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

Clinopyroxenesfrom the KeweenawanLavas 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 volcanic series: tholeiitic olivine basalt, basaltic andesite,quartz , and , which is assumed to be derived from a tholeiitic parental . These lavas contain calcic clinopyroxene which ranges from CarnMgnuFe,,to CaoMgr"Fenu.The crystallization trend of these clinopyroxenes is characterized by marked iron enrichment and is similar to that of the associated Beaver Bay Gabbro Complex and of the Skaergaard Intrusion. The mineral assemblageof 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.

fntuoduction gabbro far to the northeastand southwestof those The compositionaltrends and the phaserelations consideredhere, all of the rocks (including the of pyroxenesin intermediateand felsic volcanic BeaverBay Complex)which intrude the North Shore -+15 rocks of tholeiitic derivationhave beenrcported by Volcanic Group are within m.y. of the same many investigators(e.g., Carmichael,1960; Kuno, ageas the lavas (Silverand Green, 1972), showthe 1969; Emeleus,Dunham, and Thompson,l97l). same magneticpole position (Beck and Lindsley, One suchrather complete series of lavasof generally 1969; Books, 1972; Books and Green, 1.972) and tholeiitic affinitiesis the North ShoreVolcanic Group areassumed to be cogeneticwith thelavas. of Late Precambrianage, about 1130 million years The basaltsof the North ShoreVolcanic Group old, which underlies(Fig. 1) most of the shore of are nearly alLnon-porphyritic, the pyroxenesoccur- Lake Superior in northeasternMinnesota, U.S.A. ring generallyeither as poikilitic crystals in ophitic (Green, 1972). This series,which includesapproxi- or subophiticintergrowth with plagioclaseor as small mately7500 meters(25,000 feet) of lavas,contains equantcrystals in an intergranulartexture. The inter- sometypes which are more potassicthan is typical mediate and felsic lavas, however, are nearly all in pure tholeiiticsuites, and it rangesin composition porphyritic, with small ,clinopyroxene, from olivine tholeiites through q\artz tholeiites, and commonly also magnetiteand olivine pheno- basaltic andesites, trachybasalts,andesites, and crysts.The olivinehas in all casesbeen destroyed by quartz to . The olivine tholeiites are secondaryalteration, but the clinopyroxeneis fresh the most abundantsingle rock type. The flows are in manyrocks. intruded by the immenseDuluth Gabbro Complex To understandthe completetrend of compositions and by locally abundant shallow intrusive rocks, of the clinopyroxenesduring fractionationof tholei- principally olivine diabase.The Beaver Bay Com- itic magma,partial analysesof many crystalsfrom plex, describedby Grout and Schwartz(1939) and this volcanicseries were obtainedusing an electron Green (1972) and in part by Gehman(1957) is microprobe (Green), and the clinopyroxenesfrom one such group of intrusions.Muir (1954) and four large sampleswere separatedand analyzedby Konda (1970) have studiedthe clinopyroxenesof conventionalwet chemical methods (Konda). the Beaver Bay Complex. Except for lavas and The samplesstudied for this report are from lava I 190 CLINOPYROXENES FROM KEWEENAWAN LAVAS 11,91 flows of the uppermost 2000 meters of the North laths; subhedral olivine now completely altered to Shore Volcanic Group, plus a dike that crosscutsthe serpentine or chlorite; and both ilmenite and Ti- flows roughly 2600 meters from the top of the ex- magnetite. This basalt is the most "primitive" posed section. The sample localities are shown on (highest Mg/Fe ratio, lowest K2O) of the olivine Figure 1. This paper presents the results of these tholeiites analyzed in the North Shore Volcanic analyses and discussesthe petrologic significance of Group, and, according to calculations by Phinney the pyroxene crystallization trend thus outlined. (I970), it may be close to the parental magma com- position for the Duluth Gabbro Complex as well as Petrography for the North Shore Volcanic Group. Chemical analyses of most of the rocks sampled (b) Olivine basalt (T-22); on old road, NW 1/4 are given in Green (1972); they are plotted on an sec. 2, T.58N., R.5W. Medium-grained, subophitic; AFM diagram in Figure 2. Brief petrographic de- the upper-central part of a very thick flow. About scriptions of the rocks from which pyroxenes were 3 percent small olivine crystals, completely altered; studied follow. augite up to 3 mm diam., some as skeletal inter- (a) Olivine basalt (T-56) : lakeshore at E. edge growths with plagioclase; both ilm. and Ti-mag; sec. 12, T.59N., R.4W. Slightly porphyritic, fine- to plagioclase laths zoned, An65-An2p.Trace of inter- medium-grained basalt flow between 7-8 -m thick stitial granophyre. Saponite and zeolites fill dikty- (Fig. 3,A.). Major minerals: clots of small, euhedral taxitic cavities. bytownite phenocrysts zoned Ans5-Anool poikilitic (c) Ophitic olivine basalt (DY-3): lakeshore at augite up to 2 mm across enclosing small labradorite north edge of Sec. 15, T.60N., R.2W. The lower

Flc. 1. Generalizedgeologic map of Keweenawanrocks in the western Lake Superior area showing location of samplesstudied. Check pattern, intrusive rocks; vertical ruling, Lower Keweenawanrocks (magnetically reversed); stippled, Middle Keweenawan lavas (magnetically normal). After Green (1972, Fi'e. V-3). t192 T, KONDA, AND I. C. GREEN

maryinal zoning by changesin extinction angle and birefringence.Some augite shows very thin exsolu- tion lamellae(pigeonite?). Groundmass intergranu- lar in texture, with plagioclase, alkali-, augite, pigeonite, opaques, and greenish clay min- erals.The pigeonitetends to be oxidized along rims and fractures. il-z .xt-4 (f ) Basalticandesite (MI-4): lakeshore,SE 1'/4 " Fal ' 1-16 Sec. 22, T.62N., R.3E. Fine-grained red-brown aphyric andesiteor trachy-andesiteflow showingthin oxidation banding. Intergranular to intersertal tex- ture with small plagioclaselaths (An 53-45), ffi?8- netite octahedra,and 8 percent blocky, fresh ferro- augitesin a groundmassof oxidized glass,magnetite, alkali-feldspar, quartz, and amphibole (Fig. 3D). Cavities are filled with quartz, chlorite, clay, and Fro. 2. AFM diagram for rocks from which pyroxene zeolites.Flow thicknessunknown. analyses aro given in this paper. Point labelled GH-7a is (g) Porphyriticandesite or trachyandesite(H-4): derived from rock analysis S & G-5-9 (from the same flow Highway61, centerline Sec.12,T.62N, R.4E. Very but a diferent locality), and point DY-3 is derived frorn fine-grained,dark brown dike, which transectsthe rock analysis DY-6b (from a similar flow slightly lower in lavas about 2600 m below the top of the exposed the sestion than DY-3), Rock analyses from Green (1912). section. About 3 percent of small phenocrysts of subhedraloligoclase (Anzu), ferroaugite,and mag- middle part of a very thick (200 m) flow of netite, in a groundmassof plagioclase,alkali-feld- fine- to medium-grainedbasalt. Tabular plagioclase spar, ferroaugite, magnetite, anhedral hornblende, predominates, with fresh ophitic augite, olivine possibledevitrified brown glass,and alteration prod- pseudomorphs,magnetite and ilmenite (Fig. 3B). ucts. Interstitial cavitiescontain zeolites and calcite. (h) Porphyritic intermediate qaartz latite (MI- (d) Porphyriticbasalt (GH-7A): lakeshore500 2): Highway61 at W. edgeSec. 