American Mineralogist, Volume 67, pages M3452, 1982

Review of the behavior of plagioclase under metamorphic conditions

Julrex R. Gor-osurru

Department of the Geophysical SciencesThe University of Chicago

Chicago, Illinois 60637

Abstract

The behavior of plagioclase feldspars under metamorphic conditions is reviewed. Theliterature on plagioclase composition as related to reactions involving zoisite (epidote) andother calcic phases, the peristerite gap, and the coexistence ofintermediate plagioclases ofvarious compositions is presented. The various ideas on the relative effects of structuralstates and of the role of chemical reactions on subsolidus relations are considered. Twoviews, not mutually exclusive, relative to plagioclase compositions, are apparent: a"homogeneous" view, emphasizing control by crystal-structural factors, and a "heteroge-neous" view, dominated by sequential mineralogical reactions.

Mineralogy and petrology are supplementary andcomplimentary disciplines. Much of our under-standing of the petrology and ultimately the geologyof any portion of the earth is dependent on ourknowledge of the properties and characteristics ofminerals and of mineral interactions. The petrolo-gist examines the minerals to help unravel thehistory of a rock; the mineralogist gains insight onmineral behavior from accumulated information onthe rocks. Some minerals with limited composition-al and structural variation have little to reveal.Others, such as the feldspars, may contain a greatdeal of information, but can be complex and stub-born in revealing their secrets. A voluminous litera-ture on plagioclase has not yet clearly characterizedthe phase relations. The equilibrium picture is over-lain by confusion compounded of and dominated bykinetic factors. Furthermore, heterogeneous miner-alogical reactions can be affected by the varyingenergetics of different structural states. This reviewbriefly covers and is restricted to the highlights ofpapers on plagioclase feldspars in metamorphicrocks, and is followed by a report on an experimen-tal investigation that bears on plagioclase stabilityat_ elevated temperatures and water pressures.

Petrologists have long been aware of an increasein anorthite content in some metamorphic rocksaccompanying an increase in metamorphic grade(Becke, 1913). Although earlier workers had con-sidered a relationship between plagioclase and epi-dote or other Ca-rich minerals (see Christie, 1959),Ramberg (1943), largely on the basis of field evi-

dence, attempted to explicitly define the nature ofthe relationship. Ramberg proposed an equilibriumbetween epidote and "anorthite-molecule" in theplagioclase:

2CaAl2Si2O8 + llzHzO = Ca2Al3Orz(OH)anorthite "epidote"

+ l/2Al2sio5 + 1/2sio2kyanite quartz

He then wrote a reaction in which calcic plagioclaseplus water breaks down to a more albitic plagioclaseplus epidote, kyanite, and quartz and constructed atheoretical curve for the equilibrium in which, withdecreasing temperature, an increasingly sodic pla-gioclase is stable with epidote. Ramberg was fullycognizant ofthe role ofpressure and drew schemat-ic diagrams of both isobaric and isothermal surfacesfor the reaction. He considered this reaction rela-tion to be a continuous one all the way from thegreenschist to the granulite facies.

Ramberg (1949) extended his ideas on the plagio-clase-epidote equilibrium to potassium-rich rockswith orthoclase and muscovite. A second continu-ous reaction occurs:

4CaAlzSizO8 + KAlSi3Os + H2Oanorthite orthoclase

? KAhSigOro(OH)z + 2Ca2Al3SirOrz(OH) + 2SiO2muscovite epidote quartz

Figure I is a reproduction of his (2-dimensional)diagram. The dashed curve, representing the equi-

0003-004x/82/0708-0643$02. 00 il3

644 GOLDSMITH: REVIEW OF PLAGIOCLASE

HOMOGENEOUS FIELDof

P L A G I O C L A S E

ar%\-^\

EPIDOTE Aprlllocusr 4 U

(Kyon i t e , t lMuscovi te ) Yo'

t-

0 20 40 60 80 t00C o , A l N o , A l

S i l i c o t e s S i l i c o t e sFig. l. Reproduction ofRamberg's (1949) subsolidus diagram

of the plagioclase P epidote equilibrium, deduced in part fromfield observations.

librium of plagioclase and epidote in the presence ofmuscovite and potash feldspar is drawn somewhatabove the K-free equilibrium because muscovite inassociation with K-feldspar and quartz has some-what lower Al-activity than kyanite. Rambergpointed out that increased H2O pressure displacesthe equilibrium toward epidote, as does increasingthe Fe3+/Al ratio because of the ability of epidote toincorporate Fe3* in its structure. Ramberg consid-ered the reactions of Figure I to span a temperatureinterval of approximately 500" C, and facies, repre-senting temperature, are ploted on the vertical axis.He felt that the epidote-plagioclase equilibrium inmost quartzo-feldspathic rocks cannot register P-Zconditions higher than those that would developAnsg-ao plus epidote, due to the low CalNa ratio inthese rocks. Ramberg (1952) further extended theplagioclase-epidote relations to a more complexsystem containing CaCO3 and indicated that rockswith the same plagioclase compositions may haveformed at different temperatures, depending on thepartial pressure of water.

Lyons (1955) plotted plagioclase compositionsagainst metamorphic grade in the Hanover quadran-gle (New Hampshire-Vermont) as a check againstRamberg's (1943) diagram. He observed that in ageneral way the separation of the plagioclase-epi-dote and plagioclase fields followed Ramberg's

curve, and the data clearly indicated a gradualincrease in the anorthite content of the plagioclaseswith increasing grade of metamorphism. He notedsignificant deviations, however, and indicated thatthere are problems, as recognized by Ramberg, inits use as an index of intensity of metamorphism. Inaddition, Lyons noted, ". . . a great paucity ofmetamorphic plagioclase in the range of AnrrAn2o."

