-ies of polymetamorphic Alpine terranes (e.g., Frey et al.,1976; Hoernes and Friedrichsen, 1978) suggest that,through a judicious choice of samples, it is possible tocharacterize the conditions and isotope compositionsprevalent at the time of the growth of new minerals duringdiferent metamorphic events. From studies on samplescontaining minerals that obviously formed during differ-ent events, it is apparent that overprinting by later eventscan obscure earlier compositions either by recrystalliza-tion or by interaction with infiltrating fluids. In the Beit-bridge samples, the only obvious evidence for recrystal-lization or new mineral growth during later metamorphicevents is the presence of thin cordierite rims or the insig-nificant quantities of symplectic intergrowths present insome samples.

Contemporaneous crystallization of mineral phases mayresult in isotopic disequilibrium ifthe exchange rates be-tween various minerals and fluids are substantially dif-ferent. When a mineral passes through a temperature whereit ceases to exchange anions or cations, its compositionbecomes fixed. Parameters pertinent to oxygen-isotopeexchange that are available for the phases of interest inthis study are shown in Figure 4 and in Table 4 along withcalculated diffusion rates at 650'C. Most of the diffirsiondata have been obtained under hydrothermal conditionsand may not reflect conditions prevalent during high-grademetamorphism. Touret (197l), in a study of fluid inclu-sions in metamorphic rocks, has suggested that the fluidspresent during high-grade metamorphism are CO, rich.Muehlenbachs and Kushiro (1974) provided limited datafor oxygen-isotope diftrsion in rocks and minerals in the

Table 5. Diffusion rates of 18O in various minerals

T D O Q D(kJ/mol) (65eC)

5 .71 x 10 18

3.69 x 10 ,b

1 .61 x 1g - ' o2.'12 x 10-18g.g3 x 10 18

3.97 x 10 "--1 .32 x 10 , 1

4.5 x 10-zz2.77 x 1g-"'1 . 34 x 10 37

.References: (1)Giletti and Yund (1984), (2)Giletti et al. (1978), (3)Freer and Dennis (1982)'(4) Giletti and Anderson (1977), (5) Muehlenbachs and Kushiro (1974).

.- Diffusion measured at 873 K.

(m'ls)(K)

OuartzQuartzQuartzFeldspar (Ab,?)Feldspar (Anr6)Feldspar (Ano)PhlogopiteGrossularEnstatiteDiopside

1 073-8731 073-873873-773623-873673-1 073

873773-1073

1123-13231 5531 553

1 .0 x 10 -44.0 x 10- '1.9 x 10'?

2.31 x 10 '3

1 '39 x 10-"

1 .03 x 10 - ' o2.67 x 10 16

6 .0 t 10 t u

2 .4 x 10 16

23414228489

109

122102377377

11I22343J

tot g

Table 6. Calculated D18O and dD values ofwaters in equilibrium with hydrous

minerals from the Beitbridoesamoles

Sample r("c)

r349

",r. r. - ".*", u,r u,*ri*'L";r rr.:""",

", o.,n*

from the Beitbridge metapelites (O) and the field of values forwaters calculated to be in equilibrium with biotite (A) and horn-blende (B) from the Beitbridge samples. Also plotted are valuesof fluids calculated from the metamorphic teffanes of the InyoMountains (Shieh and Taylor, 1969) and Damara orogen (Hoernesand Hotrer. 1979).

peratures calculated from both anion and cation ther-mobarometry, are evidence that the isotope ratio of qtartzdid not change significantly as the temperature fell belowthe closure temperature of garnet. The isotopic compo-sition of the fluid present must have been buffered by therock under conditions ofa low fluid/rock ratio. That gar-net and quartz effectively buffered the isotopic composi-tion of the fluid is substantiated by the modal composi-tions which indicate that these two minerals comprise thebulk of the oxygen in most samples.