6, T.62N., R.3E. m westof powerplant at GrandMarais; SE 1/4 SW Fifty m-thick lava flow rather similar to rocks termed l/4 Sec.20, T.61N., R.lE. The chilledbase of a "icelandites"by Carmichael(1964). Reddishbrown, thick (110 m) basaltic to andesiticsheet that is fine-grained;about 7 percent of small phenocrysts probablya flow but may be a sill. Abundant 1-2 mm of stout tabular, unzoned andesine(Atuo), ferro- phenocrystsof zonedplagioclase, abott 2 percent of augite, altered olivine, and magrretite.The ferro- fresh prismatic augite crystals (0.5-1.5 mm) and augite phenocrystsare fresh, about 1 mm across; fewer small altered olivine phenocrystsset in a sub- some are aggregatedwith the other phenocrysts spherulitic, fine-grained groundmassof plagioclase (Fig. 3E). The groundmassminerals are plagioclase, tablets,clinopyroxene prisms, and magnetiteblocks, alkali-feldspar,interstitial quartz,opaques, and green with interstitial alkali feldspar in the coarser areas. alteration products, probably chlorite. Although Original small miarolitic cavities filled with quartz there may have been clinopyroxenein the ground- andan oxidizedchloritic mineral. mass,it is not now recognizable. (e) Porphyritic basaltic andesite(F-7a): lake- (i) Porphyritic rhyolite (GM-10): lakeshore, shore,NE l/4 Sec.21, T.57N, R.6W. The basal, centerline Sec. 14, T.61N, R.1E. Basal spherulitic chilled portion of a very thick (at least 100 rr) zone of a rhyolite lava flow that is as much as 155 m flow. Red-brown, fine-grained potassic basaltic thick. Orange-red,aphanitic, about 3 percent small andesitewith 10 percent small phenocrystsof plagi- blocky phenocrystsof plagioclase(An35-26) and rare oclase (Anss), augite, rare pigeonite,and olivine qvartz, ferroaugite, magnetite and altered olivine. (Fig. 3C). The olivine phenocrysts(completely The ferroaugite occurs as small, fresh, subhedral alteredto iddingsite)are prismaticand someare in phenocrysts(0.1 x 0.8 mm) (Fig. 3F). Ground- parallel intergrowth with augite.Augite phenocrysts mass microcrystallinefeldspar and quartz with mi- euhedral, pale green, nonpleochroic. A few show nuteopaque grains. CLINOPYROXENES FROM KEWEENAWAN LAI/AS rt93

(k) Porphyritic intermediatequaftz latite (F-21): small, blocky albite phenocrysts,about 3 percent NE 1/4 Sec. 28, T.57N., R.7W. Red-brown,very spongy augite prisms and about 2 percent each of fine-grained"andesite" assumedto be a lava al- magnetiteand oxidizedolivine. Groundmassdomi- though contacts not exposed.Contains 10 percent nated by K-feldspar, with lesser amounts of plagi-

Flc. 3. Photomicrographs of clinopyroxenesin KeweenawanLavas. A, ophitic augite enclosing plagioclaseand pseudomorphs of olivine in olivine tholeiite T-56; B, ophitic augite enclosing plagioclase with olivine pseudomorph in olivine tholeiite DY-6b (flow closeto and similar to pyroxenesample DY-3);C,augite and plagioclasephenocrysts, olivine pseudomorphin basalticandesite F-|a;D,equant augite crystals with magnetite and plagioclasein andesiteMI-4; E, prismatic ferroaugite, plagioclase,and olivine- pseudomorphphenocrysts with abundant apatite in intermediate quartz latite