Ramberg's (1943, 1949, 1952) deduced curveswere continuous in nature, implying an equilibriumthat varies continuously with temperature (or meta-morphic grade). A number of field studies in region-ally metamorphosed terranes had given some indi-cation of a rapid conversion of albite to oligoclaseor sodic andesine over a short distance, but it islikely that the first clear-cut evidence for a discon-tinuous increase in plagioclase composition withincreasing grade was described from greywackes ofthe Missi series, Manitoba (Ambrose, 1936; seeBrown, 1962). Ambrose commented on abundantepidote, and assumed that original intermediateplagioclase of volcanic rocks has reacted, ".according to the familiar equation: plagioclase +water * iron : albite + epidote * alumina *silica." He noted that water-clear albite persistswith little change through the biotite zone, but withthe appearance of garnet, the anorthite content ofthe plagioclase rises abruptly from An64 to An2s-32and that epidote is greatly diminished in amountabove the garnet isograd, concomitant with theappearance of oligoclase.

Since the earlier years of the 20th century, whenthe nature of the isomorphous replacement of Naand Si for Ca and Al in the plagioclase feldsparsbecame known, petrologists had long consideredthe plagioclases to be an ideal solid solution series.The dogma of one rock, one plagioclase (excludingzoning, or other non-equilibrium phenomena) pre-vailed. Chao and Taylor (1940) were the first toobserve superstructure reflections in X-ray diffrac-tion patterns of intermediate plagioclases, indicat-ing structural complexities. Chao and Taylor con-sidered this as evidence against a complete solidsolution series, although a strong hint had earlierappeared when Taylor et al. (1934) found the c-axisof anorthite to be twice that of albite. Kohler(1942a, b) and Larsson (1941) were the first toexplicitly describe optical diferences between pla-gioclases ofplutonic and volcanic rocks, and thus todifferentiate high-temperature from low-tempera-ture plagioclases. Cole et al. (1951) showed thatlow-temperature plagioclase from An3q to AnT2had

I

UF

z

1tF

a

==

z

-i

U

o

: +

1 lt ll t

E 9r 9

1 lt lt

GOLDSMITH: REVIEW OF PLAGIOCLASE 645

complex superstructures. Laves (1951, 1954) re-ported that there are plagioclases in the composi-tional range of approximately An5-Ar17, &rlorgthem peristerites, that show two phases in "single-crystal" X-ray photographs. Laves estimated thecompositions of the exsolved phases to be approxi-mately Ane and An3s. Gay and Smith (1955) con-firmed this observation, and refined the estimationof composition of the two phases as An3*2 andAnzt=z.This region of composition was termed the"peristerite gap". Additional crystallographic workand refinement was done by Ribbe (1960, 1962),Brown (1960), and Fleet and Ribbe (1965).

Christie (1959) pointed out the significance of theperisterite gap in low-temperature plagioclases, andthat as a consequence Ramberg's smooth unbrokencurve needed modification. He also assumed animmiscibility gap in the intermediate region delin-eated by Cole et al. (1951) in the low-temperatureplagioclases. On the basis of the discontinuity pro-duced by the peristerite phenomenon, Christie stat-ed that the plagioclase epidote thermometer wasuseless.

At this same time de Waard (1959) published onbasic and pelitic schists of the Usu massif in Timorand noted changes in the An content of the plagio-clases over relatively short distances, reflecting asuccession of narrow zones of metamorphism. Hefound the frequency of occurrence of plagioclasebetween An5 and An2s to be remarkably low, partic-ularly in the basic schists. He was almost certainlyunaware of the evidence of a peristerite gap, andsuggested that the abrupt change is coupled to theformation of garnet in pelites and hornblende inbasic rocks. The papers by Christie (1959) and deWaard (1959) set off a small flurry of activity in theearly 1960's on the niceties of the plagioclase com-positional problem from two points of view: (a)facies boundaries, and (b) crystal-chemical aspects.

From this point until the present there have beennumerous field observations, a number of interpre-tive papers, rather few experimental studies, andmany crystallographic investigations, all related tothe plagioclases, and directly or indirectly, to theirbehavior in metamorphic and low-temperature en-vironments. Some of the earlier interpretive papersare by Rutland (1961, 1962), Christie (1962), Noble(1962), Sen (1963) and Kretz (1963). Noble (1962)summarized the previous field and theoretical workrelative to the peristerite gap and presented a fur-ther refinement of the Ramberg-Christie diagrams.He also considered the effect of intermediate meta-stable plagioclases, suggesting that the increased

free energy raises and flattens the plagioclase-epidote equilibrium curve in the An35-An5s r&ng€,reducing the temperature interval in which calcicandesine is stable.

Brown (1962) presented an excellent review andclarification of the subject of peristerite unmixing inthe plagioclases and related metamorphic and faciesseries. His summary of the field observations indi-cated that the rate of increase in the composition ofthe most basic plagioclase with increasing metamor-phic grade is continuous or nearly so in rocks of theandalusite-sillimanite type and perhaps in the low-pressure intermediate type. A jump occurs betweenabout An7 and about An26 in the kyanite-sillimanitetype and the high-pressure intermediate type, and inthe jadeite-glaucophane type only albite seems toexist. He mentioned that the relation between gar-net and the discontinuity in plagioclase compositionis not clear; there is an implication that reactionsinvolving garnet and other Ca-rich phases may ormay not be important in this connection. Sen (1963)pointed out that not all plagioclases unmix, thusbreaks in composition are not always present, andat any rate are of no avail in delineating metamor-phic facies. Demarcation must be made on empiri-cal grounds based on petrographic observations,and checked against other independent evidence offacies transitions, such as reactions involved in themetamorphic process.