Quartz-biotite and quartz-feldspar pairs provide ap-parent temperatures that differ substantially from thoseof the quartz-gamet pairs and, in most cases, that differslightly from each other. The difiirsion data suggest thatbiotite should reach its closure at a temperature muchhigher than that of either quartz or feldspar. Such largedifferences in the closure temperatures between biotiteand feldspar may be an artifact of the two diferent meth-ods that were used to calculate the diffusion coefficients,although most certainly the relative sense of the diffusiondata is correct. Quartz-biotite pairs provide slightly higherapparent equilibration temperatures than do the quartz-feldspar pairs. If the isotopic composition of quartz hasnot changed significantly with time as has been suggested,the equilibration temperatures calculated using the ex-perimentally determined quartz-feldspar curyes of Mat-suhisa et al. ( I 979) and the approximation for biotite fromthe curves of muscovite (Garlick and Epstein, 1967; Gra-ham, l98l) may have some relevance. These tempera-tures (Table 4) are much lower than those that have beensuggested for upper-amphibolite facies and are probablyrepresentative of one of the lower-grade events that oc-curred in the Central zone. It is difficult to assess whetherisotope diftrsion or recrystallization was predominant inthe re-equilibration process. Because most mineral ages

HUEBNER ET AL.: STABLE-ISOTOPE GEOCHEMISTRY OF METAPELITES

d160

B b 1 l bBb1 1d( l )Bb12aBb13bBbl 6cBbl 7aBb20Bb25cBb4d-Bb1 b-

532 7.3539 8.7539 8.7505 8.5561 8.4518 8 .5528 8.7643 9.0649 8.0482 7.7

-22- 1 3-26-32- z l

20-22- 3 1-45-30

Note: d'6O calculated from Bottinga and Javoy (1 973)using temperatures obtained from quartz-biotite andquartz-amphibole empirical curves; all 6D calculatedfrom fractionation factors given by Suzuoki and Ep-stein (1976). - Hornblende; rest are biotite.

presence ofCO, (Table 5), which when calculated to 650"Cyield very low difusion rates. In addition, the experi-mental determination of diffusion coefficients for quartz,garnet, and feldspar were done by examining diftrsionprofiles using secondary ion mass spectrometry (srrvrs),whereas diffusion rates for phlogopite were measured us-ing the bulk isotope method, which often yields slowerrates ofdiffi.rsion than the srrras method. An estimation forbulk diflusion in quartz is probably a value between therates for diffusion parallel to the c axis and that perpen-dicular to the c axis. From these results, the order ofoxygen-isotope closure among minerals from earliest tolatest is garnet, qtrartz, and feldspar, with the closure forphlogopite perhaps being intermediate between that ofgarnet and quartz.

The diffusion data suggest thatquarlz and feldspar con-tinue to exchange to lower temperatures than does garnetor, perhaps, biotite, which has important implications forthe significance of the calculated isotope equilibrationtemperatures. Because the calculated temperatures arebased on the partitioning ofoxygen between two phases,ifone phase ceases oxygen-isotope exchange with the fluidwhile the other phase continues the exchange, the tem-peratures determined from the pair ofphases do not reflectequilibrium values. If the phase that continues to ex-change oxygen with a fluid to lower temperatures does sowith a fluid whose isotope composition is bufered by lowwater/rock ratios, the isotopic composition of the mineralwill remain constant because it is buffering the fluid.

The isotopic disequilibrium seen in the minerals fromthe Beitbridge samples may be the result of differentialexchange rates of the minerals with fluids. Freer and Den-nis (1982) suggested that isotopic geothermometry usingthe quartz-garnet pair may record near-peak metamorphictemperatures because of slow diffirsion rates, whereas thequartz-biotite and quartz-feldspar pairs would not. Thelimited diffusion data on garnet, which suggest that it hasthe highest closure temperatures, in conjunction with thesimilarity between the quartz-garnet equilibration tem-

M e t a m o r p h i c W a t e r s

. \-_-/ / ,/.r / //\lnyo

I 350

for the Central zone have been reset to 2. I Ga (Watkeys,1983) regardless of which isotopic system is used, recrys-tallization was probably ongoing at the same time as dif-fusion.

