American Mineralogist, Volume 70, pages 1l0l-l l l3, 1985

The Oka carbonatite complex, Quebec: geology and evidence forsilicate-carbonate liquid immiscibility I

AlraN H. TnrnvreN2 .q.No Enlc J. EssENr

Department of Geological SciencesUniuersity of Michigan

Ann Arbor, Michigan 48109-1063

Abstract

Associated carbonate and silicate igneous rocks from the northern intrusive center of theOka complex (Quebec, Canada) probably formed from immiscible carbonate and silicatemagmas in situ, as their structure, petrography, isotope ratios and mineral chemistry areconsistent with immiscibility. The intrusion consists of intimately intermixed rocks derivedfrom carbonate magma (some jacupirangites, okaites, melilite-bearing carbonatites) and sili-cate magma (some jacupirangites, melilitites). Some jacupirangites and okaites are crystalcumulates. Inferred order of crystallization is: jacupirangites; nepheline okaite and nephelinemelilitite; okaite and melilitite; and melilite-diopside carbonatite and diopside melilitite.Carbonate-rich ocelli are common in melilitites, and silicate-rich ocelli are present in rocksderived from carbonate magma. Rare earth elements are six times more abundant in carbon-atite than in melilitite, consistent with REE partition experiments on immiscible carbonateand silicate liquids. Initial strontium isotope ratios (-0.7033) are identical in silicate andcarbonate rocks, as required by immiscibility. The strongest evidence for immiscibility comesfrom identity of compositions of magmatic minerals in rocks derived from contemporaneousimmiscible fractions (required by liquid immiscibility, but inconsistent with other modes offractionation). Fassaite pyroxenes in the jacupirangites (derived from both silicate-rich andcarbonate-rich magmas) span identical compositional ranges (9.3-L6o/o Alro.; 3.24j%TiOr). Diopsidic pyroxenes in carbonatites and melilitites also have identical ranges (1.45-2.0o/o AlrOt;0.314.45% TiOr). Perovskites with magmatic compositions are preserved onlyin nepheline okaite and nepheline melilitite; here, the perovskites are of identical composition.other minerals (calcite, melilite, nepheline) do not preserve magmatic compositions.

Introduction

Silicate-carbonate liquid immiscibility has been con-sidered an important process in carbonatite genesis (e.g.,Koster van Groos and Wyllie, 1963; Middlemost, l9Z4;LeBas, 1977), primarily because of the common associationof carbonatites and alkaline silicate rocks. Experimentalstudies indicate that immiscible silicate and carbonatemagmas are likely in geologically reasonable compositions(Koster van Groos and Wyllie, 1966, 1968, 1973; Kostervan Groos, 1975b; Verwoerd, 1978; Freestone and Hamil-ton, 1980; Harris and Koster van Groos, 1981), and im-miscible silicate and carbonate magmas may be preservedin melt inclusions (Rankin and LeBas, 1974) and in pyro-clastics (Hay, 1978; Mariano and Roeder, 1983). The casefor liquid immiscibility in the genesis of plutonic carbon-atite complexes is less certain because magmatic and sub-

I Contribution No. 413 from the Mineralogical Laboratory,The University of Michigan.

solidus reactions may have destroyed much of the evidencethat is available in experimental charges or volcanic rocks.

The Oka pluton, a Cretaceous intrusive complex of theMonteregian Hills petrographic province (Gold, 1963,1972; Gold et al., 1967; Gold and Vall6e, 1969; Fig. 1)includes a wide range of silicate and carbonate rocks, isfairly well exposed, and has not been deformed or meta-morphosed. Many workers (e.g., Eby, 1975: Koster vanGroos, 1975b; Cullers and Medaris, 1977; Verwoerd, 1978)have infered that immiscibility between silicate and car-bonate magmas was important at Oka, but these studiesdid not specifically state which lithologies formed fromcontemporaneous immiscible magmas. Nor did any ofthese studies consider more than one line of evidence tosupport their conclusions. This paper evaluates liquid im-miscibility in the Oka complex using all pertinent field,petrographic and chemical data, and delineates which li-thologies may have been formed contemporaneously by insitu immiscibility between silicate and carbonate magmas.

Geology2 Current Address: Departnient of Geology, Boston University, The geology of the Oka complex is described by Gold

Boston, Massachuserts 02215. (1963, 1972) and Gold and vall6e (1969): their work forms

0003-{04x/E5/t I 12-1 101$02.00

1102

f l Moderate-Sodium Carbonati teFrrH Low -Sodium Carbonati teE Pslem11s Carbonati te

I n l fa l ineS i t i ca te Rocklllij 41n611sF,,il cneiss

Fig. 1. Simplified geologic map of the Oka pluton, modifiedfrom Gold (1963, 1972) and Gold and Vallee (1969). The complexis emplaced into Grenville-age metamorphic rocks and is cut byalnciite breccia pipes. The complex is composed of four plutons intwo intrusive centers, northern and southerh. On the map, the twocenters are not differentiated. At each intrusive center. the earlierpluton is a ring-complex of "moderate-alkali carbonatite" (con-taining the alkali-bearing silicate minerals melilite, nepheline, ac-mitic pyroxene, atdlor alkali amphibole) and alkaline silicaterocks (acupirangites, okaites, melilitites, ijolites and their alter-ation products). Cutting each ring-complex is a pluton of "low-alkali carbonatite" (containing no alkali-rich silicate mineralsexcept phlogopite and/or hornblende). Sample locations shownare: A, Husereau Hill; B, Bond Zone;and C, Open Pit A2. SeeAppendix for exact sample locations.

the basis for Figure I and the following description. TheOka complex is a composite pluton emplaced into meta-morphic rocks and anorthosite 40 km west of Montreal,Quebec. The complex includes two distinct intrusive cen-

Fig. 2. Outcrop of dolite/melilitite (dark) and melilite-diopsidecarbonatite (light). Note carbonatite foliation wrapping aroundijolile. Field of view is 2 meters across. Trench 7, Bond, Zone(southwestern margin of complex; Gold, 1972).

TREIMAN AND ESSENE: OKA ALKALINE COMPLEX, LIQUID IMMISCIBILITY

ters, which give the complex its "figure-eight" plan. Bothintrusive centers include an earlier outer annulus of alkalicsilicate-rich rocks and carbonatite 0abelled "moderate-

alkali carbonatite" on Fig. l; Fig. 2), and a later centralplug of carbonatite (labelled "low-alkali carbonatite" onFig. 1). The complex is cut by alnoitic breccia pipes anddikes of dolomitic carbonatite (Treiman and Essene, 1984).All of the following descriptions and ilata refer to the north-ern intrusiue center at Oka. The rock names used here(Table 1) have been chosen to be compatible with previous

descriptions of the Oka complex, and do not always followthe recommendations of Streckeisen (1980). Many of thesilicate-rich rocks at Oka are alteration products, and theseare not discussed here.

Petrography

Jacupirangites are defined as rocks consisting predomi-nantly of titanaugite (see Williams et al., 1954, p. 82 fordefinition of type jacupirangite); at Oka, the pyroxene is atitaniferous fassaite (Peacor, 1967). Two varieties ofjacupi-rangite occur at Oka; coarse- and fine-grained. The coarse-grained jacupirangite (C-Jac, Fig. 3a) consists ofcentimeter-sized euhedral fassaite crystals and minor mag-netite in a cumulate texture with interstitial melilite, calcite,and hauyne (probably altered nepheline). The fassaite ispartially altered to an aggregate of phlogopite, melilite andperovskite; alteration is most intense near calcite. Fine-

Table 1. Rock types, sample numbers, abbreviations and averagemodal mineraloeies of Oka rocks

tocL ttp. atrd [ub.t Abbrcv. Uod.l XlEcrllo8y

Coraa Jaccplrr!8lta C-Jac790K12

FlEJ.cuplrroglt . F-J!c7 9019

n ph.ll!. OLrll. il.-OL79tr8

N.ph.IlD. f,clllltlt. nc-x.l790f,r3

Ok.ltc Ok790K14

N.l l I l te-c.rbourlre l {c-C

xcltut l ! . I le lt90K10

t.ltllte-dtop.td. l{c-Dt-CC!rbout l !a

79OX23, EOB9-418 0 8 9 - 7 , 8 0 8 1 1 - 3

Dtop. ld. k l l l . t r l t . Dl-xe160B1-2

I jol l tc IJ7 9 O X 2 5 , 7 9 f r 2 8 ,E089-r2

liontlc.llllF f,Fdt-Cdloprl l l . Carbmtl tc

t0A2-8,60A2-40

F....lte 50-60?, rcltllt€ 5-20U, phlogopt!. I-5t , cr lc l ! . 1- lOl, petov.t l tc 1-3I, h.uyne l-31, pyrrhot l t . t r , , .pr! l t . t r .

Xelllt!! !0-501, n.ph.llre 30-501, phloaoptlc1 0 : , u 8 o c l l t c 5 1 , c r l c l t c 3 l , . p r ! t t ! l t ,p€rowHtc 12, ptrrhot l le tr .

N c p h e l l n c 5 0 I r ! c I t I l t a 3 0 t , p e r o v . k i t . 5 1 ,D r g n c t l t e 5 1 , p h l o S o p t t e 5 2 , c r l c l t a 3 Z ,.pr! I l . t r , , pytrhol l te tr .

l tc l t l t tc 30-701, c. Ic l t . 15-502, !{ne! l t . 5-l0l , p€rd. l l t . l -5t , .Fl l iG 2-51, phlotoplt€1-5U, pFrhol l t . t r . , dlop. ld! tE.

