American Mineralogist, Volume 59, pages 1190-1197, 1974

Clinopyroxenes from the Keweenawan Lavas of Minnesota

T.qolsHI KoNna

Department of Geology, Yamagata Uniuersity,Yamagata 990, Iapan

eNo JonN C. GnnrN

Department of Geology, Uniuersity ol Minnesota, Duluth,Duluth. Minnesota 5 58 I 2

Abstract

The Keweenawan (I-ate Precambrian) lava flows in northern Minnesota include the volcanicseries: tholeiitic olivine basalt, basaltic andesite, quartz latite, and rhyolite, which is assumedto be derived from a tholeiitic parental magma. These lavas contain calcic clinopyroxene whichranges from CarnMgnuFe,, to CaoMgr"Fenu. The crystallization trend of these clinopyroxenes ischaracterized by marked iron enrichment and is similar to that of the associated Beaver BayGabbro Complex and of the Skaergaard Intrusion. The mineral assemblage of olivine f fer-roaugite + quartz and the crystallization trend of the ferroaugites are characteristic of equi-librium crystallization of pyroxene during the late stage of difterentiation of tholeiitic magma.

fntuoductionThe compositional trends and the phase relations

of pyroxenes in intermediate and felsic volcanicrocks of tholeiitic derivation have been rcported bymany investigators (e.g., Carmichael, 1960; Kuno,1969; Emeleus, Dunham, and Thompson, l97l).One such rather complete series of lavas of generallytholeiitic affinities is the North Shore Volcanic Groupof Late Precambrian age, about 1130 million yearsold, which underlies (Fig. 1) most of the shore ofLake Superior in northeastern Minnesota, U.S.A.(Green, 1972). This series, which includes approxi-mately 7500 meters (25,000 feet) of lavas, containssome types which are more potassic than is typicalin pure tholeiitic suites, and it ranges in compositionfrom olivine tholeiites through q\artz tholeiites,basaltic andesites, trachybasalts, andesites, andquartz latites to rhyolites. The olivine tholeiites arethe most abundant single rock type. The flows areintruded by the immense Duluth Gabbro Complexand by locally abundant shallow intrusive rocks,principally olivine diabase. The Beaver Bay Com-plex, described by Grout and Schwartz (1939) andGreen (1972) and in part by Gehman (1957) isone such group of intrusions. Muir (1954) andKonda (1970) have studied the clinopyroxenes ofthe Beaver Bay Complex. Except for lavas and

gabbro far to the northeast and southwest of thoseconsidered here, all of the rocks (including theBeaver Bay Complex) which intrude the North ShoreVolcanic Group are within -+15 m.y. of the sameage as the lavas (Silver and Green, 1972), show thesame magnetic pole position (Beck and Lindsley,1969; Books, 1972; Books and Green, 1.972) andare assumed to be cogenetic with the lavas.

The basalts of the North Shore Volcanic Groupare nearly alL non-porphyritic, the pyroxenes occur-ring generally either as poikilitic crystals in ophiticor subophitic intergrowth with plagioclase or as smallequant crystals in an intergranular texture. The inter-mediate and felsic lavas, however, are nearly allporphyritic, with small plagioclase, clinopyroxene,and commonly also magnetite and olivine pheno-crysts. The olivine has in all cases been destroyed bysecondary alteration, but the clinopyroxene is freshin many rocks.

To understand the complete trend of compositionsof the clinopyroxenes during fractionation of tholei-itic magma, partial analyses of many crystals fromthis volcanic series were obtained using an electronmicroprobe (Green), and the clinopyroxenes fromfour large samples were separated and analyzed byconventional wet chemical methods (Konda).

The samples studied for this report are from lava

I 190

flows of the uppermost 2000 meters of the NorthShore Volcanic Group, plus a dike that crosscuts theflows roughly 2600 meters from the top of the ex-posed section. The sample localities are shown onFigure 1. This paper presents the results of theseanalyses and discusses the petrologic significance ofthe pyroxene crystallization trend thus outlined.

Petrography

Chemical analyses of most of the rocks sampledare given in Green (1972); they are plotted on anAFM diagram in Figure 2. Brief petrographic de-scriptions of the rocks from which pyroxenes werestudied follow.