MI-2; F, prismatic ferroaugite, plagioclase,and small magnetite phenocrystsin rhyolite (felsic quartz latite) GM-10. l194 T, KONDA. AND I, C. GREEN oclase, quartz, yellow-green secondary minerals, apatite,magnetite, and ilmenite. t. ^ oo ^rto r-22- -l'fri_, .sf*i "" Analytical Results H-4 The clinopyroxene phenocrysts of three large x T-56 + Ml-z samplesand subophitic augite of a fourth were . DY-3 o F-21 separated,using the isodynamic separator and o GH-7o r GM-IO A Clerici's solution, and analyzedby conventionalwet MI-4 chemicalmethods (Konda). Groundmasspyroxenes in the porphyritic rocks either wgre not present Frc. 4. Atomic ratios of analyzed calcic clinopyroxenes on pyroxene quadrilateral. Points shown by circle around (MI-z) or were avoidedby sizeclassification during dot are wet-chemically analyzed (Table l); others are indi- separation.In addition,about forty partial analyses vidual point analysesby electron microprobe, exceptGM-10 (for Mg, Fe, Ca) of pyroxenesfrom six other lavas which is the average of 4 points on each of two grains. All describedabove were obtainedby Green using the are phenocrysts except T-56 and DY-3 (poikilitic augite), (equant groundmass py- Mac electron microprobe at the Department of T-22 (subophitic augite), MI-4 roxene). Geologyand Geophysicsat the University of Minne- sota, Minneapolis,with the assistanceof Paul W. Weiblen, who also analyzedthe pyroxene from the positions with a slight increase in iron content. The rhyolite GM-10. The standards for microprobe compositional variation within the pyroxenes from analyseswere pyroxenesanalyzed by wet chemical saturated,intergranular basalts and basaltic andesites methods.The resultsof the wet chemicalanalyses (e.g. Ml4) is mostly in Mg/Fe ratio, nearly parallel areshown in Table 1, andthe Mg/Fe/Ca relationsof to the Di-Hd join. The ferroaugites of the intermedi- all analyzedpyroxenes are shownin Figure 4. ate quafizlatite,Ml-z, show enrichment in both Ca A weak zoning in the ophitic pyroxenes of the and Fe at the margin, and the qnartz latite F-21 olivinetholeiites T-56 and DY-3 is suggested,trend- shows principally Mg/Fe zoning. Microprobe study ing away from diopsidic and toward subcalciccom- of the clinopyroxene from rhyolite GM-10 showed no zoning present. TesI,n 1. The Chemical Compositions and Atornic Ratios of Calcic Clinopyroxenes from Keweenawan Flows, North Discussion Shore Volcanic Group, Minnesota Figure 5 shows the correlation between the atomic T. Konda, Analyst (wet chem.) Fe/(Fe*Mg) ratios of the clinopyroxenes and the ratio of total FeO to total FeO*MgO in the ana- lyzed host lavas. Figure 6 shows the pyroxene Fe/ si02 !5.51+ !8.11 si 1.880 r.923 r.852 ),.9]6 50.74 51.00 (Fe+Mg) plotted the Tto2 I. 16 0.93 0.82 0.80 Alrv 0.120 O.0?7 o,ror 0,061 ratio against SiOz content of |JzO3 2,98 3.rB 2.18 r. 31 T1 o.o2l+ o.023 host rock. Because the percentage of pyroxene phe- Fe203 2.O, z.j9 L.j2 r.Lo 0.023 ----- nocrysts in the porphyritic lavas is very small, the Feo ro-6j t9.zz 22.\o lvl o,oto o. 06lr 1,3t1 whole-rock Fe/Mg ratios do not depend greatly l,{no 0,16 o.29 0.37 0.1+7 Tt o.o32 0.a26 Mso L5.53 13.08 10.91 6.26 Fe+3 o. 05? o. o?3 0.022 0.041 on the pyroxene phenocryst compositions. Conse- cao t9.66 16.58 15.09 r8.37 Fu+2 o.22'l 0.336 o.530 0.1\5 quently, the general correlations shown in Figures 5 0,009 o, or2 o. o!1 Na20 O.51 0,611 0.50 a,25 un 0.003 and 6 can be interpreted as indicating that the KzO 0.09 0.1? 