Ar rbd. but el th celcl tc > r l l l t t . .

H . l 1 l l t . 7 0 1 , p h l o g o p t t c l 5 t , h r u F e 5 lperovl l l te: |1, l r t rct l ta 31, pyrrhott tc 2l lepst l ! . 21, t rqsulr t t ! .

c s l c l t e 5 0 - E 0 I , ! c l t l l t r 5 - 3 0 1 , d l o p s l d e 2 -20t, rpBt l te 5-2otr Elocsl l t€ 2-I5t , lsSnett tc1 - 5 2 , p h l o g o p l t . l - 5 1 , p € r o v B k l l . I - 5 1 ,p y r o c h l o r e t r . - 5 1 , p y r r h o t t l . 1 2 , D ! t r o I l t e ,t r . , o o l e a o t r . , S i o a a u L r t r .

DtQstdc 30-501, ! . t t l l ! . 20-5d, phloSopit .5-15t, grd8ular 2-51, pcrov.kl tc 21, [agn.-t l tc, 0-5!, pFrhot l te 2-5?, ep.t i t . t t . ,

pyrochlore tr , , .ph€ne tr .

D t o p . l d . 2 0 - 6 0 2 , ! . l t l l t e 0 - 2 0 t , n e p h e l l n €2 0 - 4 0 2 , g r o . . u l e r 5 1 , p h l o S o p l r e 5 - 1 0 U ,p e r o v s k l t e 2 l , . p . t l t e t r . . p y r o c h l o r . t r . ,loa.ed !r . r olocal l te t ! .

C.tct t . 751, .ont lc. l l l t . 72, D.gn. l l te 5t,p h l o g o p l ! . 5 : , d l o p s t d e 3 t , . p r t l t c 2 : ,pyrochlor. 2U, peroskl te 12, pFrholt t . t r .

grained jacupirangite (F-Jac; Fig. 3b, c) consists of euhe-dral to subhedral fassaite (l mm long) with interstitial, poi-kilitic melilite, phlogopite, hauyne (probably altered neph-eline) and perovskite. Within the fine-grained jacupirangiteare ocelli of coarse-grained, calcite-rich rock (Fig. 3b)which resemble nepheline okaite (see below). Relict fassaitewithin the ocelli suggest that they were originally coarse-grained jacupirangite.

Okaite and nepheline okaite (Ok and Ne-Ok) are coarse-grained rocks (Figs. 3d, e; I cm grain size) composed ofeuhedral to anhedral silicate (melilite, nepheline, hauyne)and oxide (magnetite, perovskite) minerals with interstitialcalcite, apatite, and phlogopite. The proportion of inter-stitial calcite varies from only a few percent (Fig. 3e) tonear 507o (Fig. 3d). The latter rocks are essentially melilitecarbonatites; an arbitrary cutoff between okaite and meli-lite carbonatite is 50% (volume) calcite. Hauyne in theokaites commonly occurs between nepheline and calcite,and is probably reaction product. In the classification ofStreckeisen (1980), the okaites would be called varieties ofmelilitolites.

Melilitite (Mel), nepheline melilitite (Ne-Mel), diopsidemelilitite (Di-Mel), and ijolite (Ij) are fine-grained (0.1-lmm grain size) silicate-rich rocks with hypidiomorphicgranular textures (Figs. 3f, g). Mineral distribution is com-monly irregular, with patchy concentrations of melilite orpyroxene. Hauyne, commonly present, is probably an alter-ation of nepheline. In the classification of Streckeisen(1980), these rocks would be called varieties of micro-melilitolites. As used at Oka, "ijolite" encompasses a widerange of mafic rock types, fine and coarse grained, somewithout nepheline, and some which are almost certainlyaltered okaites. The ijolites referred to here are fine-grained, hypidiomorphic rocks, commonly intermixed withcarbonatite (Fig.2), and are best exposed on the south-westside of the complex. Ijolite is texturally identical to themelilitites discussed above, and differs only by addition ofnepheline, garnet and commonly wollastonite or pectolite,elimination of magnetite, and partial or complete elimina-tion of melilite.

Carbonatites are all coarse-grained, calcite-rich rocks(sovites), containing euhedral to subhedral oxide and sili-cate minerals in a hypidiomorphic matrix of calcite. Atleast four mineralogic varieties are distinguished.Monticellite-diopside carbonatite (Mo-Di-C) occurs asdikes cutting other rocks of the complex and as the centralstock in the northern intrusive center. Melilite carbonatite(Me-C) is gradational to okaite (Fig. 3d) by decrease incalcite content. Melilite-diopside carbonatite (Me-Di-C;Fig. 3h) contains diopsidic pyroxene, commonly adheringto melilite, but also as discrete grains. Melilite is commonlyaltered. The two most common alteration assemblages are:diopside * grossular i calcite + natrolite; and vesuvia-nite * calcite + diopside + natrolite. The grossular-bearing alteration is earlier than the vesuvianite alteration.Late carbonatites have no melilite, but contain diopsidethat is chemically similar to that in the melilite-diopsidecarbonatite.

1 103

Interpretation

The outer annulus of the northern intrusive center atOka, here informally called the Jacupirangite-Okaite-Melilitite-Ijolite-Carbonatite (rourc) suite, consists of in-termixed silicate-rich and carbonate-rich rocks in incom-plete concentric bands. The banding shown on the map(Fig. 1) is a gross simplification of the structure; see Goldand Vall6e (1969) for a more detailed map of the complexas a whole, and Gold (1912) tor maps of individual out-crops. Lithologic units shown on Figure 1 are really zoneswhere that rock type is the most common lithology. Forexample, the outcrop of Figure 2 is in a zone mapped asijolite, although much of the exposure is carbonatite. De-spite the structural complexity, there is some regularity tothe distribution of rock types. Jacupirangites, okaite, andmelilitite are more common to the north and east in theannulus; carbonatite and ijolite are predominant to thesouth and west.

Regular gradations in mineralogy and mineral chemistryamong rocks of the JoMrc suite suggest that they are prod-ucts of a single magma. Based on intrusive contacts andmineralogic changes (Table 2), the sequence of emplace-ment is interpreted to be: jacupirangites; nepheline okaiteand nepheline melilitite; okaite, melilite carbonatite andmelilitite; diopside melilitite and melilite-diopside carbon-atite; and diopside-carbonatite (without melilite). It is notcertain whether ijolite is magmatic and contemporaneouswith the diopside carbonatite or is an alteration of thediopside melilitite. Ijolite could form from diopside meli-litite by replacement of magrnatic magnetite by garnet *perovskite and by replacement (partial or complete) ofmagmatic melilite by diopside * nepheline f wollastonite* calcite. Textures indicative of these reactions are ob-served in the okaites and carbonatites, where they have notprogressed to completion.

Liquid immiscibility in thelourc pluton

The rorrltc intrusion is a good candidate for an investi-gation of silicate-carbonate liquid immiscibility because itconsists of closely associated silicate-rich and carbonate-rich rocks. Among the silicate-rich lithologies, it is essentialto contrast those derived from silicate-rich magmas fromthose formed by crystal accumulation from carbonate-richmagmas.

The only criteria for distinguishing silicate-rich crystalcumulates from solidified silicate magmas are textural.Rocks such as the coarse-grained jacupirangites andokaites (Figs. 3a, d; 3e lower half) may be interpreted ascumulates from carbonate-rich magma because they con-tain large, euhedral crystals of silicate and oxide mineralswith interstitial space-filling carbonate. Other silicate-richlithologies from the Oka complex are fine-grained, havelittle carbonate, and are either phenocrystic with poikiliticsilicate minerals (fine-grained jacupirangite; Fig. 3b) or arehypidiomorphic granular in texture (melilitites and ijolite;Figs. 3e upper half, 3f, g). These rock types are interpretedto have formed from silicate-rich magmas.

TREIMAN,4ND ESSENE: OKA ALKALINE COMPLEX, LIQUID IMMISCIBILITY

I 104 TREIMANAND ESSENE: OKA ALKALINE COMPLEX, LIQUID IMMISCIBILITY

Within this spectrum of rock types which may have beenderived from contemporaneous immiscible silicate-rich andcarbonate-rich melts, four pairs of lithologies can bechosen, based on mineralogy, as follows (Table 2, rockderived from carbonate magma given first): coarse-grainedand fine-grained jacupirangites; nepheline okaite and neph-eline melilitite; okaite and melilitite; and melilite-diopsidecarbonatite and pyroxene melilitite. Because all of the pairsof lithologies are interpreted as products of a singlemagma, evidence from one pair is permissive evidence forthe rest.

Immiscibility between magmas, even if not observed di-rectly, should leave structural and chemical imprints onrocks derived from the magmas; such imprints may beused as tests for an origin by liquid immiscibility. Amongthe proposed imprints and tests are: textures and field rela-tions; major element rock chemistry; trace element andlight isotope chemistry; radio-isotope ratios (including ra-diochronology); and mineral chemistry. These tests are ap-plied to the above pairs of lithologies from the rourcpluton, and applicable tests are satisfied in so far as dataand preservation of magmatic characteristics allows. Notall of these tests are of equal value, and the limitations ofeach test will be discussed.