(a) Olivine basalt (T-56) : lakeshore at E. edgesec. 12, T.59N., R.4W. Slightly porphyritic, fine- tomedium-grained basalt flow between 7-8 -m thick(Fig. 3,A.). Major minerals: clots of small, euhedralbytownite phenocrysts zoned Ans5-Anool poikiliticaugite up to 2 mm across enclosing small labradorite

11,91

laths; subhedral olivine now completely altered toserpentine or chlorite; and both ilmenite and Ti-magnetite. This basalt is the most "primitive"(highest Mg/Fe ratio, lowest K2O) of the olivinetholeiites analyzed in the North Shore VolcanicGroup, and, according to calculations by Phinney(I970), it may be close to the parental magma com-position for the Duluth Gabbro Complex as well asfor the North Shore Volcanic Group.

(b) Olivine basalt (T-22); on old road, NW 1/4sec. 2, T.58N., R.5W. Medium-grained, subophitic;the upper-central part of a very thick flow. About3 percent small olivine crystals, completely altered;augite up to 3 mm diam., some as skeletal inter-growths with plagioclase; both ilm. and Ti-mag;plagioclase laths zoned, An65-An2p. Trace of inter-stitial granophyre. Saponite and zeolites fill dikty-taxitic cavities.

(c) Ophitic olivine basalt (DY-3): lakeshore atnorth edge of Sec. 15, T.60N., R.2W. The lower

CLINOPYROXENES FROM KEWEENAWAN LAVAS

Flc. 1. Generalized geologic map of Keweenawan rocks in the western Lake Superior area showing location of samples studied.Check pattern, intrusive rocks; vertical ruling, Lower Keweenawan rocks (magnetically reversed); stippled, Middle Keweenawanlavas (magnetically normal). After Green (1972, Fi'e. V-3).

t192 T, KONDA, AND

Fro. 2. AFM diagram for rocks from which pyroxeneanalyses aro given in this paper. Point labelled GH-7a isderived from rock analysis S & G-5-9 (from the same flowbut a diferent locality), and point DY-3 is derived frornrock analysis DY-6b (from a similar flow slightly lower inthe sestion than DY-3), Rock analyses from Green (1912).

middle part of a very thick (200 m) flow offine- to medium-grained basalt. Tabular plagioclasepredominates, with fresh ophitic augite, olivinepseudomorphs, magnetite and ilmenite (Fig. 3B).Interstitial cavities contain zeolites and calcite.

(d) Porphyritic basalt (GH-7A): lakeshore 500m west of power plant at Grand Marais; SE 1/4 SWl/4 Sec. 20, T.61N., R.lE. The chilled base of athick (110 m) basaltic to andesitic sheet that isprobably a flow but may be a sill. Abundant 1-2 mmphenocrysts of zoned plagioclase, abott 2 percent offresh prismatic augite crystals (0.5-1.5 mm) andfewer small altered olivine phenocrysts set in a sub-spherulitic, fine-grained groundmass of plagioclasetablets, clinopyroxene prisms, and magnetite blocks,with interstitial alkali feldspar in the coarser areas.Original small miarolitic cavities filled with quartzand an oxidized chloritic mineral.

(e) Porphyritic basaltic andesite (F-7a): lake-shore, NE l/4 Sec.21, T.57N, R.6W. The basal,chilled portion of a very thick (at least 100 rr)flow. Red-brown, fine-grained potassic basalticandesite with 10 percent small phenocrysts of plagi-oclase (Anss), augite, rare pigeonite, and olivine(Fig. 3C). The olivine phenocrysts (completelyaltered to iddingsite) are prismatic and some are inparallel intergrowth with augite. Augite phenocrystseuhedral, pale green, nonpleochroic. A few show

I. C. GREEN

maryinal zoning by changes in extinction angle andbirefringence. Some augite shows very thin exsolu-tion lamellae (pigeonite?). Groundmass intergranu-lar in texture, with plagioclase, alkali-feldspar,augite, pigeonite, opaques, and greenish clay min-erals. The pigeonite tends to be oxidized along rimsand fractures.