Kretz (1963) restated Ramberg's (1943, 1949,1952) emphasis on the fact that by the nature oftheir compositions plagioclase-epidote reactionsmust involve one or more additional phases, andthat knowledge of the reactions is essential to anunderstanding of plagioclase thermometry. Throughthe use of some simplifying assumptions, Kretzderived equations to interrelate the composition ofplagioclase, temperature, pressure, chemical poten-tial of HzO and the composition of the other solidsolutions that take place in six univariant universalreactions, three of which were considered by Ram-berg:

epidote * muscovite + quartzP anorthite * potash feldspar + H2O (l)

epidote + kyanite + quartz<= anorthite + H2O (2)

epidote + CO2a anorthite * calcite + H2O (3)

epidote * chlorite + quartz? actinolite * anorthite + H2O (4)

epidote * actinolite + quartz

e anorthite + calcic pyroxene + H2O (5)

GOLDSMITH: REVIEW OF PLAGIOCLASE

epidote + quartz

€ anorthite * grossularite + Hro (6)

He then modified the equilibria by imposing feld-spar ordering, or non-ideal behavior in the reac-tions, and thus for the first time attempted to treattwo of the factors that are involved in the behaviorof plagioclase under metamorphic conditions.

Rambaldi (1973) found a general increase in theanorthite content of plagioclase and a decrease inthe Al content of epidote with increasing metamor-phic grade in metasediments and metavolcanicsnear Bancroft, Ontario. Plagioclase crystals arecommonly zoned, becoming richer in anorthite to-ward the rims, and epidote crystals may also bezoned, enriched in iron toward the rims. The plagio-clases associated with epidote are mostly oligo-clases and andesines, with a maximum of Ana6. Theplagioclase in rocks without epidote is generallyalbitic. Misch (1968) also observed that plagioclasemore calcic than An37-as is rare in the epidote-bearing subfacies of the amphibolite facies SkagitGneiss of Washington State. Rambaldi attributesthe increasing An content to the breakdown ofepidote, and considered fifteen possible mineralogi-cal reactions that may be involved in the variouscompositional changes. More complex reactionsthan those discussed by Kretz (1963) are recog-nized, and generally involve sphene and Fe-Tioxides. Hoy (1976) also observed "positive zoning"of metamorphic plagioclases (increase of An con-tent toward the rim) associated with prograde meta-morphism, and "negative zoning" with retrogradereactions in the Riondel area of British Columbia.These phenomena are associated with the modelreaction clinozoisite * tremolite + quartz ? diop-side + plagioclase + H2O. The distribution ofisograds and the variation in plagioclase composi-tion in the above assemblage is modeled using T-X66, diagrams calculated at 5 kbar pressure, indi-cating that An content increases with increasing Tand Xss,; Hoy indicated that in the assemblage epi-dote - actinolite - qo rtz - diopside - plagioclase,plagioclase compositions ) An6e are reached attemperatures above 600'C and at X66, values of0.5-0.6. Ghent and DeVries (1972) had also consid-ered the role of X6s, in plagioclase compositions.Bird and Helgeson (1981), in an extensive reviewand treatment of the system NazO-KzO-CaO-FeO-Fe2O3-AlzO3zO-HCl-2 at temperatures to 5kbar and 600' C, conclude that the complex zoning

commonly exhibited by both plagioclase and epi-dote solid solutions in geologic systems can beattributed to minor isothermal/isobaric changes inthe composition of coexisting aqueous solutions.

It is apparent that there have been two differentviews on mechanisms controlling plagioclase com-positions, particularly in metamorphic rocks. Onemight be called an internal, or homogeneous mecha-nism, controlled by structural discontinuities, or-der{isorder, or other crystallographic features inthe feldspar. The other is the result of external, orheterogeneous factors, such as progressive or retro-gressive sequences of mineral reactions induced bychanging P-? conditions, producing discontinuitiesin plagioclases as a consequence of "abrupt"changes in mineral assemblages. Either is feasibleand both might be operative: discussion of thehomogeneous view began with Chao and Taylor(1940), and evidence for compositional discontinui-ties in natural plagioclases presumably produced bystructural control is given, for example, by Domanet al. (1965), and DeVore (1956). The concept ofheterogeneous control is implicit in the writings ofRamberg (1943,1949,1952) and espoused by otherssince that time.

Up until this point there had been no recognitionof the existence of two optically distinguishableplagioclase phases in any rock other than presumeddisequilibrium associations (see for example, Sen,1963). Evans (1964), in a combined microprobe anduniversal stage study (microscopy by F. J. Turner),was the first to report clear-cut duality in plagio-clase composition from schists of the almandine andoligoclase zones in New Zealand. The albites arevery pure, the bulk of the grains below Anr.o.Epidote is present in all of the rocks. The coexistingoligoclase ranges from An1s.3 to An25.6, averagingAnzr-zq.s, and the two plagioclases occur side by sideas distinct grains. There is some tendency foroligoclase to occur as thin irregular strips alongalbite grain boundaries.

Observations of two plagioclase rocks becamerelatively common after Evans' (1964) note, per-haps because the ice was broken, but also becauseof the availability of the electron microprobe. Tsuji(1966) studied the Higo metamorphics in Kyusyu,Japan, where two plagioclases, one with albite andthe other with oligoclase or andesine composition,occur side by side in the same thin section. Interme-diate compositions do not occur. Although thecompositional gap seems to decrease with risingtemperature and minimum and maximurn An con-

GOLDSMITH : REVIEW OF PLAGIOCLASE 647

tents converge to Anla in the central part of thehighest temperature (garnet) zone, Tsuji cautionedagainst the use of a presumed peristerite gap as thesole explanation, pointing out the possibility of areaction at the boundary between the epidote am-phibolite and the amphibolite facies where albite,chlorite, and actinolite react to form a component ofamphibole that enriches the plagioclase in Ca.