I 105

Textures

Immiscible melts will have a phase boundary, a me-niscus, between them; the presence of menisci or relict me-nisci constitutes a permissive test for in situ formation asimmiscible melts. Phase relations in immiscible silicate-carbonate melt systems (e.g., Freestone and Hamilton,1980) show decreasing mutual solubility with decreasingtemperature, so that cooling of immiscible silicate and car-bonate melts will produce an emulsion of droplets of onemelt within drops of the other melt. On cooling, such anemulsion should produce an ocellar texture.

Many,rocks of the rourc pluton do exhibit ocellar tex-tures that may reasonably be interpreted as relics of im-miscible magmas. Rocks solidified from silicate-rich melt(fine-grained jacupirangite, melilitite, ijolite) commonlycontain ellipsoidal ocelli (Figs. 3b, e, f, g) which are richerin calcite and coarser grained than their silicate-rich hosts.The compositions and textures of the ocelli resemble thoseof the carbonatites.

Evidence for silicate-rich ocelli in the carbonatites is lim-ited. The best examples are the melilitite stringers in cumu-late okaite (Fig. 3e); okaites are interpreted as cumulatesfrom carbonate magma. The melilitite stringers do notcross-cut the okaite, but fit among its large crystals. The

TREIMAN AND ESSENE: OKA ALKALINE COMPLEX, LIQUID IMMISCIBILITY

G) (h)

Fig. 3. Thin-section photomicrographs of routc rocks; plane liglt. (a) Coarse-grained jacupirangite, consisting of euhedral zonedtitanaugite (dark) with interstitial calcite and melilite (light). Sample 790K9. Scale bar 2 mm. (b) Fine-grained jacupirangite; ocellus ofcalcite-rich, coarser-grained material (top right) similar in texture to coarse-grained jacupirangite or okaites. Sample 790K12. Scale bar 2mm. (c) Fine-grained jacupirangite, detail of texture. Euhedral titanaugite (Px) poikilitically enclosed in melilite, pholgopite, andperovskite. Sample 790K12. Scale bar 1 mm. (d) Calcite-rich okaite, with euhedral melilite (light) enclosed poikilitically in calcite (darker,stained). Sample 790K14. Scale bar 1 mm. (e) Stringer of nepheline melilitite (top, small grain size) in nepheline okaite (bottom, largegrains). Euhedral melilite (Me) and interstitial calcite (Cc) of the nepheline okaite abut on melilitite; melilitite does not cross-cut okaitetexture. Coarser calcite-rich patch (arrow) in nepheline melilitite is texturally similar to okaite (Fig. 3d). Sample 790K8. Scale bar 1 mm.(f) Melilitite, with typical hypidiomorphic texture to lower right, melilite phenocryst at left, and coarser ocellus (in center), rich in calcite(orthogonal cleavage traces). Sample 790K10. Scale bar ll2 mm. (g) Diopside melilitite, with diopsidic pyroxene (intermediate gray),melilite and calcite (light), and perovskite and garnet (dark). Coarser, calcite-rich ocellus between arrows. Sample 8081-2. Scale bar 2mm. (h) Melilite-diopside carbonatite, with phenocrysts of melilite (Me) and diopside (Di) in calcite (with twin lamellae and cleavage).Sample 8089-7. Scale bar 1 mm.

1 106

Table 2. Silicate rnineralogy of routc rocks*, in inferred order ofemplacement

Rock fype !l1nere13

T l tasug l te Nephe l te Me l t l t te DtoPstde l fo l l la to [1 te ]Garnet

c - J e c , F - J e x x x

}le-OL, Ne-Uel X X

u€-c, ok' uel x

UeDt-C, Dt-Mel x X

r J x ( x ) x x

Pare l theseg tnd lca te P teaence ln soE sdPlea '

r A11 ouples cotrteln phloSoPlte.

stringers contain carbonate-rich ocelli, and are thus likeother rocks derived from silicate magma. A possible inter-pretation of this okaite is that it formed from a carbonatemagma as a cumulate of silicate crystals and silicate

magma droplets.The intimate mixture of dolite and carbonatite in some

portions of the roulc pluton (Fig. 2) may be interpreted as

representing large ocelli of ijolite magma in carbonatitemagma, but field relations are not definitive. Rock in this

area is strongly deformed (note foliation in Fig. 2 bending

around ijolite; Gold, 1963, 1972), and the ijolite-

carbonatite contacts are replaced by phlogopite.Thus, rock textures and structures in the rourc pluton

are consistent with in situ immiscibility between silicateand carbonate magmas. Ocellar or emulsion textures arenot preserved in all rock types, particularly in the carbon-

atites, but are common enough to suggest that liquid im-miscibility was an important process in the routc pluton.

The rarity in the carbonatites of textures indicative of sili-

cate magma droplets may be ascribed to two possible ef-

fects: carbonatite magma has low viscosity (Treiman and

Schedl, 1983), so silicate droplets may settle quickly; and

carbonatite magma may dissolve little silicate, and thus

produce few immiscible droplets on cooling.

BuIk composition

If the bulk compositions of associated carbonatite and

silicate rocks are similar to compositions of known immisc-

ible liquids, one has permissive evidence for a liquid immis-

cibility relationship between the rock types (e.g., Kostervan Groos, 1975a; Verwoerd, 1978; Freestone and Hamil-ton, 1980). This test of liquid immiscibility is difficult to

apply because: igneous rocks may not retain the compo-sitions of their parental magmas, and definitely haue losttheir magmatic volatiles; and comparable compositionswhich show immiscibility may not have been studied,

either in nature or the laboratory. For the JoMIc pluton,

both problems prevent use of bulk compositions as a test

of silicate-carbonate liquid immiscibility.Rocks of the rolttc pluton clearly do not retain their

magmatic volatiles (possibly including alkali carbonates)and many are interpreted as crystal cumulates. Examples

include the coarse-grained jacupirangite (Fig. 3a) and the

okaites (Figs. 3d, e). The melilite-diopside carbonatite (Fig.

3h) may represent a magma composition, or be a calcite-

rich cumulate (cf. Nesbitt and Kelly, 1977; Twyman and

TREIMANAND ESSENE: OKA ALKALINE COMPLEX, LIQUID IMMISCIBILITY

Gittins, 1982). Lacking magmatic inclusions, one cannot

tell if the melilite-diopside carbonatite retains a magmatic

composition.The test of bulk compositions is difficult to apply to the

lovrc rocks because no relevant experiments have been

reported and no relevant natural samples have been stud-

ied. Most experimental studies of silicate-carbonate liquid

immiscibility have involved feldspar-bearing synthetic sys-

tems (Koster van Groos and Wyllie, 1966, 1968, 1973;

Harris and Koster van Groos, 1983), or syenitic or rhyolitic

natural compositions (Koster van Groos, 1975a, b; Free-

stone and Hamilton, 1980). Two studies which found

silicate-carbonate liquid immiscibility in systems contain-

ing ijolitic or nephelinitic magma also used Na-rich car-

bonate, and are inapplicable to the JoMIc rocks because

their silicate melt fractions contained 12-19 wt.% NarO,

while the rorr,rrc melilitites and ijolite contain only 3-5

wt.% NarO (Verwoerd, 1978; Freestone and Hamilton,

1980: Table 3)!Natural samples showing possible immiscibility between

carbonatite and ijolite or melilitite are not suffrciently well-

known for use as references in tests of liquid immiscibility.

Most of these have the same status as the rorralc rocks:

unproven. A few pyroclastic rocks contain spheres of meli-

litite in a carbonatite matrix (LeBas, 1977; Hay, 1978;

Mariano and Roeder, 1983), but necessary bulk and

magma compositions are unavailable.

Trace elements and light isotoPes

As with major elements, trace elements and light iso-

topes may be fractionated in characteristic patterns be-

tween immiscible melts and, thus, form a test of origin by

liquid immiscibility. In principle all trace elements and

light isotopes might serve as tests for immiscibility, but so

far only data on rare earth element partitioning are appli-

cable to the JoMIc rocks. Partition of Sr between immisc-

ible silicate and carbonate melts has been studied, but only

for silica-rich compositions inapplicable to the rowc rocks

(Koster van Groos, 1975b). Light isotope ratios are avail-

able for rocks of the roulc pluton (Deines and Gold' 1973);

no partition data are known.Rare earth element distributions between the silicate and

carbonatite rocks of the roulc pluton are consistent with

derivation by liquid immiscibility, although this test is

weak. The rare earth elements (REE) are partitioned into

carbonate melt over coexisting silicate melt, with D(REE)- 2 3 (Cullers and Medaris, 1977; Wendlandt and Harris-

on, 1980). Carbonatites and ijolites of the routc pluton

have D(REE) - 5-6 (Eby, 1975;Table 3, Fig.4). Similarity

of REE distribution between the loutc rocks and the REE

partition between known immiscible fractions may be

taken as evidence of silicate-carbonate liquid immiscibility

in the routc pluton (Cullers and Medaris, 1977)-

But REE partition as a test of liquid immiscibility has

two serious problems. First, it is not certain that any of the

lorurc carbonatites retain magmatic compositions. Second,

most carbonatites in the world have similar REE content(Loubet et al., 1972\ and could be interpreted as immiscible