(f ) Basaltic andesite (MI-4): lakeshore, SE 1'/4Sec. 22, T.62N., R.3E. Fine-grained red-brownaphyric andesite or trachy-andesite flow showing thinoxidation banding. Intergranular to intersertal tex-ture with small plagioclase laths (An 53-45), ffi?8-netite octahedra, and 8 percent blocky, fresh ferro-augites in a groundmass of oxidized glass, magnetite,alkali-feldspar, quartz, and amphibole (Fig. 3D).Cavities are filled with quartz, chlorite, clay, andzeolites. Flow thickness unknown.

(g) Porphyritic andesite or trachyandesite (H-4):Highway 61, center line Sec. 12,T.62N, R.4E. Veryfine-grained, dark brown dike, which transects thelavas about 2600 m below the top of the exposedsection. About 3 percent of small phenocrysts ofsubhedral oligoclase (Anzu), ferroaugite, and mag-netite, in a groundmass of plagioclase, alkali-feld-spar, ferroaugite, magnetite, anhedral hornblende,possible devitrified brown glass, and alteration prod-ucts.

(h) Porphyritic intermediate qaartz latite (MI-2): Highway 61 at W. edge Sec. 6, T.62N., R.3E.Fifty m-thick lava flow rather similar to rocks termed"icelandites" by Carmichael (1964). Reddish brown,fine-grained; about 7 percent of small phenocrystsof stout tabular, unzoned andesine (Atuo), ferro-augite, altered olivine, and magrretite. The ferro-augite phenocrysts are fresh, about 1 mm across;some are aggregated with the other phenocrysts(Fig. 3E). The groundmass minerals are plagioclase,alkali-feldspar, interstitial quartz, opaques, and greenalteration products, probably chlorite. Althoughthere may have been clinopyroxene in the ground-mass, it is not now recognizable.

(i) Porphyritic rhyolite (GM-10): lakeshore,center line Sec. 14, T.61N, R.1E. Basal spheruliticzone of a rhyolite lava flow that is as much as 155 mthick. Orange-red, aphanitic, about 3 percent smallblocky phenocrysts of plagioclase (An35-26) and rareqvartz, ferroaugite, magnetite and altered olivine.The ferroaugite occurs as small, fresh, subhedralphenocrysts (0.1 x 0.8 mm) (Fig. 3F). Ground-mass microcrystalline feldspar and quartz with mi-nute opaque grains.

i l -z .xt-4" Fal ' 1-16

CLINOPYROXENES FROM

(k) Porphyritic intermediate quaftz latite (F-21):NE 1/4 Sec. 28, T.57N., R.7W. Red-brown, veryfine-grained "andesite" assumed to be a lava al-though contacts not exposed. Contains 10 percent

KEWEENAWAN LAI/AS rt93

small, blocky albite phenocrysts, about 3 percentspongy augite prisms and about 2 percent each ofmagnetite and oxidized olivine. Groundmass domi-nated by K-feldspar, with lesser amounts of plagi-

Flc. 3. Photomicrographs of clinopyroxenes in Keweenawan Lavas. A, ophitic augite enclosing plagioclase and pseudomorphs

of olivine in olivine tholeiite T-56; B, ophitic augite enclosing plagioclase with olivine pseudomorph in olivine tholeiite DY-6b

(flow close to and similar to pyroxene sample DY-3);C,augite and plagioclase phenocrysts, olivine pseudomorph in basaltic andesite

F-|a;D,equant augite crystals with magnetite and plagioclase in andesite MI-4; E, prismatic ferroaugite, plagioclase, and olivine-

pseudomorph phenocrysts with abundant apatite in intermediate quartz latite MI-2; F, prismatic ferroaugite, plagioclase, and small

magnetite phenocrysts in rhyolite (felsic quartz latite) GM-10.

l194 T, KONDA. AND I, C. GREEN

oclase, quartz, yellow-green secondary minerals,apatite, magnetite, and ilmenite.