Crawford (1966) studied two suites of regionallymetamorphosed semi-pelitic schists from Vermontand New Zealand in order to investigate the para-genesis of low-temperature plagioclase. On the ba-sis of coexisting albite and oligoclase, and takinginto consideration reactions to produce differentassociated Ca-rich phases, Crawford outlined aperisterite solvus with a vertical albite limb and asloping oligoclase limb, at pressures on the order of5-7 kbar. The solvus maximum occurs at about450-500" C (corresponding to the almandine-stau-rolite isograds in the Vermont area) and spans theplagioclase compositions Anl to An2aat lower tem-peratures.

Cooper (1972) followed progressive mineralogicalchanges in metabasic rocks of a Barrovian-typeseries, from greenschist to amphibolite facies in thesouthern Alps of New Zealand. This area is thesame as that studied by Evans (1964) and Crawford(1966). As did Crawford, Cooper used coexistingcompositions of albite and oligoclase through thegarnet zone to outline the form of the peristeritesolvus. At the oligoclase isograde the compositionsare An6-1 and An23-25. In Vermont the calciummineral participating in the development of oligo-clase is a carbonate, in New Zealand, epidote.Cooper noted that oligoclase first appears as thindiscontinuous rims around an albite nucleus at thegarnet-oligoclase-hornblende isograd, and althougholigoclase is soon in evidence as individual grains,zoned grain relationships persist throughout thegarnet zone. With increasing metamorphic gradethrough the garnet zone oligoclase increases inamount, becoming more sodic in composition,while albite decreases, becoming more calcic. Coo-per stated that true amphibolite facies status isattained by the final disappearance of albite at thesecond oligoclase isograd, either by the completereplacement of albite cores by oligoclase, or (proba-bly simultaneously) by the rise of sodic plagioclaseabove An1s, presumably due to closure of theperisterite solvus.

Morteani and Raase (1974) found abrupt changesin plagioclase compositions, seen especially as re-

verse zoning, even three-fold zoning, in epidote-bearing, amphibole-free rocks of the eastern Alps.The plagioclase cores are commonly 0-6Vo An, with14-19% An in the rims. Compositions between An5and An15 are rare, presumably representing theperisterite gap. Plagioclases more calcic than An35are also rare. Only calcite-free or low-calcite rockswere studied, as the An-content of the plagioclase issignificantly increased in rocks with greater than20Vo carbonate. They noted that the presence ofmargarite may also influence the Ca-content of theplagioclase, as indicated by Ackermand and Mor-teani (1973). In epidote-containing rocks, they callupon the peristerite gap plus an ill-defined two-stage metamorphic process with increasing tem-peratures to account for the feldspar compositions.

Frey and Orville (1974) studied a margarite-bear-ing black shale from the central Swiss Alps with thefollowing unusual features:

1) The first plagioclase to appear in the lower-grade greenschist facies is not albite, but oligo-clase-andesine (-An:o).

2) The An-content commonly remains constantwith increasing metamorphic grade.

3) Some rocks at the transition from the green-schist to the amphibolite facies show unusual-ly high An-contents (up to Anzo). Graphicalanalysis of the system CaO-Na2O-AlzOr-SiO2-H2O-COz (with SiO2, H2O, and CO2 inexcess) with the phases pyrophillite, parago-nite, margarite, zoisite, calcite, and plagio-clase satisfactorily explains the variable com-positional nature of the plagioclase and addi-tionally points to the importance of certainspecific reactions under varying conditions ofP-T-X in controlling the composition of thefeldspar.

The importance of the Fe content of epidote andthe oxidation potential on the composition of theassociated plagioclase was discussed by Hormannand Raith (1973), who investigated the Fe3+ contentof epidotes in a metamorphic series from Tyrol,Austria. Fe3* in epidote as well as the oxidationratio (Fe3+ x 100/Fe3+ + Fe2*1 decreases fromgreenschist to almandine-amphibolite facies. Re-dox reactions were suggested that account for ob-served mineral assemblages and that may explainthe increase of An-content of coexisting plagioclasewith increasing metamorphic grade. The impor-tance of oxidation and of the value of P66,/Pq,s wasalso considered by Rambaldi (1973).

E. Wenk (1962) observed that the An content of

GOLDSMITH: REVIEW OF PLAGIOCLASE

plagioclase associated with calcite is almost uni-form in a given region and tends to be high, al-though it varies from area to area dependent on thegrade of metamorphism. Wenk and Keller (1969)examined 700 amphibolites from the central Alps,and distinguished zones of albite amphibolites, oli-goclase amphibolites, andesine amphibolites, andlabradorite amphibolites, apparently as discontinu-ous sequences of plagioclase composition. Al-though heterogeneous control was not specificallydiscussed, the observations (with calcite present inparticular) and the zonal relationship implied amechanism controlled by mineralogical reactions.This view became strongly modified in later papersby H.-R. Wenk and coworkers when transmissionelectron microscopy and single-crystal X-ray workwere used in association with the earlier microscop-ic and field petrographic techniques. Wenk et al.(1975) described intergrowths of andesine (Any)and labradorite (Anoo) with textures indicating si-multaneous equilibrium growth. Although mineralassociations (chiefly zoisite and/or scapolite) indirect contact with the plagioclases are considered,emphasis is placed on structure. Wenk et al. clearlystated that the microscopic intergrowths representend members of the intermediate plagioclase misci-bility gap. They noted DeVore's (1956) thesis thatAn33 and An66 €r€ crystallographically "stable"compositions and Wenk et al. felt that in an inter-growth both phases have similar crystal structurebut with a reverse Ca-Na and Al-Si arrangement,symmetrical to An56. Transmission electron micro-scope observations display anti-phase boundaries(APB's) that record structural transformations, andare indicative of pre-existing but now unstablesingle-phase structures. Wenk et al. noted, that thesuperstructure produced by ordering is maintainedregardless of chemical composition, and that theboundary between labradorite and andesine istherefore a kind of chemical APB superposed in theintermediate plagioclase superstructure with peri-odic APB's on a smaller scale which extend throughboth structures. They also pointed out that theorientation of the feldspars makes this intergrowthdifficult to see in ordinary thin sections and that itmight go undetected in many rocks.