TREIMAN ,4ND ESSENE: OKA ALKALINE COMPLEX, LIQUID IMMISCIBILITY

Table 3. Chemical analyses of rocks from the Oka complex

rt07

Mel Ok Ne-llel

7 9 0 K 1 0 7 9 0 K . 4 7 9 0 K 1 3

Ne-Ok C-Jac

790K8 790K12 790K23

lteDi-C

80B9-4 80B9-7

I J790K25 790K28 8089-12

MqDt-C

80A2-8

s to2 3r .3 22 .9T l o 2 2 . 0 8 I . 1 3z to2 0 .04 0 .00A l r o r 1 1 . 3 5 . 8 8rnEn io , 0 .28 0 .26F e O I I . 3 1 8 . 1H n O 0 . 5 3 1 , 5 0H g o 5 . 7 4 5 . 2 LC a o 2 2 . 3 2 7 . 5S r 0 0 . 4 1 O . 7 3B a O 0 . 3 0 0 . 3 2N a 2 O 4 . 3 4 1 . 6 3K 2 0 1 . 1 4 O . 4 7P 2 O 5 2 . 0 0 0 . 6 9szoS 0 .05 0 .04L O I * t . 2 3 1 0 . 6Tota l 94 .53 96 .96

49L!3 563l28L!4 L279

384133 3254 5 . 5 ! . 2 3 4 . 01 6 . 3 1 . 3 1 0 . 34 . 4 ! . 4 2 . A6 . 3 1 . 3 4 . 70 . 5 1 . 4 < 0 . 5

34.2 33 .41 . 8 8 1 . 1 10 . 0 5 0 . o 2

t 4 . 4 1 8 . 4o , 2 0 0 . 1 3

1 0 . 9 9 . O 40 . 5 6 0 . 4 54 , 1 6 2 . 7 5

2 0 . 5 1 6 . 80 . 4 9 0 . 4 60 . 0 0 0 . 0 55 . 4 6 7 . 3 8r . 8 9 2 . 9 5L . 2 3 0 . 6 10 . 0 8 0 . 0 6o . 7 7 6 , 9 2

96.77 rOO.s3

418 286981 587240 1342 L , r 2 3 . 91 l . r 7 . 43 . 6 2 . r5 , 2 3 . 9

< 0 . 4 0 . 3

3 4 . 2 1 6 . 23 . 5 0 0 . 6 90 . 0 5 0 . 0 6

1 0 . 5 1 . 3 50 . 1 3 0 . 7 0

1 5 . 2 1 0 . 9o . 2 3 1 . 0 39 . 0 1 2 . 2 9

2 r . o 3 6 . 60 . 0 8 0 . 7 50 . 0 5 0 . 0 60 . 4 8 0 . 5 50 . 5 3 0 , 1 1r . 0 5 2 . 2 3o . o 2 2 . 4 4r , 0 8 t 7 , 5

9 7 , 1 5 9 3 . 5 6

<6 L278269 3184<84 877

r 5 . 7 L 2 6 .5 . 7 4 0 . 8

< r . 4 L 2 . 51 . 8 9 20 . 3 3 . 6

r 2 . 8 4 . 2 30 . 4 9 0 . 0 50 . 0 5 0 . 0 01 , 0 8 0 . 4 31 . 3 3 0 . 8 69 , 8 8 1 . 5 90 . 8 6 0 . 2 4L . 7 9 0 . 1 4

37.7 49 .40 . 8 2 r . 4 00 . 0 6 0 . 0 00 . 5 0 0 . 1 60 . r l 0 . 0 56 . 4 L 4 . 4 62 . 0 6 0 . 1 0

t 4 . 4 3 3 . 88 9 . 8 9 9 7 . 5 r

2672 192867L7 41901572 11702 0 8 . t 2 6 .

5 6 . 5 1 1 . 71 5 . 9 . 58 0 t 4 . 2< 0 . 9 0 . 9

40.4 37 .5r . 0 0 0 . 5 50 . 0 3 0 . 0 1

1 0 . 2 1 1 . 1o , 2 2 0 . t 98 . 5 6 5 . 4 7r . 2 3 1 . 0 97 . 7 3 6 . 7 3

20.6 20.70 . 3 0 0 . 3 50 . 0 0 0 . r 33 . 8 9 4 . 4 22 . 4 7 2 . 1 0L . 2 9 0 . 9 00 . 0 0 0 . 1 82 . 3 1 6 . t 6

9 9 . 8 0 9 7 . 2 2

440 386774 902645 2AO

45.4 37 .61 5 . 5 L 2 . 45 . r 3 , 9

t 2 . 5 8 . 82 . 4 1 . 8

4 0 . 0 3 . 0 40 . 6 0 0 . 0 50 . 0 I 0 . 0 19 . 5 8 0 . 0 5o . 2 2 0 . 2 57 . 8 0 1 . 3 7l . l 9 0 . 6 48 . 1 2 2 . 8 6

2L.4 47 .5o . 4 4 t . 1 80 . 1 2 0 . l l3 . 2 7 0 . 0 22 . 3 6 0 . 0 20 . 9 5 2 . 9 7o . 2 4 0 . 4 60 . 3 1 3 6 . 7 0

9 6 . 6 1 9 7 . 2 3

247 5r4868 1591240 3r3

35.2 41 .71 3 . 6 1 3 . 96 . 5 4 . 38 . 1 6 . 8t q r t

La ppECe ppDNd ppnsD ppEu ppDTb ppDYb ppuLu pF

*Loss on lgnl t lon

Analyse€ by X-ray f luotescence (X-ray Assay Laborator les, Don U111s, Ontar lo) except Nb, Ba, REE byI N A A ( P h o e n l x L a b o r a t o r y , U n l v e r s l t y o f M l c h l g a n . L o w t o t a l s f o r s o m e a n a l y a e a p r o b a b t y r e f l e c ta b s e n c e o f S r a n d N b f r o n d a t a - c o r r e c t l o n p r o g r a m . S a n p l e n u m b e r s a n d a b b r e v L a t l o n 6 a 9 l n T a b l e 1 .

fractions with most ijolites. As an example, themonticellite-diopside carbonatite at Oka (see above) isyounger than the JoMrC pluton, has diflerent bulk compo-sition (e.g., Fe/Mg, Table 3), different minerals (Table 1),and different mineral compositions (see below). Yet REEdata are consistent with liquid immiscibility between thiscarbonatite and the roMtc ijolite (Fig. a) !

Radiogenic isotopes

Radiogenic isotope ratios in associated silicate and car-bonate rocks provide a necessary condition for an origin asimmiscible liquid fractions: at the time of unmixing, im-miscible fractions must have had the same ratios of parentand daughter isotopes. Identity of isotope ratios and crys-tallization ages merely shows that the rocks may be coge-netic, but non-identity of isotope ratios shows that therocks were not cogenetic and could not have originated asimmiscible fractions. Initial isotope ratios should be inde-pendent of crystal fractionation.

Only Rb/Sr isotope data are available for silicate andcarbonate rocks at Oka, and do indicate that the rocks arecogenetic. Powell et al. (1966) found that ijolite and car-bonatite from Oka (no petrography given) both have initial87sr78usr : 0.7032, and Grunenfelder et al. (1982) reportthe same initial ratio for carbonatites. Preliminarv analvses

of silicate-rich lithologies described here (790K9, 790K10,8089-12) give initial Sr isotope ratios of 0.7032-0.7034.

Although further isotopic work remains to be done,available data are consistent with liquid immiscibility inthe rorurc pluton.

Mineral assemblages

Rocks crystallized from immiscible magmas must haveidentical liquidus minerals, so long as equilibrium wasmaintained (Bowen, 1928, p. 15-16), and mineral assem-blages thus constitute a test for origin as immiscible frac-tions. A weakness of this test is that only vanishingly smallquantities of a mineral need be present, and might not beencountered or recognized in thin section.

Mineral assemblages in the pairs of suspect rock types(the jacupirangites; nepheline okaite and nepheline meli-litite; okaite, melilite carbonatite, melilitite; melilite-diopside carbonatite and diopside melilitite) are essentiallyidentical (Table 1), consistent with origins as immiscibleliquid fractions. A few minerals are exceptions. Diopside inokaite and nepheline okaite occurs only as reaction rimson melilite, and may have formed after the coexisting meli-litites had solidified. Magnetite in the melilite-diopside car-bonatite is commonly rimmed by andradite * perovskite;in the ijolites this reaction has gone to completion leaving

1 108

Fig. 4. Rare-earth element contents of Oka rocks, normalizedto Cl chondrites. Analyses by instrumental neutron activation.Key : melilite-diopside carbonatite, slanted hatchures ; ijolite, verti-cal hatchure; and monticellitFdiopside carbonatite, horizontalhatchure. High Yb values are probably in error.

only rounded intergrowths of andradite + perovskitepseudomorphing magnetite.

Mineral colnpositions

Just as rocks crystallized from equilibrium immisciblefractions must have the same liquidus minerals, so must theliquidus minerals in each rock have identical compositions(Bowen, 1928, p. 15-16). This test has been used as evi-dence for silicate-carbonate immiscibility by LeBas andHandley (1979). Identity of mineral compositions providesa powerful test for origin by liquid immiscibility because:(1) minerals present are easily analyzed (e.g. by micro-probe); (2) chemical analysis of complex minerals (like py-roxenes) tests many independent chemical variables; (3)minerals from crystal cumulates can be used; (4) one is notdependent on flnding rare mineral grains; and (5) otherpetrogenetic processes (particularly crystal fractionation)are unlikely to produce magmas of differing bulk compo-sitions with minerals of identical composition. However, atest by identity of mineral compositions is susceptible tothe effects of failure to maintain magmatic equilibrium, andpost-magmatic chemical reaction, re-equilibration, andalteration.