Analytical Results

The clinopyroxene phenocrysts of three largesamples and subophitic augite of a fourth wereseparated, using the isodynamic separator andClerici's solution, and analyzed by conventional wetchemical methods (Konda). Groundmass pyroxenesin the porphyritic rocks either wgre not present(MI-z) or were avoided by size classification duringseparation. In addition, about forty partial analyses(for Mg, Fe, Ca) of pyroxenes from six other lavasdescribed above were obtained by Green using theMac electron microprobe at the Department ofGeology and Geophysics at the University of Minne-sota, Minneapolis, with the assistance of Paul W.Weiblen, who also analyzed the pyroxene from therhyolite GM-10. The standards for microprobeanalyses were pyroxenes analyzed by wet chemicalmethods. The results of the wet chemical analysesare shown in Table 1, and the Mg/Fe/Ca relations ofall analyzed pyroxenes are shown in Figure 4.

A weak zoning in the ophitic pyroxenes of theolivine tholeiites T-56 and DY-3 is suggested, trend-ing away from diopsidic and toward subcalcic com-

TesI,n 1. The Chemical Compositions and Atornic Ratiosof Calcic Clinopyroxenes from Keweenawan Flows, North

Shore Volcanic Group, Minnesota

T. Konda, Analyst (wet chem.)

s i 0 2 5 0 . 7 4

Tto2 I . 16

|JzO3 2,98

F e 2 0 3 2 . O ,

Feo 1,3t1

l ,{no 0,16

Mso L5.53

cao t9.66

N a 2 0 O . 5 1

KzO 0.09

H 2 o ( + ) o . L 5

H 2 o ( - ) o . l o

P z O , 0 . o o

5 1 . 0 0 ! 5 . 5 1 + ! 8 . 1 1

0 . 9 3 0 . 8 2 0 . 8 0

3 . r B 2 . 1 8 r . 3 1

z . j 9 L . j 2 r . L o

ro-6j t9.zz 22.\o

o . 2 9 0 . 3 7 0 . 1 + 7

1 3 . 0 8 1 0 . 9 1 6 . 2 6

1 6 . 5 8 1 5 . 0 9 r 8 . 3 7

0 , 6 1 1 0 . 5 0 a , 2 5

0 . 1 ? 0 . 2 8 o . 2 6

0 . 8 6 o . 8 ! 0 . 5 3

0.3! ----- -----

o . 0 r 0 . 2 0 0 . 4 0

s i 1 . 8 8 0 r . 9 2 3

A l r v 0 . 1 2 0 O . 0 ? 7

T1

lvl o,oto o. 06lr

T t o . o 3 2 0 . a 2 6

F e + 3 o . 0 5 ? o . o ? 3

Fu+2 o.22' l 0.336

u n 0 . 0 0 3 0 , 0 0 9

M s 0 . 8 5 8 0 . ? 3 5

c a 0 . 7 8 0 0 . 5 ' f o

N a 0 . 0 ! 0 0 . 0 4 7

K o . o o l r 0 . 0 0 5

Frc. 4. Atomic ratios of analyzed calcic clinopyroxeneson pyroxene quadrilateral. Points shown by circle arounddot are wet-chemically analyzed (Table l); others are indi-vidual point analyses by electron microprobe, except GM-10which is the average of 4 points on each of two grains. Allare phenocrysts except T-56 and DY-3 (poikilitic augite),T-22 (subophitic augite), MI-4 (equant groundmass py-roxene ).

positions with a slight increase in iron content. The

compositional variation within the pyroxenes fromsaturated, intergranular basalts and basaltic andesites(e.g. Ml4) is mostly in Mg/Fe ratio, nearly parallelto the Di-Hd join. The ferroaugites of the intermedi-ate quafizlatite,Ml-z, show enrichment in both Caand Fe at the margin, and the qnartz latite F-21shows principally Mg/Fe zoning. Microprobe studyof the clinopyroxene from rhyolite GM-10 showedno zoning present.

Discussion

Figure 5 shows the correlation between the atomicFe/(Fe*Mg) ratios of the clinopyroxenes and theratio of total FeO to total FeO*MgO in the ana-lyzed host lavas. Figure 6 shows the pyroxene Fe/(Fe+Mg) ratio plotted against SiOz content of thehost rock. Because the percentage of pyroxene phe-nocrysts in the porphyritic lavas is very small, thewhole-rock Fe/Mg ratios do not depend greatlyon the pyroxene phenocryst compositions. Conse-quently, the general correlations shown in Figures 5and 6 can be interpreted as indicating that thepyroxene phenocryst compositions were controlledby the compositions of the lavas in which we nowfind them. Thus, the phenocrysts were probablyindigenous, and not included xenocrysts.