The isotopic compositions ofwaters from various meta-morphosed terranes and those calculated to be in equilib-rium with the biotites from the metapelites and horn-blendes from associated metabasites from the Beitbridgearca aI temperatures equivalent to those provided by theempirical quartz-biotite or quartz-hornblende fractiona-tion curves of Bottinga and Javoy (1975) are listed inTable 6 and shown in Figure 5. If the experimentallyderived fractionation factors and resulting temperaturesare used, the 6D values for the calculated waters are morepositive by l2%n and D'8O values are lower by about2Vm.Exchange rates for D/H between water and hydrous min-erals are much faster than those for oxygen-isotope ex-change (O'Neil and Kharaka, 19761, Graham, 1981) sothat hydrogen may continue to exchange well below theclosure temperature of oxygen. The '80/160 partitioningbetween quartz and biotite may represent too high anapparent temperature for H-isotope exchange so that the6D values ofthe fluids calculated to be in equilibrium withthe hydrous minerals represent the minimum values. Thefluids at Beitbridge lie near the metamorphic water box(Taylor, 197 4)buI because the DD represented in this dia-gram are minimum values and the d'8O are probably max-imum values, the fluids had a distinct component of me-teoric water.

Fluid dynamic controls on isotopic compositions

The nearly concordant temperatures from the qrrarlz-biotite and quartz-feldspar pairs may have implicationsfor the geologic history of the Beitbridge region. Graham(1981) reviewed two different processes and their impli-cations for diffusion rates and stable-isotope exchange inhigh-grade metamorphic rocks: (l) prograde metamor-phism and (2) catastrophic loss of fluid from a system.

During prograde metamorphism the fluid/rock ratio de-creases as fluid is expelled. The fluid/rock ratio can be-come so small as to no longer be able to change the bulkisotopic composition of the rock and may only serve tocalalyze exchange by acting as an exchange medium.Thompson (1983) argued that fluid-absent metamor-phism may be the norm for all metamorphism exceptwhere reactions take place such as at isograds. By "fluid-absent" he suggested that the mineral grain boundariesmay be coated with a water layer that is only one or twomolecules thick, which would correlate well with the con-cept ofvery low fluid/rock ratios. In rocks with low fluid/rock ratios, the isotopic composition of the fastest-difus-ing mineral is somewhat controlled by the closure of thesecond-fastest-diffusing mineral. When the second-fast-est-diffusing mineral reaches its closure temperature, thefastest-diffusing mineral effectively stops exchanging asthe fluid cannot affect the isotopic composition of themineral. Later examination of the isotopic compositionsreveals a pattern of "equilibrium" at least among the two

HUEBNER ET AL.: STABLE-ISOTOPE GEOCHEMISTRY OF METAPELITES

fastest-diffusing minerals. Ifa third mineral, such as garnetor quartz, had reached the closure temperature prior tothe second-fastest mineral, the pattern would be one ofapparent isotopic disequilibrium such as that observed inthe Beitbridge samples. Infiltration of late fluids wouldalso produce this pattern because the minerals with thefastest diffi.rsion or reaction rates would be most affected.

Catastrophic loss offluid essentially depletes the systemofthe buffering agent or catalyst for isotopic exchange andshould therefore allow effective preservation of the iso-topic compositions at the time of the loss. Later exami-nation of the phases may reveal equilibrium or disequi-librium depending on the closure temperatures of theindividual minerals and and the temperature at which theloss offluid occurred. Loss offluid can occur in the fol-lowing ways: (1) the fluid may be consumed by retrogradereactions during the cooling, (2) the opening ofcracks andfractures during the exhumation of a metamorphic orogenmay allow fluid escape, and (3) extraction of fluids by thedevelopment of a melt phase. Because the temperature offluid loss is likely to be different for each ofthese processes,different patterns of mineral isotopic compositions shouldresult.