TREIMAN ,4ND ESSENE: OKA ALKALINE COMPLEX, LIQUID IMMISCIBILITY

To apply the test of mineral composition, one must com-pare only minerals which have retained their magmatic

compositions. Clinopyroxenes, melilite, and perovskite are

reasonable choices, and compositions of all three may be

consistent with liquid immiscibility in the JoMIc pluton.

Both melilite and perovskite show evidence of post-

magmatic re-equilibration, and identity of clinopyroxene

compositions forms the best evidence for liquid immisci-

bility.Clinopyroxene is the best rorutc mineral to use for com-

parison of compositions because it retains magmatic oscil-

latory zoning (Fig. 5) and is known to have very low in-

terdiffusion coeflicients (Brady and McCallister, 1983; Riet-

meijer, 1983). Clinopyroxenes in the lourc rocks include

fassaite in the jacupirangites and diopside in the melilite-

diopside carbonatite, diopside melilitite, and ijolite.

Fassaite compositions from the coarse- and fine-grainedjacupirangites are essentially identical (Table 4; Figs. 6, 7),

supporting the hypothesis that these rocks originated from

immiscible melts. The pyroxenes contain oscillatory growth

zones (Fig. 5a, b), and thus have a range of compositions.Compositions of pyroxenes from the melilite-diopside

carbonatite, diopside melilitite, and ijolite are also essen-tially identical (Table 4; Figs. 6, 7), consistent with anorigin as immiscible liquid fractions. A few ijolite pyrox-

enes are strongly enriched in Na and Fe and were notplotted on the figures; these analyses are from partially

altered rocks with much phlogopite and garnet.

Trace element content of pyroxenes should be as useful

as their major-element composition for testing liquid im-

miscibility. The REE and Sc data of Eby (1973,1975) for

Oka pyroxenes are not, on the surface, consistent with

liquid immiscibility between silicate and carbonate melts

because pyroxenes from the carbonatite are richer in REE

and Sc than those from the ijolites and melteigites. How-ever, all of Eby's analyzed ijolite and melteigite sampleswere strongly altered, containing 15-57% volume of alter-ation products (including analcime, natrolite, cancrinite,nosean and gonnardite). Trace element analyses of pyrox-

enes from less-altered rocks would be most useful.Melilites in the routc rocks are essentially of constant

composition, and are primarily solid solutions between so-damelilite (NaCaAlSirOr) and akermanite components(Table 5; Watkinson, 1972;Yoder,1973). Chemical zoningis seen as a gradual change in birefringence, with zoneboundaries parallel to grain boundaries. Oscillatory zoningis not visible, although it might be present. It is more likelythat melilite compositions reflect sub-solidus (and/or latemagmatic) diffusional zonation. Thus, melilite compo-sitions are consistent with liquid immiscibility in the rorvrlcpluton, but weakly so because they may not be magmatic.

Perovskite may be useful in the test of mineral compo-sitions because it can accept many substituents in readilymeasurable proportions (Table 6). However, perovskite inthe rorurc rocks is either unzoned or indistinctly zoned,suggesting that diffusional re-equilibration may be signifi-cant.

Perovskite composition in nepheline melilitite and neph-

TREIMAN IND ESSENE: OKA ALKALINE COMPLEX, LIQUID IMMISCIBILITY l 109

Fig. 5. Fine-scale optical zoning in pyroxenes from the rorrrrc intrusion.Zonation corresponds to range of chemistry shown in Figures7 and 8. Scale bars 0.1 mm. (a) Coarse-grained jacupirangite (790K9). Oblique illumination, plane light. (b) Fine-grained jacupirangite(79OKl2). Crossed polars. (c) Melilite-diopside carbonatite ('190K23). Plane light. Other minerals are calcite (twin lamellae) and apatite.(d) Dopside melilitite (80B1-2). Plane light.

eline okaite are identical (Table 6), consistent with the hy-pothesis that these rocks originated as immiscible fractions.In both lithologies, perovskite is a liquidus mineral and iszoned (although not sharply). Therefore, similarity in com-positions is unlikely to be the result of post-magmaticequilibration.

Compositions of perovskites in the melilite-diopside car-bonatite, diopside melilitite and ijolite vary widely (Table6), but none of the perovskites is zoned and there is otherevidence of post-magmatic equilibration. In a single ijolitesample (790K25, Table 6) compositions range from 16 to32 wt.'/o Nb2O5 within millimeters. The analyses with theleast Nb are from andradite * perovskite psuedomorphsafter magnetite; because the magnetite contained little Nb,the replacement perovskite must have been in partialchemical communication with the high-Nb, magmatic per-ovskite. Perovskite from the diopside melilitite is differentfrom that in the ijolite, but is partially replaced by sphene.

Perovskites from the melilite carbonatite and melilite-diopside carbonatite vary widely from sample to sample(Table 6). The best match along perovskite analyses is be-tween the carbonatite 790K23 and the high-Nb variety inthe ijolite 790K25. These samples are from adjacent unitsin the field, and the similarity of compositions may implythat these two rocks crystallized from contemporaneousimmiscible melts, or that perovskites in the two samplesequilibrated with each other after magmatic emplacement.

Perovskite in the coarse-grained jacupirangite is compo-sitionally distinct from that in the fine-grained jacupi-

rangite (Table 6). In neither rock is the perovskite zoned,and in both rocks it crystallized quite late. It seems reason-able that the two samples might have been out of chemicalcontact when perovskite grew andfor equilibrated.

Other minerals in the JoMIC pluton (including calcite,magnetite, nepheline and pyrochlore) are not zoned, andthus may have experienced post-magmatic chemical eflects.

1 1 1 0

Table 4. Chemical analyses of pyroxenes from the Oka complex,and normalizations to four cations

5 r M n O 2.5r T l02

J l c r L J i c r L J i c r L J l

Fig. 7. Compositions of pyroxenes from the JoMlc pluton, plot-ted as ranges of minor-element contents (weight percent). Averageis shown as central line within each bar. Key: J, coarse-grainedjacupirangite; j, fine-grained jacupirangite; I, ijolite and diopside-melilitite ; C, melilitediopside carbonatite ; L, monticellite-diopsidecarbonatite.

enes). Because the oldest JoMIc rocks (jacupirangites) show

evidence for coexisting immiscible magmas, one cannot tellif immiscibility occured prior to emplacement of the JoMIcpluton or in the magma chamber presently exposed.

Among the proposed tests of liquid immiscibility, littlefaith may be placed in those that rely on bulk rock compo-sitions, including bulk trace element abundances. Many ofthe silicate-rich lithologies at Oka are crystal cumulatesfrom carbonate-rich magmas, and thus reflect neither thecompositions of silicate-rich melt nor carbonate-rich melt.The carbonatites may also be crystal cumulates (cf. Nesbittand Kelly, 1977;Twyman and Gittins,1982). Use of REEabundances has an added pitfall: most carbonatites world-wide have similar REE abundances (Loubet et al., 1972).Thus, unrelated ijolite or melilitite and carbonatite mayhave the REE distributions of immiscible magmas.

Comparison of mineral compositions is the most usefuland powerful test for origin as immiscible liquid fractions(Bowen, 1928). Comparison of minerals with complex solidsolutions tests many independent chemical parameters, andchemical analysis of these minerals is easy and precise withmodern methods. Mineral compositions consistent withpetrogenesis by liquid immiscibility, i.e., identical compo-sitions in rocks derived from silicate and carbonatemagmas, are extremely unlikely if the rocks are unrelatedor are related by fractional crystallization, partial melting,assimilation, etc. Thus, if mineral compositions and tex-tures are consistent with two rock types originating as im-miscible melt fractions, essentially no other petrogeneticprocesses can be invoked.

To use mineral compositions as tests of immiscibility, tieminerals must retain magmatic compositions. Presence ofsmall-scale oscillatory zoning is taken as evidence of mag-matic compositions (Fig. 5); homogeneous minerals may

TREIMAN ?4ND ESSENE: OKA ALKALINE COMPLEX, LIQUID IMMISCIBILITY

l O x N a 2 O At 2o3

C-Jrc P-J.c ll

79(9 790K12 790X25 ?90U8

NG-H-C Xo-Dt-C

79fl23 6089-7 E0A2{0

sto2rlo2cr203dzo3Fe0horgo

&zoTot. l

39.66 42,964 . 6 5 3 . 5 40 . 0 3 0 , 0 1

L 3 . 4 7 1 1 . r 5

7 , E 2 5 . 0 10 . 1 3 0 . r 09 . 7 r 1 0 . 8 7

24.64 24.36

o . 2 9 0 . 2 71 0 0 . 4 2 9 9 . 3 9

4 2 . 5 1 t 3 . 5 23 . 9 6 3 . 2 L0 . 0 0 0 . o o

1 1 . 9 7 1 0 . 6 8

6 . 2 3 5 . 6 30 . 0 9 0 . r 7

1 0 . 3 5 r 0 . 6 924,42 24.34o . 2 7 0 . 3 3

99.79 98.71

52.93 s2.030 . 2 0 0 . 2 50.00 0.040 . 9 6 r . t 7

3 . 6 9 4 . 4 4

1 . 8 5 2 . 4 31 4 , 8 8 1 3 , 4 02 4 . 6 1 2 t . 9 5

0 . 2 9 0 . 3 99 9 . 6 1 9 9 . 6 4

5 r . 5 4 5 1 . 2 1 5 5 . 7 10 . 3 5 0 . 5 1 0 . 0 30 . 0 3 ! . r ,r . 5 9 1 . 2 8 0 . O l4 , 8 3 5 . l r l . 2 r

1 . 8 3 1 . 5 4 0 , 9 3r 3 . 4 9 1 4 . 2 r 1 7 . 1 525. rE 24.88 25.8ro . 2 8 0 . a 2 0 , 0 4

99,22 r00.t9 r00.E9

b

C!