Even more complex relations were described byWenk and Wenk (1977) from a single outcrop fromVal Carecchio in the central Alps. In addition tostepwise change in An-content from An13_25 inleucocratic gneiss to Anes*3 in marble bands, theydescribed microscopic intergrowths of andesine

Anr:=o with labradorite An67*5 (as in Wenk et al.,1975) and of labradorite An62-76 with bytownite/anorthite Ans8-sz. In one sample three plagioclasesare intergrown.

Phillips et al. (1977) found intergrowths of An61and An6e.5 in plagioclase from a plagioclase-quartz-biotite gneiss from Broken Hill, New South Wales.This intergrowth spans the range between B/ggild(An+z and An5s) and Huttenlocher (An67 and Anes)intergrowths, in a region generally considered "ho-mogeneous". Grove (1977) studied plagioclases involcanic, shallow plutonic, deep-seated plutonic,and metamorphic environments, and calculated ap-proximate cooling rates. All plagioclases were inthe range An65*Aq5. In the most rapidly cooledplagioclases there was no evidence of exsolution.With progressively slower cooling the compositionrange of two-phase intergrowths increases. In all ofthe plagioclases there is evidence of local disequi-librium, and there may be second generation lamel-lae with variation of composition on a submicro-scopic scale. Grove feels that "Huttenlocherplagioclases" are useful as indicators of coolinghistory.

Garrison (1978) found eight discrete domainswithin the compositional range of the B/ggild andHuttenlocher intervals in metamorphosed igne-ous plagioclases in metabasalts from the LlanoUplift, central Texas. It is estimated that the (shal-low) amphibolite-grade metamorphism took placenear 3.5 kbar and 650' C (-1200 m.y. ago), and thecompositions of the eight domains are Anrr,An42-44, An55 5s, An6s, An65, An7s, Anss, andAnaa-sr. He suggested that (1) both B/ggild andHuttenlocher intergrowths could be present in thesame zoned grain, (2) the immiscibility region maybe extremely complex, allowing more than the fourmetastable end-member compositions to exist, or(3) a disequilibrium situation exists within the un-mixing intervals due to unfavorable metamorphicconditions. Garrison feels the second to be the mostprobable.

One of the tenets of petrology is that equilibriumis closely approached in metamorphic rocks. Obser-vations suggestive of disequilibrium, as above, areoccasionally reported. Hunahashi et al. (1968)found a "surprisingly wide" difference in the An-content of coexisting plagioclases in some plutonicand metamorphic rocks in Japan, even in one thinsection. They later indicated that the compositions4n22,4n27, An33, An43, An5g, An55, An62, An67, andAnTs crystallize more readily than others, and they

GOLDSMITH: REVIEW OF PLAGIOCLASE

correlated these compositions with presumed struc-turally-favored regions of composition, althoughthe concept of local or "mosaic" equilibrium on thescale of a single thin section was also espoused'Byerly and Vogel (1973) studied grain-boundaryeffects accompanying changes in the composition ofplagioclase in the progressively metamorphosedCross Lake gneiss in southern Ontario. The lower-grade plagioclases of the granodiorite are chemical-ly inhomogeneous with zoned rims containing dis-tinct compositional levels of An0-3, Arl7, and An25.As the grade increases, the plagioclase becomesmore homogeneous, with Ans-3 rims dominating.Braun and Miiller (1975) found very large composi-tional variations over distances of 10-20 p.m inlower-grade metamorphosed "basaltoid" and"granitoid" rocks. They stated that only at tem-peratures greater than 600'C (upper amphibolitefacies) is homogenization of the plagioclase (andgarnets) initiated, so that at lower temperaturesthermometry dependent upon An-contents in theplagioclases or Mn/Fe relations in the garnets is notpossible.

Nord et al. (1978) also reported on compositional-ly-zoned plagioclase microporphyroblasts in phyl-lites from southwestern Massachusetts. Rocks fromthe lowest metamorphic grade yield unzoned albite.At slightly higher grade albite cores and thin outerrims have an intervening narrow zone of An13-n. Athigher grades, plagioclase is complexly zoned inone of two ways: either the composition dropsstepwise from an oligoclase to an albite rim, or thecomposition of the core gradually increases out-ward from Anle to Anz+, then abruptly drops to analbite rim. All of these phenomena are in rocks atgrades below the appearance of chloritoid and gar-net. Above the garnet isograd the zoning is simple,with a core of An13 and a gradual outward increaseto a maximum of An25. It is estimated that the rockswere metamorphosed at no more than 350" C, and afew kilometers overburden, well below the peak ofthe peristerite solvus as estimated by Crawford(1966) and others. Metastable crystallization is indi-cated, although there are no data to distinguishwhether a single or double stage of metamorphismwas involved.