0.28 o.26 Ms 0.858 0. ?35 O.6t!2 O.3-lr H2o(+ ) o. L5 0.86 o. 8! 0.53 ca 0.780 0.5'fo 0.685 o.783 pyroxene phenocryst compositions were controlled H2o(-) o.lo 0.3! ------Na 0.0!0 0.047 o,ol+5 0.019 by the compositions of the lavas in which we now PzO, 0.oo o.0r 0.20 0.40 K o.oolr 0.005 0.013 0.or3 find them. Thus, the phenocrysts were probably P ------0,005 0.013 Total L00.83 1O0,32 99,57 Loo.59 z 2.oo0 2.D00 2.000 2.000 indigenous,and not included xenocrysts. wrt 2.0rr 1.96T 1.9q1 r.9q9 The clinopyroxenes in samplesF-7a, GH-7a, and At@ic tr H-4 show slightly low Fel(Fe+Mg) ratios (Fig. 5) Ca !0.5 36.'t 3!.1{ !o.1 Ma LI+.6 l0.l 32.2 19.0 as compared to the general trend. Under the micro- Fe* f!,9 22.9 33.1 !0.9 scope, these rocks contain small amounts of pheno- 2 crystic magnetite, suggesting that the magnetite may have precipitated during or before the crystallization Ferroeugite phenocrysts fr@ 6wite edestte (H-l+) FeFoau8ite phenocitrsts fron lntemealiate qustz Iatlte (Mf-2) of clinopyroxene and prevented further iron enrich- CLINOPYROXENES FROM KEWEENAWAN LAVAS 1195

and Lindsley (197I). We interptet this lack of a "quench trend" to be the result of relatively slow crystallizationof these thick, flood basalts. Slow 3 crystallization is also indicated by their relatively coarse grain size and their flat surfaces and wide LP o extent. The observedtrend is instead inferred to be the "intratelluric equilibrium trend" referred to by Muir and Tilley. Clinopyroxenesfrom the consanguineousBeaver 05 Bay Complex have been studiedby Muir (1954) ta^?6r.o 7 ttotarLo - Mso), host rock and more recentlyby Konda (1970) and Stevenson Frc. 5. Atomic Fel(Fe*Mg) of clinopyroxenesplotted (1,973,1974).Because the pyroxenephenocrysts in FeOfMgO) (wt. percent) of host againsttotal FeO/(total the lavas could be expectedto have formed during rocks. Augite DY-3 is compared with analyzed DY-6b, a similar flow slightly lower in the section. Augite GH-7a is hypabyssaldifferentiation under conditions similar compared with analyzed rock S & G-59 from another Io' to those obtaining during crystallization of the cality in the same flow. Rock analysesfrom Green (1972). BeaverBay intrusiverocks, it is interestingto com- pare the clinopyroxenesof the lavas with those of ment in the pyroxene. Although Eggler and Burn- the BeaverBay Complex. Although the lava samples ham (1973) find that magnetiteshould not o€curat describedin this papercame from a muchwider area the liquidus in andesitic , it evidently did than the BeaverBay Complex(Fig. 1), the controls startto precipitatein theselavas at sometime during on and characterof differentiationof both may well the first 10 to 20 percentof crystallization.flowever, have been similar. Some of the sampledlavas may there is less magnetite than pyroxene present as actually have been surface extrusions of the same phenocrystsin theserocks, and there is no evidence magmasthat producedparts of the BeaverBay Com- to indicate that it siarted to crystallize before the plex, although it would be difficult to demonstrate clinopyroxene.Furthermore, these samples have no sucha relationship. more magnetiteas phenocryststhan the samples In chemical characteristics,the clinopyroxenes falling on the "general trend." The relatively low from the lavas of the North Shore Volcanic Group Fe/(Fe+Mg) ratios of these three clinopyroxenes are slightly richer in Al2O3 and Na2O, but have comparedto their host rocksmay be a reflectionof about the same amount of MnO and TiOp as those the fact