The clinopyroxenes in samples F-7a, GH-7a, andH-4 show slightly low Fel(Fe+Mg) ratios (Fig. 5)as compared to the general trend. Under the micro-scope, these rocks contain small amounts of pheno-crystic magnetite, suggesting that the magnetite mayhave precipitated during or before the crystallizationof clinopyroxene and prevented further iron enrich-

r .852 ), .9]6

o , r o r 0 , 0 6 1

o . o 2 l + o . 0 2 3

0 . 0 2 3 - - - - -

0 . 0 2 2 0 . 0 4 1

o . 5 3 0 0 . 1 \ 5

o , o r 2 o . o ! 1

O.6t!2 O.3- lr

0 . 6 8 5 o . 7 8 3

o , o l + 5 0 . 0 1 9

0 . 0 1 3 0 . o r 3

Total L00.83 1O0,32 99,57 Loo.59P - - - - - - - - - - 0 , 0 0 5 0 . 0 1 3

z 2 . o o 0 2 . D 0 0 2 . 0 0 0 2 . 0 0 0

w r t 2 . 0 r r 1 . 9 6 T 1 . 9 q 1 r . 9 q 9

At@ic tr

C a ! 0 . 5

Ma LI+.6

F e * f ! , 9

36. ' t 3 ! .1 { !o .1

l0 . l 32 .2 19 .0

22.9 33 .1 !0 .9

2

Ferroeugite phenocrysts fr@ 6wite edestte (H-l+)

FeFoau8ite phenocitrsts fron lntemealiate qustz Iatlte (Mf-2)

t . ^ o o ^ r t or-22- .sf*i ""

-l'fri_,H - 4

x T-56 + Ml -z. DY-3 o F-21o GH-7o r GM- IOA M I - 4

3

LPo

05 ta^?6r.o 7 ttotar Lo - Mso), host rock

Frc. 5. Atomic Fel(Fe*Mg) of clinopyroxenes plottedagainst total FeO/(total FeOfMgO) (wt. percent) of hostrocks. Augite DY-3 is compared with analyzed DY-6b, asimilar flow slightly lower in the section. Augite GH-7a iscompared with analyzed rock S & G-59 from another Io'cality in the same flow. Rock analyses from Green (1972).

ment in the pyroxene. Although Eggler and Burn-ham (1973) find that magnetite should not o€cur atthe liquidus in andesitic magmas, it evidently didstart to precipitate in these lavas at some time duringthe first 10 to 20 percent of crystallization. flowever,there is less magnetite than pyroxene present asphenocrysts in these rocks, and there is no evidenceto indicate that it siarted to crystallize before theclinopyroxene. Furthermore, these samples have nomore magnetite as phenocrysts than the samplesfalling on the "general trend." The relatively lowFe/(Fe+Mg) ratios of these three clinopyroxenescompared to their host rocks may be a reflection ofthe fact that their small content of pyroxene pheno-crysts represents only the first, most magnesian partof their normative pyroxene content; whereas inMI-4, an aphyric andesite, the analyzed pyroxene isin the groundmass and should more nearly approachthe normative values. In the three more felsic sam-ples (F-21, MI-2, GM-10) the analyzed phenocrystsprobably contain most of the potential pyroxene inthe rock.

Also, these three intermediate samples GH-7a,F-7a, and H-4 show a high content of magnetite inthe groundmass as compared to the other rocks,which would increase the total FeO/(FeO*MgO)total ratio ofthe rock analysis.