Fe

h

Irs

CrAst

0 . 0 2 1 0 . 0 2 0 0 , 0 1 9 0 , 0 2 40 , 9 8 6 0 . 9 t 8 0 . 9 7 8 0 . 9 E 2

o . 2 2 4 0 , 1 t 8 0 . r 9 5 0 . r 8 30.004 0.003 0.003 0.0050 . 5 4 0 0 . 6 0 7 0 . 5 7 1 0 . 6 0 00 . r 3 1 0 . 1 0 2 0 . r r r 0 . 0 9 10.oor 0.000 0.000 0,0000 . 5 9 3 0 . L 9 2 0 . 5 2 8 0 . 4 7 4

1 . 4 8 0 r . 6 r 0 1 . 5 9 0 1 . 5 3 8

0 , 0 2 1 0 , 0 2 8 0 . 0 2 0 0 . 0 2 30 . 9 E 2 0 . 9 9 6 t . 0 0 8 0 , 9 7 80 . 1 1 4 0 . r 4 9 0 . 1 5 1 0 . 1 5 9

0 . 0 5 8 0 . 0 7 7 0 . 0 5 8 0 . 0 4 80 . 8 2 0 0 . 7 4 4 0 . 7 5 2 0 . 7 7 80.005 0.007 0.010 0,0090 . 0 0 0 0 . 0 0 r 0 . 0 0 1 ! . . .0 . o 4 2 0 . o 5 0 0 . 0 7 5 0 . 0 5 6

1 . 9 5 7 r . 9 3 8 1 , 9 2 1 1 . 9 5 3

o.0030 , 9 9 80.0360 . 0 2 8o.9220 . m l

0 . 0 0 12 . 0 1 0

-o .a . - lad lca tea Dot sa ly td .

Ana lyses by c lec t toa D lc roProbe, c tys t6 l aPcc t roEetc t . S tandarda lnc ludcd

llDerrls lnd rytrthetlc 8les.e3. Sce lable I for aeople. and rbbtevlltloo8.

DiscussionThe routc pluton of the Oka complex includes an evol-

utionary sequence of rocks derived from silicate-rich andcarbonate-rich magmas. In situ immiscibility between thesemagmas is suggested by, or consistent with, their textures,radiogenic isotope ratios, REE abundances, mineral assem-blages, and mineral compositions (especially clinopyrox-

5 1 0 1 5% F e

Fig 6. Compositions of pyroxenes from the lorvrtc pluton plot-

ted as molar Ca, Fe and Mg normalized to I cation;end members

do not represent wollastonite, enstatite and ferrosilite components.Pyroxenes from the coarse-grained and fine-grained jacupirangites

(open and filled triangles respectively) plot together, and the py-

roxenes from the melilite-diopside carbonatite, diopside-melilititeand ijolite (open and closed circles respectively) plot together. Py-roxenes from the monticellite-diopside carbonatite (open squares)are shown for comparison.

45

t 5

5

E-.t I

tl-H

TREIMAN /ND ESSENE:OKA ALKALINE COMPLEX, LIQUID IMMISCIBILITY

Table 5. Chemical analyses of melilites from the JoMIc intrusion, and normalizations to five cations

l l 1 1

Hel

790K10

0k

790K14

Ne-Uel

790r(13

Ne-0k

790K8 790K25 790K28

Ue-D1-C

790K23 8089-4

s to2 43 .25

[2o3 9 .24

FeO 2 .53

MnO 0 .14

UgO 7. 19

C8O 33.07

M 2 0 4 ' 1 5

KrO 0.00

Tota l 99 .58

44.2r 43.738 . 5 8 8 . 8 52 . t 4 1 . 8 30 , 2 6 0 . 6 77 .44 6 .94

33 .38 32 .853 . 9 9 5 . ? 50 .11 0 .02

ro1 .31 100 .75

43.15 42.568 . 4 0 8 . 7 5r . 6 2 3 . 3 60 ,55 0 .447 .20 6 .55

32.O9 33.695 . 6 3 4 . 8 70 ,00 0 .05

99 .3 r r 00 .30

0.004 0,003o.489 0.4941 .539 r . 6190 .061 0 .128o.o21 0,0250 .481 0 .438o .444 0 .4631 .960 r .909

42.63 41 ,2L 42 .72

9 . 1 6 1 0 . 7 1 9 . L 7

3 . 0 9 2 . 3 4 1 . 2 4

0 . 5 0 0 . 4 2 0 , 5 r

6 . 6 3 5 . 9 3 6 . 3 4

3 3 . 8 1 3 1 . 8 1 3 2 . 6 1

4 . 9 8 5 . 1 3 5 . 5 3

0.06 0 .05 0 .06100.E8 100.60 100. l7

43 .04 46 .70

8 . 2 7 E . 5 3

2 . 8 3 1 . 9 4

r . 2 4 1 . 5 5

5 . 9 9 5 . 8 63 0 . 9 3 3 0 . 9 6

5 . 5 2 5 . 2 4

0.o4 0 .03

9 7 . 7 5 1 0 0 . 8 0

4 3 . 0 4 4 4 . 7 L

8 . 2 7 5 . 6 3

2 . 8 8 2 . 8 L

L . 2 2 r . 0 7

5 . 5 9 7 . 5 0

1 1 . 1 2 3 3 . 5 3

5 . 4 5 3 . 8 0

0.o0 0 .00

99.45 100.05

K

Na

CaFe

hMS

AI

s1

0.000 0 .007 0 .0010 . 3 6 4 0 . 3 4 8 0 . 4 9 4

1 . 6 0 1 1 . 5 0 7 1 . 5 5 60.096 0 ,068 0 .o72

0.005 0 .o lo 0 .025

0 . 4 8 5 0 . 4 9 9 0 . 4 5 7

0.492 0 .454 0 .461

1 . 9 5 6 1 . 9 8 7 1 . 9 3 3

0 . 0 0 3 0 , 0 0 3

0 . 4 3 0 0 . 5 2 6

r . 6 1 3 1 . 5 0 7

o . 1 1 6 0 . 0 7 80 . 0 1 9 0 . 0 1 6

0 . 4 4 0 0 . 3 9 1

0 . 4 E 2 0 . 5 5 8

1 . 8 9 6 1 . 9 1 1

0.003 0 .002

o . 4 7 9 0 . 4 9 1

L,562 r ,52r

0 . t 2 I 0 . 1 0 5

0.019 0 .048

0 . 4 2 3 0 . 4 1 0

0.4E3 0 .467

1 . 9 1 0 L , 9 7 5

0.002 0 .000 0 .000

0.454 0 .479 0 .317

1.484 1 .521 L .642

0 . 0 7 3 0 , 1 1 3 0 . r 0 9

0.059 0 .047 0 .041

0 . 4 4 9 0 . 3 7 7 0 . 5 1 r

0 . 4 4 9 0 . 4 5 9 0 . 3 5 7

2,085 2 ,006 2 .044

Analyaes by electron olcroprobe, cry8tsl apectrmter. Standards lncluded Dlnerala and syothetlcglaases, See Table I for eauplee and abbrevlat loE.

Table 6. Chemical composition of perovskites from the JoMrc intrusion, and normalizatiin{on eqa!!g!L .

C-Jac F-Jac Ne-Ok Ne-UeI790(9 790K12 790K8 790KE

ileDt-C8089-7 80811-3 790K23

IJ790K25

Dt-ltelEoB1-2

slo,Tro;zfrtA1203FeO

UnoCaO

Total

0 .04 0 ,0653 .13 s4 .68o .24 0 .o20 . 5 0 0 . 2 21 . 6 5 r . 2 80 .o9 0 .o5

34.56 36.670 . 2 6 0 . 2 01 .61 0 .402 . 7 0 0 , 6 90 . 7 3 0 . 2 10 . 8 2 1 , 9 30 .01 0 .o00 .00 0 .08

e354 e6;5f

0.00 0.o05t .52 50,720 .09 0 .050 . 4 3 0 . 3 92 .08 2 .O50 .07 0 .07

36 .10 35 .400 . 3 8 0 . 4 31 . 1 5 1 . l 81 . 9 r 2 . L 40 .61 0 ,631 . 2 6 3 . 5 30 .08 0 .120 .00 0 .00