that their small content of pyroxene pheno- of similar Fe/Mg ratio from the BeaverBay Com- crystsrepresents only the first, most magnesianpart plex (Konda, l97O). The wet chemicalanalyses of of their normative pyroxene content; whereasin Table 1 show that the CaO contentsof the clino- MI-4, an aphyric andesite,the analyzedpyroxene is pyroxenesfrom most of the lava flows are slightly in the groundmassand shouldmore nearlyapproach the normativevalues. In the three more felsic sam- ples (F-21, MI-2, GM-10) the analyzedphenocrysts probably contain most of the potential pyroxene in I GM-1O the rock. o Also, these three intermediatesamples GH-7a, ",-,{1.-o F-7a, and H-4 show a high contentof magnetitein rl the groundmassas comparedto the other rocks, ,9 xH-4 I which would increasethe total FeO/(FeO*MgO) \o *Mr-4 xr-fo total ratio ofthe rock analysis. .o Gu-zoI ot^tl The compositionalzoning of theseclinopyroxenes *r-r, (Fig. a) is similar to the pyroxenetrend direction T-56 observedin many layered intrusions (e.B'.,Skaer- gaard and Beaver Bay), and althoughthese rocks 45 50 55 60 65 70 75 wt host rock are obviously lavas, the basaltic groundmasspy- "/. SiO2, roxenes analyzeddo not show any "quench trend" Frc. 6. Atomic Fel(Fe*Mg) of calcic clinopyroxenes referred to by Muir and Tilley (1964) or Smith plotted againstweight percentSiO, in host rocks. 1196 T. KONDA, AND J. C. GREEN lower than those from the Beaver Bay intrusion The assemblage of mafic minerals present is (Fig. 7). This might indicate slightly higher tem- closely related to the crystallization trend of py- peratures of crystallization for the phenocrysts of roxenes at the late stages of fractionation, as sug- the lavas than for the intrusive rocks. Ilowever, gested from experimental work (Lindsley and microprobe analysesof clinopyroxenes from DY-3, Munoz, 1969) andfrom natural rocks (Kuno, 1969; F-21, and MI-4 (Fig. 4) show essentiallythe same Emeleus et al, l97l). The most common phenocryst or lower CaO content compared to those from the assemblages of the mafic minerals in the analyzed Beaver Bay Complex, and the two populations samples are as follows: augite * olivine in basalt would be difficult to distinguish. (early stage; no phenocrysts); augite * olivine The crystallization trends of the clinopyroxenesin phenocrysts in basaltic andesite (GH-7a); augite * the Keweenawan flows and the cumulus augites in pigeonite * olivine * magnetite in basaltic andesite the Beaver Bay Complex are similar, as shown in (F-7a); ferroaugite * olivine * magnetite in inter- Figure 7. However, in the Beaver Bay Complex, mediate quaftz latite (MI-2); and ferroaugite, oli- Konda (1970) shows that the calcic clinopyroxenes vine, and magnetite in rhyolite (GM-10) at a late are locally rimmed by pigeonite-ferropigeonite, stage.Ferroaugite phenocrysts from MI-2 and GM- which is inferred to have crystallized by the reaction 10 have Fe/(Fe*Mg) ratios of 0.63-0.73, which between cumulus Ca-rich clinopyroxenes and the are similar to those of ferroaugite from porphyritic interstitial liquid. On the other hand, the calcic felsite and granophyre in Scotland (Emeleus er a/, clinopyroxenes from the related flows (this paper) l97I,no.8 of group 3 and nos. 5a,5b of group 4, do not have such a mantle of Ca-poor pyroxene, Table 2). Ilowever, the mineral assemblagesare probably becausesuch a reaction point had not been different: ferroaugite * olivine in the Keweenawan reachedwhen the lavas were extruded and quenched. lavas, and