The compositional zoning of these clinopyroxenes(Fig. a) is similar to the pyroxene trend directionobserved in many layered intrusions (e.B'., Skaer-gaard and Beaver Bay), and although these rocksare obviously lavas, the basaltic groundmass py-roxenes analyzed do not show any "quench trend"referred to by Muir and Tilley (1964) or Smith

CLINOPYROXENES FROM KEWEENAWAN LAVAS 1 195

and Lindsley (197I). We interptet this lack of a"quench trend" to be the result of relatively slowcrystallization of these thick, flood basalts. Slowcrystallization is also indicated by their relativelycoarse grain size and their flat surfaces and wideextent. The observed trend is instead inferred to bethe "intratelluric equilibrium trend" referred to byMuir and Tilley.

Clinopyroxenes from the consanguineous BeaverBay Complex have been studied by Muir (1954)and more recently by Konda (1970) and Stevenson(1,973,1974). Because the pyroxene phenocrysts inthe lavas could be expected to have formed duringhypabyssal differentiation under conditions similarto those obtaining during crystallization of theBeaver Bay intrusive rocks, it is interesting to com-pare the clinopyroxenes of the lavas with those ofthe Beaver Bay Complex. Although the lava samplesdescribed in this paper came from a much wider areathan the Beaver Bay Complex (Fig. 1), the controlson and character of differentiation of both may wellhave been similar. Some of the sampled lavas mayactually have been surface extrusions of the samemagmas that produced parts of the Beaver Bay Com-plex, although it would be difficult to demonstratesuch a relationship.

In chemical characteristics, the clinopyroxenesfrom the lavas of the North Shore Volcanic Groupare slightly richer in Al2O3 and Na2O, but haveabout the same amount of MnO and TiOp as thoseof similar Fe/Mg ratio from the Beaver Bay Com-plex (Konda, l97O). The wet chemical analyses ofTable 1 show that the CaO contents of the clino-pyroxenes from most of the lava flows are slightly

GM-1O

ot^t l * r - r ,T-56

45 50 55 60wt " / . S iO2, hos t

Frc. 6. Atomic Fel(Fe*Mg) of calcic clinopyroxenesplotted against weight percent SiO, in host rocks.

I

" , - , {1.-or l

I x H - 4

* M r - 4Gu-zo I xr-fo

o

,9

\ o.o

65 70 75rock

1196

lower than those from the Beaver Bay intrusion(Fig. 7). This might indicate slightly higher tem-peratures of crystallization for the phenocrysts ofthe lavas than for the intrusive rocks. Ilowever,microprobe analyses of clinopyroxenes from DY-3,F-21, and MI-4 (Fig. 4) show essentially the sameor lower CaO content compared to those from theBeaver Bay Complex, and the two populationswould be difficult to distinguish.

The crystallization trends of the clinopyroxenes inthe Keweenawan flows and the cumulus augites inthe Beaver Bay Complex are similar, as shown inFigure 7. However, in the Beaver Bay Complex,Konda (1970) shows that the calcic clinopyroxenesare locally rimmed by pigeonite-ferropigeonite,which is inferred to have crystallized by the reactionbetween cumulus Ca-rich clinopyroxenes and theinterstitial liquid. On the other hand, the calcicclinopyroxenes from the related flows (this paper)do not have such a mantle of Ca-poor pyroxene,probably because such a reaction point had not beenreached when the lavas were extruded and quenched.As mentioned above, the potential Ca-poor (and Fe-rich) pyroxene components probably remain in theunanalyzed groundmass ferromagnesian minerals.

Figure 7 further shows for comparison the trendof clinopyroxene compositions for a variety of Ter-tiary and Quaternary volcanic rocks (Carmichael,1967); it is essentially identical to that of the ana-lyzed Keweenawan pyroxenes. From a study of theassociated oxide minerals, Carmichael (1967) inferspyroxene crystallization temperatures ranging from970" to 915"C for similar intermediate and felsicporphyritic lavas.

FIo. 7. Calcic pyroxene trends compared. Crosses, wet-chemically analyzed clinopyroxenes frorn Keweenawanlavas; circles, average compositions of clinopyroxenes fromKeweenawan lavas analyzed by microprobe; dots, analyzedpyroxenes from Beaver Bay complex (Konda, 1970); solidline, trend of clinopyroxenes from various Tertiary andQuaternary volcanic rocks (Carmich ael, 19 67 ) ; dashed line,trend of clinopyroxenes from Shiant Isles sill (Gibb, 1973).