The partial adjustment of a mineral assemblage to anew set of conditions has been noted in several of theBeitbridge samples as reaction rims of cordierite aroundgarnet, sillimanite, and orthopyroxene. Quartz and feld-spar symplectites also suggest a change in the chemicalcompositions of the feldspars toward more albite-richcompositions. Although no hydrous minerals per se wereproduced, cordierite can contain water in its structurelattice (Martignole and Sisi, 1981). The reactions thereforecould be the result of a fluid phase being consumed duringretrograde cooling. However, the isotopic systematics ofthe Beitbridge samples with or without these reactionproducts are the same, which suggests that the consump-tion of fluids by retrograde reactions was not a majorprocess affecting the isotopic composition of garnet, feld-spar, and biotite.

There is field evidence that large amounts of fluids havepassed through the shear zones that form the boundariesofthe Central zone (Tankard et al., 1982). It is not knownif the fluids were meteoric or metamorphic, so that theestimation of the contribution of this mechanism to theisotopic systematics is very diftcult. In the Beitbridgearea, fluid loss could have occurred as a result ofmassiveoverthrusting (Coward, I 984).

A melt phase produced under high-grade metamorphicconditions would be undersaturated with respect to water.The partitioning of water into the melt phase would ef-fectively dehydrate the surrounding rock and thereby pre-

serve the isotopic compositions ofthe minerals at the timeof the melt production. The apparent equilibration tem-peratures provided by the quartz-biotite, quartz-feldspar,and biotite-feldspar pairs using the experimental calibra-tion curves suggest that exchange between biotite and feld-spar took place to low temperatures and therefore that afluid must have been present.

Anion and cation geothermometry inmetamorphic terranes

The experimental diffusion data, if applicable to themetamorphic process, indicate that exchange rates for cat-ions, on which most of the thermobarometers are based,are often faster than those for anion-exchange rates es-pecially if the vehicle for anion exchange involves a COr-rich fluid. The high-temperature assemblage of Fe-Mgminerals of the Beitbridge samples are garnet + biotite +hypersthene + spinel, and the cation-diffusion rates forgarnets are very slow (Freer, 1979; Cygan and Lasaga,1985). Therefore, when garnet ceased exchanging Fe andMg with the fluid, it left biotite as the major Fe-Mg min-eral along with minor hypersthene and spinel. If the onlyremaining Fe-Mg mineral were biotite, exchange wouldbe effectively stopped, and the cation compositions ofboth the garnet and biotite would reflect the time (andtemperature) of closure of garnet. The later event thatproduced cordierite would provide the next opportunityfor Fe-Mg exchange, but it is probable that the cordieritewould derive Fe and Mg ions from the consumption ofthe garnet with only subsidiary exchange with biotite. Thissuggests that the thermobarometers dependent on cationpartitioning between a limited number of minerals areless likely to distinguish polymetamorphic events becausethe exchange of ions is only between a few specific min-erals. In contrast, oxygen-isotope compositions will re-cord most polymetamorphic events, as the majority ofminerals in any metamorphic rock are oxygen bearing andclosure temperatures for exchange with fluids are variableamong silicates. The development of isotopic equilibriumtemperatures is probably contingent on the rapid coolingor fluid loss from peak metamorphic conditions.

CoNcr,usroNsl. Stable-isotope systematics of the upper-amphibolite-

facies metamorphic rocks from the Central zone of theLimpopo belt yield different apparent equilibration tem-peratures among various minerals.

2. Of the mineral pairs examined, the mineral pairquartz-garnet yields isotope equilibration temperatures thatare closest to those given by the mineralogical and cationgeothermometers. Because garnet and quartz comprisemost of the oxygen in the Beitbridge samples, the 6rEOvalue of these minerals probably did not change after theclosure temperature ofgarnet. The othertwo pairs, quarlz-biotite and quartz-feldspar, give lower isotope equilibra-tion temperatures than does quartz-garnet and are slightlydifferent from each other. They appear to represent con-ditions of later, lower-grade metamorphic events.