ferroaugite + ferro-pigeonite in the As mentioned above, the potential Ca-poor (and Fe- Scottishfelsic rocks. rich) pyroxene components probably remain in the Based on the experimental work of Lindsley and unanalyzedgroundmass ferromagnesian minerals. Munoz (1969, p. 316), the reaction Ferrosilite = Figure 7 further shows for comparison the trend Fayalite + SiO, melt is such that an increase in the of clinopyroxene compositions for a variety of Ter- activity of SiOz leads to an increasein the activity of tiary and Quaternary volcanic rocks (Carmichael, Fs under equilibrium conditions. Persistenceof oli- 1967); it is essentially identical to that of the ana- vine throughout the crystallization sequence and a lyzed Keweenawan pyroxenes. From a study of the higher content of alkalies in these Keweenawanlavas associatedoxide minerals, Carmichael (1967) infers than in the typical tholeiitic suites and Scottish felsic pyroxene crystallization temperatures ranging from rocks indicate the magma compositions to be slightly 970" to 915"C for similar intermediate and felsic sub-alkaline and suggesta slightly low SiOz activity. porphyritic lavas. However, the Keweenawan trends clearly differ from that of the mildly alkaline Shiant Isles sill (Gibb, 1973), which is more calcic and lacks the central dip in Ca content (Fig. 7). Although SiO2 activity is close to unity near the solidus for each Keween- awan lava, it would appear that SiOz activity did not increase fast enough below the liquidus to stabilize Ca-poor pyroxene. On the other hand, however, Mg/Fe distribution values between coexisting ferro- pigeonite (5c, group 4) and ferroaugite (5a and 5b, group 4) in the granophyres of Scotland are FIo. 7. Calcic pyroxene trends compared. Crosses,wet- somewhat variable, possibly suggesting metastable chemically analyzed clinopyroxenes frorn Keweenawan crystallization of ferropigeonite in the Scottish grano- lavas; circles, average compositions of clinopyroxenes from phyre. Also, from the descriptions of the ferro- Keweenawanlavas analyzed by microprobe; dots, analyzed pigeonite-bearing rocks (Emeleus et aI, 1971, p. pyroxenesfrom Beaver Bay complex (Konda, 1970); solid line, trend of clinopyroxenes from various Tertiary and 941-944), the crystals are rather variable in com- Quaternary volcanic rocks (Carmichael, 1967 ) ; dashedline, position within individual samples and some show trend of clinopyroxenesfrom Shiant Isles sill (Gibb, 1973). complex zoning. This too would suggest that the CLINOPYROXENES FROM KEWEENAWAN LAVAS lt97 ferropigeonitesdescribed result from disequilibrium Elrerpus, C. H., A. C. DuNnervr, eNo R. N. TnorrapsoN crystallization, perhaps related to an extension of (1971) Iron-rich pigeonitesfrom acid rocks in the Ter- igneous province of Scotland. Am. Mineral. 56, Muir and Tilley's "quenchtrend" (1964). The pres- tiary 940-951. ent data agreewith experimentsby Smith (1971) on GsnrvreN,H. M., Jn. (1957) The Petrology of the Beaoer compositionsalong the join Fss5En15-Wo.These in- Bay Gabbro Complex, Lake County, Minnesota. Ph.D. dicate that the assemblagecalcic clinopyroxene- Thesis, University of Minnesota. fayalitic olivine-silica is stable at low crustal pres- GnEeN, J. C. (1972) North Shore Volcanic Group and intrusions. In, The Geology o'f Minnesota, andtemperatures of 1000' to associated sures 800oC. P. K. Sims and G. B, Morey, Eds., Minn. Geol. Surv., The stable mineral assemblageduring late stages St. Paul,p.294-332. of fractionation appearsto be ferroaugite* fayalitic Grsn, F. G. F. (1973) The zoned clinopyroxenesof the olivine + qnartz,and the crystallizationtrend shown Shiant Isles sill, Scotland. I . Petol. 