T. KONDA, AND J. C. GREEN

The assemblage of mafic minerals present isclosely related to the crystallization trend of py-roxenes at the late stages of fractionation, as sug-gested from experimental work (Lindsley andMunoz, 1969) andfrom natural rocks (Kuno, 1969;Emeleus et al, l97l). The most common phenocrystassemblages of the mafic minerals in the analyzedsamples are as follows: augite * olivine in basalt(early stage; no phenocrysts); augite * olivinephenocrysts in basaltic andesite (GH-7a); augite *pigeonite * olivine * magnetite in basaltic andesite(F-7a); ferroaugite * olivine * magnetite in inter-mediate quaftz latite (MI-2); and ferroaugite, oli-vine, and magnetite in rhyolite (GM-10) at a latestage. Ferroaugite phenocrysts from MI-2 and GM-10 have Fe/(Fe*Mg) ratios of 0.63-0.73, whichare similar to those of ferroaugite from porphyriticfelsite and granophyre in Scotland (Emeleus er a/,l97I ,no.8 of group 3 and nos. 5a,5b of group 4,Table 2). Ilowever, the mineral assemblages aredifferent: ferroaugite * olivine in the Keweenawanlavas, and ferroaugite + ferro-pigeonite in theScottish felsic rocks.

Based on the experimental work of Lindsley andMunoz (1969, p. 316), the reaction Ferrosil i te =Fayalite + SiO, melt is such that an increase in theactivity of SiOz leads to an increase in the activity ofFs under equilibrium conditions. Persistence of oli-vine throughout the crystallization sequence and ahigher content of alkalies in these Keweenawan lavasthan in the typical tholeiitic suites and Scottish felsicrocks indicate the magma compositions to be slightlysub-alkaline and suggest a slightly low SiOz activity.However, the Keweenawan trends clearly differ fromthat of the mildly alkaline Shiant Isles sill (Gibb,1973), which is more calcic and lacks the centraldip in Ca content (Fig. 7). Although SiO2 activityis close to unity near the solidus for each Keween-awan lava, it would appear that SiOz activity did notincrease fast enough below the liquidus to stabilizeCa-poor pyroxene. On the other hand, however,Mg/Fe distribution values between coexisting ferro-pigeonite (5c, group 4) and ferroaugite (5a and5b, group 4) in the granophyres of Scotland aresomewhat variable, possibly suggesting metastablecrystallization of ferropigeonite in the Scottish grano-phyre. Also, from the descriptions of the ferro-pigeonite-bearing rocks (Emeleus et aI, 1971, p.941-944), the crystals are rather variable in com-position within individual samples and some showcomplex zoning. This too would suggest that the

CLINOPYROXENES FROM KEWEENAWAN LAVAS l t 97

ferropigeonites described result from disequilibriumcrystallization, perhaps related to an extension ofMuir and Tilley's "quench trend" (1964). The pres-ent data agree with experiments by Smith (1971) oncompositions along the join Fss5En15-Wo. These in-dicate that the assemblage calcic clinopyroxene-fayalitic olivine-silica is stable at low crustal pres-sures and temperatures of 1000' to 800oC.

The stable mineral assemblage during late stagesof fractionation appears to be ferroaugite * fayaliticolivine + qnartz, and the crystallization trend shownin Keweenawan lavas MI-2 and GM-10 would berepresentative of the late-stage stable crystallizationtrend of pyroxenes in tholeiitic magma under hyp-abyssal conditions.

Acknowledgments

We are grateful for discussions with members of the De-partment of Geology and Geophysics, University of Minne-sota and Geological Survey of Minnesota. Microprobeanalyses were performed using the equipment at the aboveDepartment, with the assistance of Paul W. Weiblen. I.Kushiro and N, Shimizu at the Carnegie Institution of Wash-ington reviewed an early draft of this manuscript. Part ofthis work was supported by National Science FoundationGrant GA l3-4ll to Green, which is deeply appreciated.

References

Brcr, M. E., Jn., eNo N. C. LrNosuv (1969) Paleomag-netism of the Beaver Bay Complex, Minnesota. l. Geo-phy s. R es. 7 4, 2002-20 13.