3. A combination of diffusion of oxygen isotopes andrecrystallization of mineral phases is responsible for theisotope systematics of these assemblages.

4. The waters calculated to be in equilibrium with hy-drous minerals from Beitbridge samples are similar toother regional high-grade metamorphic terranes but maycontain a component of meteoric water.

1 3 5 1

5. Small but significant differences among the whole-rock DrtO values ofdifferent lithologic units at Beitbridgesuggest a partially closed system of metamorphism andgenerally low water/rock ratios.

AcrNowlpocMENTs

This work was much improved by detailed criticism from C.M. Graham, A. Matthews, and an anonymous referee. Collectionof the samples was possible through a grant from the RoyalSociety, London, to E.G.N., in association with G. T. R. Droop,N. B. W. Harris; and T. J. B. Holland, who participated in fieidwork. We thank Tim Broderick, Tony Martin, the GeologicalSurvey of Zimbabwe and the Zimbabwe Republic police (In-spectors Magama and and Samkange) for their great help, andConstables Maquta and Khanye for their escort in a region ofhigh element activity.

In Canada the work was funded by NSERC grants to Nisbetand Kyser, and by a University of Saskatchewan scholarship toHuebner.

RrrnnnNcrsBarton, J.M., Jr., and Ryan, B. (1977) A review of the geochron-

ologic framework of the Limpopo mobile belt. Geological Sur-vey of Botswana Bulletin, 12, 183-200.

Barton, J.M., Jr., Fripp, R.E.P., and Ryan, B. (1977) Rb-Sr agesand geological setting ofancient dykes in the Sand River area,Limpopo mobile belt, South Africa. Nature, 267, 487-490.

Bottinga, Y., and Javoy, M. (l 973) Comments on oxygen isotopicgeothermometry. Earth and Planetary Science Letters, 20, 250-265.

- (1975) Oxygen isotope partitioning among the mineralsin igneous and metamorphic rocks. Reviews of Geophysicsand Space Geophysics, 13, 401-408.

Broderick, T.J. (1979) Explanation ofthe geological map ofthecountry south of Nuanetsi, Nuanetsi and Beitbridge Districts.Rhodesian Geological Survey Short Report, 46.

Chinner, G.A., and Sweatman, T.R. (1968) A former associationof enstatite-kyanite. Mineralogical Magazine, 36, I 052-1060.

Clayton, R.N., and Mayeda, T.K. (1963) The use of brominepentafluoride in the extraction of oxygen from oxides and sil-icates for isotopic analysis. Geochimica et Cosmochimica Acta,27,43-52.

Clayton, R.N., Goldsmith, J.R., Karel, K.J., Mayeda, T.K., andNewton, R.C. (1975) Limits on the effects of pressure on iso-topic fractionation. Geochimica et Cosmochimica Acta, 39,rr97-r20r.

Coward, M.P. ( 1984) Major shear zones in the Precambrian crust;examples from N.W. Scotland and southern Africa and theirsignificance. In A. Krciner and R. Greiling, Eds. Precambriantectonics illustrated, 20'1 -236. E. Schweizerbart'sche Verlags-buchhandlung, Stuttgart.

Cygan, R.T., and Lasaga, A.C. (1985) Self-difusion of Mg ingarnet at 750-900"C. American Journal ofScience, 285,328-350.

Deines, P. (1977) On the oxygen isotope distribution amongmineral triplets in igneous and metamorphic rocks. Geochim-ica et Cosmochimica Acta. 41. 1709-1730.

Du Toit, M.C., van Reenen, D.D., and Roering, C. (1983) Someaspects of the geology, structure and metamorphism of thesouthern marginal zone of the Limpopo metamorphic com-plex. In W.J. van Biljon and J.H. kgg, Eds. The Limpopobelt, l2I-142. Geological Society ofSouth Africa Special Pub-lication 8, Johannesburg.