97,69 96.70

0 .04 0 .0527.60 21.O70 . 1 5 0 . 2 60 . l l 0 . 1 35 . 6 7 5 . 7 00 .o9 0 .22

26,46 22.7L4 . r 4 4 . 6 20 .83 0 .812 . r 9 1 . 6 00 .55 0 .48

26.38 35.66L . 2 7 3 . 2 70 ,00 0 .01

97 .47 96 .55

0 . o 1 0 . 4 626.67 26 .44

0 . 1 1 0 . o 0o . 2 8 0 . 2 44 . O r 4 . 4 20 . 1 3 0 . 1 7

28.19 27 .9L4 . 3 0 4 . 0 80 . 5 7 0 . 5 90 . 6 9 0 . 9 20 . 3 0 0 . 4 5

3 2 . 4 4 3 1 . 3 5L . 2 9 0 . 2 00 . 0 0 0 . r 4

e6:r, eT36

0 .0s 0 .12 0 .o228.45 43.68 34.300 .o4 0 .08 0 .10o .24 0 .06 0 .004 . 2 7 L . 5 2 2 . L 90 , 1 4 0 . L 2 0 . 0 5

28.14 30.16 22.2r3 . 5 4 2 . 9 r 5 . 2 10 .84 1 .20 1 .490 . 8 7 t . 2 2 2 . 7 70 .56 0 .60 0 .60

28.74 15. 16 28.330 .29 0 .00 0 .010 .35 0 .00 0 .06

e7:r6- e3'=T e6:=zi

0.001 0.003 0.oo00 .555 0 .E r7 0 .6600.001 0.001 0.0010.007 0.oo2 0.0000.093 0,032 0,0470.o03 0.002 0.(x) lo . 799 0 .E04 0 .6110 . 178 0 .140 0 ,304o .o08 0 .011 0 .015o.008 0.ol l o.o270.005 0,005 0.0050 .337 0 . 171 0 .3280.002 0.o00 0.0000.002 0.000 0.000

s lTtZtAIFeMnCaNaIACeNdNbTaU

0.001 0 .001 0 .000o.975 0 .983 0 .9290.003 0 .000 0 .0010 . 0 1 4 0 . 0 0 6 0 . 0 1 20 . 0 3 4 0 . 0 2 5 0 . 0 4 10 . 0 0 1 0 . 0 0 1 0 . 0 0 10.906 0 .942 0 .9290 . 0 1 3 0 . 0 1 0 0 . 0 t 70 . 0 1 4 0 . o 0 4 0 . 0 1 1o.o24 0 .006 0 ,0170.007 0 ,001 0 .0050 . 0 0 8 0 . 0 2 1 0 . 0 3 50.000 0 ,000 0 .0000.000 0 .000 0 .000

0.000 0 .001 0 .00r 0 .oo0 0 .012o . 9 2 6 0 , 5 4 0 0 . 4 3 1 0 . 5 1 4 0 . 5 1 50.000 0 ,001 0 .003 0 .001 0 .0000 . 0 1 1 0 . 0 0 4 0 . 0 0 4 0 . 0 0 9 0 . 0 0 70.041 0 .124 0 .132 0 .086 0 .0960.001 0 .001 0 .006 0 .003 0 .004o . 9 2 4 0 . 7 4 0 0 . 6 7 3 0 . 7 7 4 0 . 7 7 40 . 0 2 1 0 . 2 r 0 0 . 2 4 7 0 . 2 t 4 0 . 2 0 50.011 0 .008 0 .009 0 .005 0 .0060 , 0 1 9 0 . 0 2 1 0 . 0 1 6 0 . 0 0 6 0 . 0 0 9o.006 0 .005 0 .o04 0 .003 0 .0040 . 0 3 9 0 . 3 3 4 0 . 4 4 5 0 . 3 7 6 0 . 3 7 60.001 0 .009 0 .024 0 .009 0 .0010.000 0 .000 0 .000 0 .000 0 .001

See Table I for rock type8 and abbrevlat lona.Amlt3es by electron nlcroprobe, crystal EpectrGter. Standards lnclude mtural and a)mthet lc

Slaasea aod3 Eetal8 for Nb, Ta, and Zr; synthet lc Blas3 (fr@ A. Cbod€) for Ui and sFthet lcglasses (Drake and l{el l l , 1972) tot the REE.

trt2

have magrnatic compositions but should be used with cau-tion. In rocks of the rorurc pluton, only clinopyroxene andperovskite have small-scale zoning and are useful in testingfor liquid immiscibility. However, few rolrrc rocks havezoned perovskites, and clinopyroxenes in some ijolites areunzoned and apparently re-equilibrated. In other carbon-atites with less post-magmatic equilibration or alteration,other minerals may be useful as indicators of immiscibility.Apatite has been used in hypabyssal rocks (LeBas andHandley, 1979) and pyrochlore and magnetite may retainmagmatic zoning (Petruk and Owens, 1975; McMahonand Haggerty,1979).

Silicate-carbonate liquid immiscibility in the Oka com-plex has been suggested by many investigators (Eby, 1975;Koster van Groos, 1975b; Cullers and Medaris, 1977 ; Yer-woerd, 1978). Their studies have each relied on single linesof evidence and have not considered textures. field rela-tions, or the possibilities of magmatic evolution or multipleintrusions at Oka. The present study has attempted to inte-grate all available field, chemical, and isotopic data on theOka complex, delineate a pluton (rorrarc) which containssilicate-rich and carbonate-rich lithologies, and apply allappropriate tests for liquid immiscibility to the rocks ofthat pluton. All available d,ata are consistent with silicate-carbonate liquid immiscibility in the evolution of the rorvrrcpluton.

It must be emphasized that these conclusions do notimply that carbonatites and ijolites everywhere originatedas immiscible liquid fractions. In most complexes, it is clearthat the carbonatites intrude the alkaline silicate rocks (e.g.,Temple and Grogan, 1965; Erickson and Blade, 1963;LeBas, 1977), precluding derivation as in situ immisciblefractions. Before carbonatite and ijolite are concluded tohave formed as immiscible liquid fractions, a study inte-grating field observations, petrography, trace and isotopechemistry, and mineral chemistry should be undertaken.

Acknowledgments

This work represents a portion of the senior author's doctoraldissertation, Department of Geological Sciences, University of Mi-chigan. It has benefitted from reviews and suggestions from Drs.P. L. Cloke, W. C. Kelly, D. H. Lindsley, T. Y. Tien, and reviewersfor The American Mineralogist. We are especially grateful to Dr.D. P. Gold and colleagues for their work on the Oka complex;without their pioneering efforts, the present study would not havebeen possible. We are grateful to the Phoenix Memorial Labora-tory (University of Michigan) for neutron activation analyses, andto the Electron Microbeam Laboratory (University of Michigan)and the Geochimica et Cosmochimica Laboratoria (Lunar andPlanetary Laboratory, University of Arizona) for maintenance ofelectron microprobe facilities. Strontium isotope analyses arecourtesy of Dr. J. Ruiz and the Laboratory of Isotope Geochemis-try, Department of Geosciences, University of Arizona. Much ofthe writing and revision of this manuscript was done while thesenior author was a research associate with Dr. M. J. Drake,Lunar and Planetary Laboratory, University of Arizona. Wethank Dr. Drake for his encouragement to continue with nonme-teoritic research, and for support through NSF grant EAR83-13425. This paper was supported in part by grant EAR79-23664from the National Science Foundation, grant 2549-79 from the

TREIMAN ,4ND ESSENE: OKA ALKALINE COMPLEX, LIQUID IMMISCIBILITY

Geological Society of America, and grants from the F. Scott

Turner Fund and Rackham Dissertation Fund (both of University

of Michigan). The senior author also is grateful for fellowships

from the Shell Foundation and the H. H. Rackham School of

Graduate Studies (University of Michigan).

References

Bowen, N. L. (1928) The Evolution of the Igneous Rocks. Dover

Publications, New York.Brady, J. B. and McCallister, R. H. (1983) Diffusion data for clin-

opyroxenes from homogenization and self-diffusion experi-

ments. American Mineralogist, 68, 95-105.Cullers, R. N. and Medaris, G., Ir. (1977) Rare earth elements in

carbonatite and cogenetic alkaline rocks: examples from Sea-brook Lake and Callendar Bay, Ontario. Contributions to Min-

eralogy and Petrology, 65, 143-153.Deines, P (1970) The carbon and oxygen isotopic composition of

carbonates from the Oka carbonatite complex, Quebec. Ge-ochimica et Cosmochimica Acta. 34. 1199-1225.

Deines, P. and Gold, D. P (1973) The isotopic composition ofcarbonatite and kimberlite carbonates and their bearing on theisotopic composition of deep-seated carbon. Geochimica etCosmochimica Acta. 37. l7O9-17 33.

Drake, M J. and Weill, D. F. (1972) New rare earth elementstandards for electron microprobe analysis. Chemical Geology,1 0 , 1 7 9 - 1 8 1 .

Eby, G. N. (1973) Scandium geochemistry of the Oka carbonatitecomplex, Oka, Quebec. American Mineralogist, 58, 819-825.

Eby, G. N. (1975) Abundance and distribution of the rare earthelements and yttrium in the rocks and minerals of the Okacarbonatite complex, Quebec. Geochimica et CosmochimicaActa, 39, 597-620.

Erickson, R. L. and Blade, L. V. (1963) Geochemistry and pet-

rology of the alkalic complex at Magnet Cove, Arkansas. U. S.Geological Survey Professional Paper 425.

Freestone, I. C and Hamilton, D. L. (1980) The role of liquidimmiscibility in the genesis of carbonatites-an experimentalstudy. Contributions to Mineralogy and Petrology, 73, lO5-117.