14, 203-230. in Keweenawanlavas MI-2 and GM-10 would be Gnour. F. F., eNo G. M. Scuwenrz (1939) The petrology of the anorthositesof the Minnesota coast of Lake Su- representativeof the late-stagestable crystallization perior. Minn. Geol. Suru. BulL 28, ll9 p. trend of pyroxenesin tholeiitic magmaunder hyp- KoNol, Teusnr (1970) Pyroxenes from the Beaver Bay abyssalconditions. gabbro complex of Minnesota. Contrib. Mineral. Petrol 29,338-344. Acknowledgments KuNo, Hrs.lsnr (1969) Pigeonitebearing andesiteand asso- ciated dacite from Asio, Japan.Am. I. Sci. 267-4,257- We are grateful for discussionswith members of the De- 268. partment of Geology and Geophysics, University of Minne- LrNosrsv, D. H., euo J. L. MuNoz (1969) Subsolidusrela- sota and Geological Survey of Minnesota. Microprobe tions along the join hedenbergite-ferrosilite.Am. l. Sci. analyses were performed using the equipment at the above 267-,^,295-324. Department, with the assistanceof Paul W. Weiblen. I. Mum, I. D. (1954) Crystallization of pyroxenesin an iron Kushiro and N, Shimizu at the Carnegie Institution of Wash- rich diabasefrom Minnesota.Mineral. Mag. 30r 376-388. ington reviewed an early draft of this manuscript. Part of eNo C. E. Trrrsv (1964) Iron enrichment and this work was supported by National Science Foundation pyroxene fractionation in tholeiite. Geol. L 4, 143-156. Grant GA l3-4ll to Green, which is deeply appreciated. PHINNEv,W. C. (1970) Chemical relations between Ke- References weenawan lavas and the Duluth complex, Minnesota' Geol. Soc.Am. BuIl.8l.2487-2496. Brcr, M. E., Jn., eNo N. C. LrNosuv (1969) Paleomag- Srrven. L. T.. eNo J. C. GnseN (1972) Time constantsfor netism of the Beaver Bay Complex, Minnesota. l. Geo- Keweenawan igneous activity (abstr.). GeoI. Soc. Am' phys. R es. 74, 2002-20 13. Abstr. Programs, 4, 665. Boors, K. G. (1972) Paleomagnetismof some Lake Su- Srurrs, Doucres (1971) Reconnaissanceexperiments with perior Keweenawan Rocks. U.S. Geo'|. Suru. Prof. Pap. compositions on the join Fs"oEn,-Wollastonite in tbe 760,42p. pyroxene quadrilateral. Carnegie Inst. Wash. Year Book, eNo J. C. GnrrN (1972) Further paleomagnetic 70,125-129. data for Keweenawan rocks in the western Lake Superior eNl D. H. LrNpsrsv (1971) Stable and metastable area (abstr.). Geol. Soc. Am. Abstr. Programs,4, 454. augite crystallization trend in a single basalt flow. ,4n. Crnrurcnent, L S. E. (1960) The pyroxenes and olivines Mineral. 56,225-233. from some Tertiary acid glasses.J. Petrol.1,309-336. SreveNsoN,R. J. (1973) A Keweenawanlayered mafic in- (1964) The petrology of Thingmuli, a Tertiary trusion near Finland, Lake County, Minnesota (abstr.). in easternlceland. I. Petrol. 5,435-460. Igth Inst. on Lake Superior Geology, University of Wis- (1967) The iron-titanium oxides of salic volcanic consin,Madison, Wisconsin, 36. rocks and their associated ferromagnesian silicates. Con- (1974\ A Keweenawan Layered Mafic Intrusion trib. Mineral. Petrol. 14.36-64. near Finlond, Lake County, Minnesota. M.S. Thesis' Uni- Eccr-en,D. H., lNo C. W. BunNuru (1973) Crystallization versity of Minnesota, Duluth. and fractionation trends in the system andesite-H,O- CO-O, at pressuresto 10 Kb. Geol. Soc. Am. BulI. 84, Manuscript receiued, September 24, 1973; accepted 2517-2532. for publication,August I, 1974.