Fourcade, S., and Javoy, M. (1973) Rapports '80/'60 dans lesroches vieux sade catazonal d'In Ouzzal (Sahara, Algerien).Contributions to Mineralogy and Petrology, 42,235-244.

Freer, R. (1979) An experimental measurement of cation diftr-sion in almandine garnet. Nature, 280,220-222.

HUEBNER ET AL.: STABLE-ISOTOPE GEOCHEMISTRY OF METAPELITES

1352

-(1981) Diffirsion in silicate minerals and glasses; a datadigest and guide to the literature. Contributions to Mineralogyand Petrology, 7 6, 440-454.

Freer, R., and Dennis, P.F. (1982) Oxygen diffusion studies I. Apreliminary ion microprobe investigation of oxygen diffirsionin some rock-forming minerals. Mineralogical Magazine, 45,r79-192.

Frey, M., Hunziker, J.C., O'Neil, J.R., and Schwander, H.W.(1976) Equilibrium-disequilibrium relations in the Monte Rosagranite, western Alps: Petrology, Rb-Sr, and stable isotop€ data.Contributions to Mineralogy and Petrology, 55, 147-179.

Friedman, I., and O'Neil, J.R. (1977) Compilation of stable iso-tope fractionation factors ofgeochemical interest. United StatesGeological Survey Professional Paper 440-KK.

Fyfe, W.S., Price, N.J., and Thompson, A.B. (1978) Fluids in theEarth's crust, Elsevier, Amsterdam.

Garlick, G.D., and Epstein, S. (1967) Oxygen isotope ratios incoexisting minerals from regionally metamorphosed rocks.Geochimica et Cosmochimica Acta. 31. 181-214.

Giletti, B. J., and Anderson, T.F. (1977) Studies in diffirsion: II.Oxygen in phlogopite mica. Earth and Planetary Science Let-ters. 28. 225-233.

Giletti, B.J., and Yund, R.A. (1984) Oxygen ditrusion in quartz.Joumal of Geophysical Research, 89, 4039-4046.

Giletti, B.J., Semet, M.P., and Yund, R.A. (1978) Studies indiffirsion: III. Oxygen in feldspars: An ion microprobe deter-mination. Geochimica et Cosmochimica Acta, 42, 45-57.

Graham, C.M. (1981) Experimental hydrogen isotope studies:III. Diffusion of hydrogen in hydrous minerals and stable iso-tope exchange in metamorphic rocks. Contributions to Min-eralogy and Petrology, 1 6, 216-228.

Harris, N.B.W., and Holland, T.J.B. (1984) The significance ofcordierite-hypersthene assemblages from the Beitbridge regionof the central Limpopo belt; evidence for rapid decompressionin the Archaean. American Mineralogist, 69, 1036-1049.

Hoernes, S., and Friedrichsen, H. (1978) Oxygen and hydrogenisotope study of the polymetamorphic area of the northernOtztal-Stubai Alps (Tyrol). Contributions to Mineralogy andPetrology, 67, 305-31 5.

Hoernes, S., and Hoffer, E. (1979) Equilibrium relations of pro-grade metamorphic mineral assemblages: A stable isotope studyof rocks of the Damara orogen, from Namibia. Contributionsto Mineralogy and Petrology, 68,377-389.

Horrocks, P.C. ( 1980) Ancient Archaean supracrustal rocks fromthe Limpopo mobile belt. Nature, 286,596-599.

Javoy, M., Fourcade , S., and All0gre, C.J. ( 1970) Graphical meth-od for examination of tEO/t6O fractionations in silicate rocks.Earth and Planetary Science ktters, 10, 12-16.

Keith, M.L., and Weber, J.N. (1964) Isotopic composition andenvironmental classification of selected limestones and fossils.Geochimica et Cosmochimica Acta. 38. 1787-1816.