Spear (1977, 1980) observed two-plagioclase as-semblages in hornblende-plagioclase calc-silicatesand amphibolites (Mt. Cube Quadrangle, NewHampshire and Vermont) that provide evidence fora miscibility gap in the range An3e*3-Ans8*3 at anestimated temperature of 530" C. There is a system-

atic partitioning of Na and Ca between plagioclase

and the M4 site of coexisting hornblende with Naenriched in the feldspar. The two-feldspar assem-blages are associated with low-Na hornblende, Na: 0.0710.04 in the M4 site. At another locality theassociation albite and oligoclase is manifested by anabsence of tie-lines from hornblende to plagioclasein the range An2-An1a; the estimated temperature is490" C.

Wenk (1979a) further estimated the influence ofmetamorphic grade on the nature of the intermedi-ate plagioclase superstructure (An3rAn7s), chieflyby 0k/ precession X-ray photography, determiningthe wavelength of the superstructure in metamor-phic plagioclases from the central Alps. Wenk ob-served wavelength variation with temperature(metamorphic grade) of a periodic APB structure inthe plagioclases, and considers this to be the resultof a continuous ordering process. The wavelengthchanges from 20 to 70A over the temperature range500-800' C, and Wenk suggested that this may be auseful indicator of metamorphic grade. Wenk(lg7gb) also reported on the assemblage albite(Ano)-anorthite (Anqs), plus actinolite and parga-

site, and clinozoisite. This ultimate plagioclase as-semblage is in low-grade poikiloblastic amphiboliteschists from an ophiolite series near the TertiaryBergell granite of the central Alps. Using TEM,Wenk's conclusion from the APB evidence is thatalbite and anorthite are formed from decompositionof intermediate plagioclase in the amphibolite withpargasitic hornblende coexisting with actinolite,and that the anorthite is not a relic. Wenk is unableto explain the assemblage, but feels that anorthitemust be stable under some metamorphic con-di t ions. l

Finally, and to return to the heterogeneous view,Lindh (1978) attempted a thermodynamic treatmentof the complex system involving plagioclase andpointed out the complexities of plagioclase-zoisitethermometry, stressing the need to know whatdecomposition products are in equilibrium withanorthite. As did many previous writers, he as-sumed a strong stabilizing influence of Na in the

rDexter Perkins III (personal communication) has observed

associations of albite and An65 in amphibolites from Quebec,also in the presence of homblende and a sodic or pargasitic

hornblende. The calcic plagioclase is in blebs at or near the

contact of the original hornblende and the sodic (and Fe-rich)

hornblende derived from it' The presumption is that the albite is

the stable feldspar, and the calcic plagioclase was formed in

association with the secondary sodic amphibole.

650 GOLDSMITH: REVIEW OF PLAGIOCLASE

plagioclase structure, and suggested that the step-wise increase in anorthite content reported by manyis in part caused by a changing mineral paragenesisand not necessarily only by discontinuities in theplagioclase crystal structure.

In addition to being structurally complex, theplagioclase feldspars in any process involving struc-tural or compositional change must undergo diffu-sive reorganization that requires breaking the verystrong Si-O and Al-O bonds (Goldsmith, 1952). Therefractory behavior of plagioclases, coupled withmultiple levels or even a continuum of metastablestates and a variety of paths available toward vari-ous lower-energy ordered configurations, breedscomplexity and variability. Newton et al. (1980), ininvestigating the thermochemistry of high structuralstate plagioclases, observed "fine structure" in theenthalpy data, particularly in hydrothermally crys-tallized specimens. It was stated that the variationsmay be real, inasmuch as plagioclase is a verycomplex substance and could well show varioussmall thermodynamic "events" at various composi-tions, such as local ordering. Recent and currentobservations on plagioclases are rapidly enlargingour descriptive bank, even if understanding lagssomewhat behind. Omitted from this review are allthe theoretical and structural papers, as well as ahost of recent transmission electron microscope(TEM) and associated electron diffraction (and X-ray diffraction) studies (going back to Mclaren,1974; McConnell, 1974a, b; Nissen, 1974; Nord etal., 1974), even though they may relate to thesubject of plagioclase stability.

AcknowledgmentThe writer's research is supported by NSF grant EAR 7g13675

(Geochemistry Program).

ReferencesAckermand, D. and Morteani, G. (1973) Occurrence and break-

down of paragonite and margarite in the Greiner SchieferSeries (Zillerthaler Alps, Tyrol). Contributions to Mineralogyand Petrology, 40, 293-304.

Ambrose, J. W. (1936) Progressive kinetic metamorphism in theMissi series near Flinflon, Manitoba. American Journal ofScience, 32. 257-286.

Becke, F. (1913) Uber Mi.reralbestand und Struktur der Kristal-linischen Schiefer. Akademie der Wissenschaften, Vienna.Mathematisch-Naturwissenschaftliche Klasse. Denkschiften,75, l-st.

Bird, D. K. and Helgeson, H. C. (1981) Chemical interaction ofaqueous solutions with epidote-feldspar mineral assemblagesin geologic systems. II. Equilibrium constraints in metamor-

phic-geothermal processes. American Journal of Science, 281.576-614.

Braun, E. and Miiller, G. (1975) Zur chemischen Variabilitiitregionalmetamorph gebildeter Plagioklase, Epidote und Gran-ate. Contributions to Mineralogy and Petrology, 52, 193-211.

Brown, W. L. (1960) The crystallographic and petrological sig-nificance of peristerite unmixing in the acid plagioclases.Zeitschrift ftir Kristallographie, I I 3, 330-344.