Gold, D. P. (1963) The relationship between the limestones andthe alkaline igneous rocks of Oka and St.-Hilaire, Quebec.Ph.D. Dissertation, McGill University, Montreal.

Gold, D. P. (1972) The Monteregian Hills: ultra-alkaline rocksand the Oka carbonatite complex. 24th International Geologi-cal Congress, Guidebook B-11.

Gold, D. P. and Vall6e, M. (1969) Excursion geologique dans laregion d'Oka. Ministere des richesses naturelles Quebec, pub-lication S-101.

Gold, D. P., Vall6e, M., and Charette, J.-P. (1967) Economic ge-ology and geophysics of the Oka alkaline complex, Quebec.Canadian Institute of Mining and Metalurgy, Bulletin 60, 1131-1r44.

Grunenfelder, M. H., Tilton G. R., Bell, K., and Blenkinsop, J.(1982) Lead-isotope relationships in the Oka carbonatite com-plex. EOS, 63:45, 1134.

Harris, G. and Koster van Groos, A. F. (1981) Liquid immiscibility in the system NaAlSi.Or-NarCO. at pressures of 30 kilo-bars. Geological Society of America, Abstracts with Program,13,443.

Hay, R. L. (1978) Melilitite-carbonatite tuffs in the Laetolil beds ofTanzania. Contributions to Mineralogy and Petrology, 67,357-367.

Koster van Groos, A. F. (1975a) The effect of high CO, pressureon alkalic rocks and its bearing on the formation of alkalic

ultrabasic rocks and the associated carbonatites. AmericanJournal of Science. 275.163-185.

Koster van Groos, A. F. (1975b) The distribution of strontiumbetween coexisting silicate and carbonate liquids at elevatedpressures and temperatures. Geochimica et Cosmochimica Acta,39,27-34.

Koster van Groos, A. F. and Wyllie, P. J. (1963) Experimentaldata bearing on the role of liquid immiscibility in the genesis ofcarbonatites. Nature, 199, 801-802.

Koster van Groos, A. F. and Wyllie, P. J. (1966) Liquid immisci-bility in the system NarO-AlrO.-SiO2{O2 at pressures up to1 kilobar. American Journal ofScience, 264,234-255.

Koster van Groos, A. F. and Wyllie, P. J. (1968) Liquid immisci-bility in the join NaAlSi.Or-Na2CO3-H2O and its bearing onthe genesis of carbonatites. American Journal of Science, 266,932-967.

Koster van Groos, A. F. and Wyllie, P. J. (1973) Liquid immisci-bility in the join CaAlrSirOr-NaAlSirO, NarCO.-HrO.American Journal of Science, 273,465-4.87 .

LeBas, M. J. (19'77) Carbonatite-Nephelinite Volcanism. JohnWiley & Sons. New York.

LeBas, M. J. and Handley, C. D. (1979) Variations in apatitecomposition in ijolitic and carbonatitic igneous rocks. Nature,279.54-56.

Loubet, M., Bernat, M., Javoy, M., and Allegre, J. C. (1972) Rareearth contents in carbonatites. Earth and Planetary Science Let-ters, 14,226-232.

Mariano, A. N. and Roeder, P. L. (1983) Kerimasi: a neglectedcarbonatite volcano. Journal of Geology,9l, 449455.

McMahon, B. M. and Haggerty S. E. (1979) The Oka carbonatitecomplex: magnetite compositions and the related role of tita-nium in pyrochlore. In F. R. Boyd and H. O. A. Meyer, Eds.,Kimberlites, Diatremes and Diamonds: Their Geology, Pet-rology, and Geochemistry, p. 382-399. American GeophysicalUnion., Washington, D. C.

Middlemost, E. A. K. (1974) Petrogenetic model for the origin ofcarbonatites. Lithos, 7, 27 5-278.

Nesbitt, B. E. and Kelly, W. C. (1977) Magmatic and hy-drothermal inclusions in carbonatite of the Magnet Cove com-plex, Arkansas. Contributions to Mineralogy and Petrology, 63,27t-294.

Peacor, D. R. (1976) Refinement of the crystal structure of a py-roxene of formula MrMil(Sir sAlo.s)O6. American Mineralogist,52,3t-41.

Petruk, W. and Owens, D. R. (1975) Electron microprobe analysesfor pyrochlores from Oka, Quebec. Canadian Mineralogist, 13,282-285.

Powell, J. L., Hurley, P. M. and Fairburn, H. W. (1966) Thestrontium isotopic composition and origin of carbonatites. InO. F. Tuttle and J. Gittins, Eds., Carbonatites, p. 365-378. J.Wiley and Sons, New York.

Rankin, A. H. and LeBas, M. J. (1974) Liquid immiscibility be-tween silicate and carbonate melts in naturally occurring ijolitemagma. Nature, 250, 206-209.

Rietmeijer, F. J. M. (1983) Inter-diffusion coelficients parallel tothe c-axis in iron-rich clinopyroxenes calculated from micro-structures. Contributions to Mineralogy and Petrology, 83,169-176.

Streckeisen, A. (1980) Classification and nomenclature of volcanicrocks, lamprophyres, carbonatites and melilitic rocks. IUGSSubcommission on the Systematics of Igneous Rocks, Rec-

I 1 1 3

ommendations and Suggestions. Geologische Rundschau, 69,194-207

Temple, A. K. and Grogan, R. M. (1965) Carbonatite and related

alkalic rocks at Powderhorn, Colorado. Economic Geology, 60,672-692.

Treiman, A. H and Essene, E. J. (1984) A periclase-dolomite-

calcite carbonatite from the Oka complex, Quebec, and its cal-culated volatile composition. Contributions to Mineralogy andPetrology, 85,149-157

Treiman, A. H. and Schedl, A (1983) Properties of carbonatitemagma and processes in carbonatite magma chambers. Journalof Geology, 91,437447.

Twyman, J. D. and Gittins, J. (1982) Differentiation of carbonatitemagma-evidence for alkali-enrichment and thenatrocarbonatite-sovite link. Geological Society of America,Abstracts with Program, 14,635.

Verwoerd, W. J. (1978) Liquid immiscibility and the carbenatite-ijolite relationship: preliminary data on the join

NaFe*3Si2O6-CaCO. and related compositions. Carnegie In-

stitution (Washington, D. C.) Yearbook,'17, 767-774.Watkinson, D. H. (1972) Electron microprobe analysis of melilite

and garnet from the Oka complex, Quebec. Canadian Mineral-

ogist, 11, 457463.Wendlandt, R. F. and Harrison, W. J. (1979) Rare earth partition-

ing between immiscibile carbonate and silicate liquids and CO,vapor: results and implications for the formation of light rare

earth enriched rocks. Contributions to Mineralogy and Pet-

rology,69, 409419.Williams, H., Turner, F. J. and Gilbert, C. M. (1954) Petrography,

Freeman and Co, San Francisco.Yoder, H. S., Jr. (1973) Melilite stability and paragenesis. Fortsch-

ritte der Mineralogie, 50, 140-173.

Manuscript receioed, December 12, 1984;acceptedfor publication, July 15, 1985.

Appendix

Sample locations

S e e F l t u r e 1 e n d T a b l e I f o r S e r e r e l l o c a t l o o 6 e n d a b b r e v l . t l o o s l o r t o c k

@es. A l l f l8ure€ l l s ted be ld l te ln Go ld (1972) .

79OKg - Ne{L - f ,uaereau HI I I , S top 2c , F18. 5 .

79OK9 - F-Jac - nusereeu n t l l , 20 E N o f S top 2b ' F18. 5 .

79oKl0 - te l - Busereau A11 l , S top 2b , F tg . 5 .

79ox l2 - c -Jdc - Huseteau Ht l I , 35 o E o f S top 2b , F t8 . 5 .

79OK13 - Ne-XeI - Huaer€eu Ht I l , 25 a E o f S top 2b , F t8 . 5 .

79oKl4 - ok - Husereau Et I l , s top 2b , F ta 5 .

79OU3 - i le -Dr -C, Bond Zo@, Tr€och 9 , u$ ternrGt carbonat l te , F lg . 8 ,

79OK25 - I l - Bo t rd zone, Trench 9 , ces le r i los t t jou te ' F t f , 8 .

790K28 - I l - Bood Zone, Trench 1 , res te raoodt t lo l t te ' F t8 .8 .

80A2-8 - Mo-Dt -C - Open P l t A2 , W se l l o f ad l t , upper hadqe leve t , F18. 7 .

8 0 A 2 - 4 O - U o - D l - C - O p e n P l l A 2 , I o s e r h s u l a S e l e v e l , N 2 0 E f r o o p l t c e u t e r l

F t B . 6 .

80BI -2 - D t -ue l - Bond Zone, Trench 1 , 20 r S o f w eDd o f t re t rch ' F l8 , L

8089-4 - t re -Dt -C - Bod Zmer Trench 9 , 7 D f roE SS edd o l t renc t , F l8 . 8 .

8089-7 - Ue-Dt -C - Bond Zone, Trench 9 , 12 r f rd S f l end o f t re t rch , F t8 . 8 .

8089-12 - I l - Bond zore , Treoch 9 ,20 o f roE S! end o f tEerch , F lg . 8 .

8081I -3 - ue-Dt -C - Bond zone, Trsch 11 , 5 ! f rm SX end o f t rench, F1g. 8 .

TREIMAN,4ND ESSENE: OKA ALKALINE COMPLEX, LIQUID IMMISCIBILITY