Kyser, T.K., and O'Neil, J.R. (1984) Hydrogen isotope system-atics of submarine basalts. Geochimica et Cosmochimica Acta,48.2123-2t33.

Kyser, T.K., O'Neil, J.R., and Carmichael, I.S.E. (1982) Geneticrelations among basic lavas and ultramafic nodules: Evidencefrom oxygen isotope compositions. Contributions to Miner-alogy and Petrology, 81, 88-102.

Light, M.P.R. (1982) The Limpopo belt: A result of continentalcollision. Tectonics, l, 325-342.

Light, M.P.R., and Watkeys, M.K. (1978) An outline of the Ar-chaean and early Proterozoic geological history of the regionaround Beitbridge. Rhodesia Geological Survey Annals, 3, 35-40.

Light, M.P.R., Broderick, T.J., and Watkeys, M.K. (1977) Apreliminary report on the Central zone of the Limpopo mobilebelt, Rhodesia. Geological Survey of Botswana Bulletin, 12,b l - / J -

Longstaffe, F.J., and Schwarcz, H.P. (1977) 180/'60 of Archeanclastic metasediments; a petrogenetic indicator for Archeangneisses. Geochimica et Cosmochimica Acta, 41, 1303-1312.

HUEBNER ET AL.: STABLE-ISOTOPE GEOCHEMISTRY OF METAPELITES

Manton, W.I. (1968) The origin of associated basic and acid rocksin the Lrbombo-Nuanetsi igneous province, southern Africa,as implied by Sr isotopes. Journal of Petrology, 9, 23-39.

Martignole, Jacques, and Sisi, J.C. (1981) Cordierite-garnet'H,Oequilibrium: A geological thermometer, barometer and waterfugacity indicator. Contributions to Mineralogy and Petrology,77 ,38-46.

Mason, R. (1973) The Limpopo mobile belt-southern Africa.Royal Society oflondon Philosophical Transactions, series A,273,463-485.

Matsuhisa, Y., Goldsmith, J.R., and Clayton, R.N. (1979) Oxy-gen isotope fractionation in the system quartz-albite-anorthite-water. Geochimica et Cosmochimica Acta, 43, 1131-l140.

Matthews, A., Goldsmith, J.R., and Clayton, R.N. (1983) Oxygenisotope fractionations involving pyroxenes: The calibration ofmineral-pair geothermometers. Geochimica et CosmochimicaActa.43. 631-644.

McNaughton, N.J., and Wilson, A.F. (1980) Problems in oxygenisotope geothermometry in mafic granulite-facies rocks fromnear Einasleigh, northern Queensland. Precambrian Research,13,77-86.

- ( I 9 8 3) The geochemical and oxygen isotope afrnities ofProterozoic mafic granulites from the Einasleigh metamor-phics, northern Queensland. Precambrian Research, 21, 2l-37 .

Muehlenbachs, K., and Kushiro, l. (1974) Oxygen isotope ex-change and equilibration of silicates with CO, or Or. CarnegieInstitution of Washington Year Book 73, 232-236.

Norris, R.J., and Henley, R.W. (1976) Dewatering of a meta-morphic pile. Geology, 4,333-336.

O'Neil, J.R., and Kharaka, Y.K. (1976) Hydrogen and oxygenisotope exchange reactions between clay minerals and water.Geochimica et Cosmochimica Acta, 40,241-246.

O'Neil, J.R., and Taylor, H.P., Jr. (1969) Oxygen isotope equi-librium between muscowite and water. Journal of GeochemicalResearch. 7 4. 6012-6022.

Perry, E.C., Jr., and Tan, F.C. (1972) Sienificance of oxygen andcarbon isotope variations in early Precambrian cherts and car-bonate rocks of southern Africa. Geological Society of AmericaBulletin, 83,647-664.

Robertson, I.D.M. (1973) The geology of the country around Mt.Towla, Gwanda District. Geological Survey of Rhodesia Bul-letin, 68.