Brown, W. L. (1962) Peristerite unmixing in the plagioclases andmetamorphic facies series. Norsk Geologisk Tidsskrift, 42(Feldspar volume), 354-382.

Byerly, G. R. and Vogel, T. A. (1973) Grain boundary processesand development of metamorphic plagioclase. Lithos, 6, 183-202.

Chao, S. H. and Taylor W. H. (1940) Isomorphous replacementand superlattice structures in the plagioclase felspars. RoyalSociety (London) Proceedings, series A, 175,76,87.

Christie, O. H. J. (1959) Note on the equilibrium between plagio-clase and epidote. Norsk Geologisk Tidsskrift, 39,268-271.

Christie, O. H. J. (1962) Feldspar structure and the equilibriumbetween plagioclase and epidote (discussion). American Jour-nal of Science, 260, 149-153.

Cole, W. F., S<irum, H. and Taylor, W. H. (1951) The structureofthe plagioclase felspars, I. Acta Crystallographica, 4,20-29.

Cooper, A. F. (1972) Progressive metamorphism of metabasicrocks from the Haast Schist group of southern New Zealand.Journal of Petrology, 13, 457-492.

Crawford, M. L. (1966) Composition of plagioclase and associat-ed minerals in some schists from Vermont, U.S.A., and SouthWestland, New Zealand, with inferences about the peristeritesolvus. Contributions to Mineralogy and Petrology, 13, 269-294.

DeVore, G. W. (1956) Al-Si positions in ordered plagioclasefeldspars. Zeitschrift ftir Kristallographie, 107, 247 -264.

Doman, R. C., Cinnamon, C. G. and Bailey, S. W. (1965) Struc-tural discontinuities in the plagioclase feldspar series. Ameri-can Mineralogist, 50, 724J 40.

Evans, B. W. (1964) Coexisting albite and oligoclase in someschists from New Zealand. American Mineralogist, 49, 173-179.

Fleet, S. G. and Ribbe, P. H. (1965) An electron-microscopestudy of peristerite plagioclases. Mineralogical Magazine, 35,165-176.

Frey, M. and Orville, P.M. (1974) Plagioclase in margarite-bearing rocks. American Journal of Science, 274,31-41..

Garrison, J. R., Jr. (1978) Plagioclase compositions from meta-basalts, southeastern Llano uplift: plagioclase unmixing duringamphibolite-grade metamorphism. American Mineralogist, 63,143-149.

Gay, P. and Smith, J. V. (1955) Phase relations in the plagioclasefelspars: composition range An6 to An7s. Acta Crystallograph-ica. 8. 64.

Ghent, E. D. and De Vries, C. D. S. (1972) Plagioclase-garnet-epidote equilibria in hornblende-plagioclase bearing rocksfrom the Esplanade Range, British Columbia. Canadian Jour-nal of Earth Sciences. 9. 618-635.

Goldsmith, J. R. (1952) Diffusion in plagioclase feldspars. Jour-nal of Geology, 60,288-291.

Grove, T. L. (1977) Structural characterization of labradorite-bytownite plagioclase from volcanic, plutonic and metamor-phic environments. Contributions to Mineralogy and petrolo-

GOLDSMITH: REVIEW OF PLAGIOCLASE 651

ey,64,273-302.Hrirmann, P-K. and Raith, M. (1973) Bildungsbedingungen von

Al-Fe (IIIFEpidoten. Contributions to Mineralogy and Petrol-ogy, 38, 307-320.

Htiy, T. (1976) Calc-silicate isograds in the Riondel area, south-eastern British Columbia. Canadian Journal of Earth Sciences,13, 1093-il04.

Hunahashi, M., Kim, C. W., Ohta, Y. and Tsuchiya, T. (1968)

Co-existence of plagioclases of different compositions in someplutonic and metamorphic rocks. Lithos, l, 356-373.

Kohler, A. (1942a) Die Abhdngigkeit der Plagioklasoptik vomvorangeagargen Wtirmeverhalten. Tschermaks Mineralo-gische und Petrographische Mitteilungen , 53, 24-49

K<ihler, A. (1942b) Drehtischmessungen an Plagioklaszwillingvon Tief-und Hochtemperaturoptik. Tschermaks Mineralo-gische und Petrographische Mitteilungen , 53, 159-179.

Kretz, R. (1963) Note on some equilibria in which plagioclaseand epidote participate. American Journal of Science, 261,973-982.

Larsson, W. (1941) Petrology of interglacial volcanics from theAndes ofnorthern Patagonia. Bulletin of the Geological Insti-tution of the University of Upsala, 28, 19l-405.

Laves, F. (1951) Relationship between exsolved plagioclase andits host. American Crystallographic Association, (abstr.)

Washington Meeting, 33.Laves, F. (1954) The coexistence of two plagioclases in the

oligoclase compositional range. Journal of Geology, 62,409-4l l .

Lindh, A. (1978) Aspects on plagioclase-epidote thermometry'Neues Jahrbuch fiir Mineralogie Abhandlungen, 133, 113-131.

Lyons, J. B. (1955) Geology of the Hanover quadrangle, NewHampshire-Vermont. Geological Society of America Bulletin,66. 105-146.

McConnell, J. D. C. (1974a) Analysis of the time-temperature-transformation behavior of the plagioclase feldspars. In W. S.MacKenzie and J. Zussman, Eds., The Feldspars, p. 460-477 .Manchester University Press, Manchester, England.

McConnell, J. D. C. (1974b) Electron-optical study of the finestructure of a schiller labradorite. In W. S. MacKenzie and J.Zussman, Eds., The Feldspars, p. 478-490. Manchester Uni-versity Press, Manchester, England.