Boors, K. G. (1972) Paleomagnetism of some Lake Su-perior Keweenawan Rocks. U.S. Geo'|. Suru. Prof. Pap.760,42p.

eNo J. C. GnrrN (1972) Further paleomagneticdata for Keweenawan rocks in the western Lake Superiorarea (abstr.). Geol. Soc. Am. Abstr. Programs, 4, 454.

Crnrurcnent, L S. E. (1960) The pyroxenes and olivinesfrom some Tertiary acid glasses. J. Petrol.1,309-336.

(1964) The petrology of Thingmuli, a Tertiaryvolcano in eastern lceland. I. Petrol. 5,435-460.

(1967) The iron-titanium oxides of salic volcanicrocks and their associated ferromagnesian silicates. Con-trib. Mineral. Petrol. 14.36-64.

Eccr-en, D. H., lNo C. W. BunNuru (1973) Crystallizationand fractionation trends in the system andesite-H,O-CO-O, at pressures to 10 Kb. Geol. Soc. Am. BulI. 84,2517-2532.

Elrerpus, C. H., A. C. DuNnervr, eNo R. N. TnorrapsoN(1971) Iron-rich pigeonites from acid rocks in the Ter-tiary igneous province of Scotland. Am. Mineral. 56,940-951.

GsnrvreN, H. M., Jn. (1957) The Petrology of the BeaoerBay Gabbro Complex, Lake County, Minnesota. Ph.D.Thesis, University of Minnesota.

GnEeN, J. C. (1972) North Shore Volcanic Group andassociated intrusions. In, The Geology o'f Minnesota,P. K. Sims and G. B, Morey, Eds., Minn. Geol. Surv.,St. Paul, p.294-332.

Grsn, F. G. F. (1973) The zoned clinopyroxenes of theShiant Isles sill, Scotland. I . Petol. 14, 203-230.

Gnour. F. F., eNo G. M. Scuwenrz (1939) The petrologyof the anorthosites of the Minnesota coast of Lake Su-perior. Minn. Geol. Suru. BulL 28, ll9 p.

KoNol, Teusnr (1970) Pyroxenes from the Beaver Baygabbro complex of Minnesota. Contrib. Mineral. Petrol29,338-344.

KuNo, Hrs.lsnr (1969) Pigeonite bearing andesite and asso-ciated dacite from Asio, Japan. Am. I. Sci. 267-4,257-268.

LrNosrsv, D. H., euo J. L. MuNoz (1969) Subsolidus rela-tions along the join hedenbergite-ferrosilite. Am. l. Sci.267-,^,295-324.

Mum, I. D. (1954) Crystallization of pyroxenes in an ironrich diabase from Minnesota. Mineral. Mag. 30r 376-388.

eNo C. E. Trrrsv (1964) Iron enrichment andpyroxene fractionation in tholeiite. Geol. L 4, 143-156.

PHINNEv, W. C. (1970) Chemical relations between Ke-weenawan lavas and the Duluth complex, Minnesota'Geol. Soc. Am. BuIl.8l.2487-2496.

Srrven. L. T.. eNo J. C. GnseN (1972) Time constants forKeweenawan igneous activity (abstr.). GeoI. Soc. Am'Abstr. Program s, 4, 665.

Srurrs, Doucres (1971) Reconnaissance experiments with

compositions on the join Fs"oEn,-Wollastonite in tbepyroxene quadrilateral. Carnegie Inst. Wash. Year Book,70,125-129.

eNl D. H. LrNpsrsv (1971) Stable and metastableaugite crystallization trend in a single basalt flow. ,4n.

Mineral. 56,225-233.SreveNsoN, R. J. (1973) A Keweenawan layered mafic in-

trusion near Finland, Lake County, Minnesota (abstr.).

Igth Inst. on Lake Superior Geology, University of Wis-

consin, Madison, Wisconsin, 36.(1974\ A Keweenawan Layered Mafic Intrusion

near Finlond, Lake County, Minnesota. M.S. Thesis' Uni-versity of Minnesota, Duluth.

Manuscript receiued, September 24, 1973; accepted

for publication, August I, 1974.