Rumble, D., III, and Spear, F.S. (1983) Oxygen-isotope equili-bration and permeability enhancement during regional meta-morphism. Geological Society of London Journal, 140, 619-628.

Savin, S.M., and Epstein, S. (1970) The oxygen and hydrogenisotope geochemistry of clay minerals. Geochimica et Cos-mochimica Acta, 34, 25-42.

Shieh, Y.N., and Taylor, H.P., Jr. (1969) Oxygen and carbonisotope studies of contact metamorphism of carbonate rocks.Journal ofPetrology, 10, 307-331.

Suzuoki, T., and Epstein, S. (1976) Hydrogen isotope fraction-ation between OH-bearing minerals and water. Geochimica etCosmochimica Acta, 40, 1229-1240.

Tankard, A.J., Jackson, M.P.A., Eriksson, K.A., Hobday' D.K.,Hunter, D.R., and Minter, W.E.L. (1982) Crustal evolution ofsouthern Africa: 3.8 billion years of earth history, Springer-Verlag, New York.

Taylor, H.P., Jr. (1968) The oxygen isotope geochemistry of ig-neous rocks. Contributions to Mineralogy and Petrology, 19,r-7 t .

- (197 4) The application of oxygen and hydrogen isotopestudies to problems of hydrothermal alteration and ore de-position. Economic Geology, 69, 843-888.

Thompson, A.B. (1983) Fluid absent metamorphism. GeologicalSociety of London Journal, 140, 533-547.

Touret, J. (1971) k facies granulite de norvege meridoniale. II.Les inclusions fluides. Lithos, 4,423-436.

van Breemen, O., and Dodson, M.H. (1972) Metamorphic chro-

HUEBNER ET AL.: STABLE-ISOTOPE GEOCHEMISTRY OF METAPELITES I 353

nology of the Limpopo belt, southern Africa. Geological So-ciety of America Bulletin, 83, 2005-2018.

Veizer, J., and Hoefs, J. (1976) The nature of 'EO/'6O and '3Cl''zC secular trends in sedimentary carbonate rocks. Geochimicaet Cosmochimica Acta, 40, 1387-1397.

Walther, J.V., and Orville, P.M. (1982) Volatile production andtransport in regional metamorphism. Contributions to Min-eralogy and Petrology, 79, 252-257 .

Walther, J.V., and Wood, B.J. (1984) Rate and mechanism inprograde metamorphism. Contributions to Mineralogy and Pe-trology, 88, 246-259.

Watkeys, M.K. (1979) Explanation of the geological map of thecountry west ofBeitbridge. Rhodesia Geological Survey ShortReport, 45.

- (1983) Briefexplanatory notes on the provisional geo-logical map of the Limpopo belt and environs. In W.J. vanBiljon and J.H. Legg, Eds. The Limpopo belt, 5-8. GeologicalSociety ofSouth Africa Special Publication 8, Johannesburg.

Watkeys, M.K., Light, M.P'R., and Broderick' T.J. (1983) Aretrospective view of the Central zone of the Limpopo belt,Zimbabwe. In W.J. van Biljon and J.H. kgg, Eds. The Lim-popo belt, 65-80. Geological Society of South Africa SpecialPublication 8, Johannesburg.

Wilson, A.F., and Baksi, A.K. (1984) Oxygen isotope fraction-ation and disequilibrium displayed by some granulite faciesrocks from the Fraser Range, western Australia. Geochimicaet Cosmochimica Acta, 48,423-432.

Windley, B.F., Ackermand, D., and Herd, R.K. (1984) Sapphi-rine/kornerupine bearing rocks and crustal uplift history oftheLimpopo belt, southern Africa. Contributions to Mineralogyand Petrology, 86, 342-358.

Mexuscnrpt REcETvED Aucusr 22, 1985Mamrscrrpr ACcEPTED JulY 8, 1986