Mcl-aren, A. C. (1974) Transmission electron microscopy of thefeldspars. In W. S. MacKenzie and J. Zussman, Eds., TheFeldspars, p. 378423. Manchester University Press, Man-chester, England.

Misch, P. (1968) Plagioclase compositions and non-anatecticorigin of migmatitic gneisses in northern Cascade Mountainsof Washington State. Contributions to Mineralogy and Petrol-ogy, 17, l-70.

Morteani, G. and Raase, P. (1974) Metamorphic plagioclasecrystallization and zones of equal anorthite content in epidote-bearing, amphibole-free rocks of the western Tauernfenster,eastern Alps. Lithos, 7, 101-111.

Newton, R. C., Charlu, T. V. and Kleppa, O. J. (1980) Thermo-chemistry of the high structural state plagioclases. Geochimicaet Cosmochimica Acta, 44,933-941.

Nissen, H.-U. (1974) Exsolution phenomena in bytownite plagio-clases. In W. S. MacKenzie and J. Zussman, Eds., TheFeldspars, p. 491-521. Manchester University Press, Man-chester, England.

Noble, D. C. (1962) Plagioclase unmixing and the lower bound-

ary of the amphibolite facies. Journal of Geology, 70 ' 234-240.

Nord, G. L., Jr., Heuer, A. H. and Lally' J. S. (1974) Transmis-

sion electron microscopy of substructures in Stillwater bytow-

nites. In W. S. MacKenzie and J. Zussman, Eds', The Feld-

spars, p. 522-535. Manchester University Press, Manchester,

England.Phillips, E. R., Chenhall, B. E., Stone, I. J. and Pemberton,

L W. (1977) An intergrowth of calcic labradorite in a plagio-

clase-quartz-biotite gneiss from Broken Hill, New South

Wales. Mineralogical Magazine, 41, 469-47 1.

Rambaldi, E. R. (1973) Variation in the composition of plagio-

clase and epidote in some metamorphic rocks near Bancroft,

Ontario. Canadian Journal of Earth Sciences, l0' 852-868.

Ramberg, H. (1943) En underspkelse av Vestrandens regional-

metamorfe bergarter. Norsk Geologisk Tidssdrift, 23, l-17 4.

Ramberg, H. (1949) The facies classification of rocks: a clue to

the origin of quartzo-feldspathic massifs and veins. Journal of

Geology, 57, 18-54.Ramberg, H. (1952) The Origin of Metamorphic and Metasomat-

ic Rocks. University of Chicago Press, Chicago, Illinois.

Ribbe, P. H. (1960) An X-ray and optical investigation of the

peristerite plagioclases. American Mineralogist, 45' 626-644.

Ribbe, P. H. (1962) Observations on the nature of unmixing in

peristerite plagioclases. Norsk Geologisk Tidssdrift, 42 (Feld'

spar volume), 138-151.Rutland, R. W. R. (1961) The control of anorthite content of

plagioclase in metamorphic crystallization. American Journal

of Science, 259,76J9.Rutland, R. W. R. (1962) Feldspar structure and the equilibrium

between plagioclase and epidote: a reply. American Journal of

Science, 260, 153-157.Sen, S. K. (1963) Some consequences ofordering and unmixing

in sodic plagioclases on epidote-plagioclase equilibrium in

regionally metamorphosed rocks. American Journal of Sci-

ence. 261. 786J89.Spear, F. S. (1977) Evidence for a miscibility gap in plagioclase

feldspar in the composition range An3e-Anss. Carnegie Institu-

tion of Washington Yearbook , 76, 619-621.Spear, F. S. (1980) NaSi = CaAl exchange equilibrium between

plagioclase and amphibole. Contributions to Mineralogy and

Petrology, '72, 33-41.

Taylor, W. J., Darbyshire, J. A. and Strunz, H. (1934) An X-ray

examination of the feldspars. Zeitschrift ftir Kristallographie,

8',1,464-498.Tsuji, S. (1966) Possible effects of peristerite solvus in plagio-

clase of the Higo metamorphics, Kyushyu, Japan. Journal of

the Geological Society ofJapan, 72' 63-68.Waard, D. de (1959) Anorthite content of plagioclase in basic and

pelitic crystalline schists as related to metamorphic zoning in

the Usu massif, Timor. American Journal of Science, 257,

553-562.Wenk, E. (1962) Plagioklas als Indexmineral in den Zentralalpen

Die Paragenese Calcit-Plagioklas. Schweizerische Mineralo-gische und Petrographische Mitteilungen ' 42, 139-152.

Wenk, E. and Keller, F. (1969) Isograde in Amphiboliterserien

der Zentralalpen. Schweizerische Mineralogische unde Petro-graphische Mitteilungen, 49, 157-198.

Wenk. E. and Wenk, H.-R. (1977) An-variation and intergrowths

of plagioclases in banded metamorphic rocks from Val Carrec-

hio (Central Alps). Schweizerische Mineralogische und Petro-graphische Mitteilungen, 57, 4l-57 '

652 GOLDSMITH: REVIEW OF PLAGIOCLASE

Wenk, E., Wenk, H.-R., Glauser, A' and Schwander, H. (1975) Wenk, H.-R. (1979b) An albite-anorthite assemblage in low-Intergrqwth of andesine and labradorite in marbles of the grade amphibolite facies rocks. American Mineralogist, 64,central Alps. Contributions to Mineralogy and petrology, 53, l2g4_l2gg-3ll-326.

Wenk, H.-R. (1979a) Superstructure variation in metamorphic Manuscript received, September 16, IggI;plagioclase. American Mineralogist, 64,71-76. acceptedfor publication, Marchg, 19g2.