r53 The Canadian Mineralogist V o l . 3 3 , p p . 1 5 3 - 1 6 3( 1 9 9 5 ) BEHAVIOR OFTHEPLATINUM.GROUP ELEMENTS DURINGDIFFERENTIATION OFTHENORTHMOUNTAINBASALT,NOVASCOTIA JOHND. GREENOUGH Depanment of Geological Sciences, University of British Columbia, Okanagan University College, 3333 College Wa-,-,Kelowna, British Columbia VIV 1V7 BRIAN J. FRYER* Centre for Eanh Resources Research, Depaftment of Earth Sciences, Memorial Universitv of New'foundlancl, St.John's, Newfoundland, AIB 3Xs ABSTRACT Differentiated rocks in thick flows ofthe Jurassic North Mountain Basalt, Nova Scotia, display evidence for fractionation of noble metals (Au, Pd, ft, Rh, Ru, and Ir). Meter-thick layers of mafic pegmatile and vesicular basalt high in the flbws are enriched in Au and Pd but depleted in Pt, Rh, Ru and Ir relative to undifferentiated basalt. Mineral precipitation (e.9., chromite removal) cannot explain Au and Pd enrichment and Pt depletion in the mafic pegmatites.The fraclionation pattem may reflect the early movement and concentration of chloride-bearing hydrothermal solutions associatedwith rising plumes of vesicles only months after extrusion. On average, basalts with orthopyroxene show Pd and Au depletions and Pt, Rh, Ru and Ir enrichments relative to undifferentiated basalt, but there are large differences in concentration of individual noble metals between samples. This intersample variability, resulting in Pt/Pd fractionation, could reflect noble-metal-bearing microinclusions in orthopyroxene. "Rhyolite" bands derived from the mafic pegmatites, possibly through silicate liquid immiscibility, contain the lowest concentrationsof noble metals in the flows. Some "rhyolite" samples show Ir enrichment, polentially reflecting selective carrying capacity for magmatic fluids segregatedfrom the "rhyolite" or, if the "rhyolites" formed through silicate liquid immiscibility, preferential partitioning of Ir into the siliceous liquid. These findings, and previous studies of other Mesozoic mafic rocks from the eastern U.S., indicate that the incompatible behavior of Pd and Au can result in substantial increasesin concentration of these noble metals in the upper reachesof thick flows and intrusions. Keywords: tholeiitic basalt, Jurassic, eastem North America, "rhyolite", platinum-group elements, differentiation, volatile phase,liquid immiscibility, Nonh Mountain Basalt Formation, Nova Scotia. SonanaatnB Les rochesdiffdrenci6es descoul6esbasalriques d'dge 6paisses de la formationde North Mountain,en Nouvelle-Ecosse, jurassique,ont subi les effetsd'un fractionnement en metauxnobles(Au, Pd, Pt, Rh, Ru et Ir). Les niveauxde pegmatite mafiqueet de basaltev6siculaire,donl l'6paisseurest e l'6chelledu mbtreet la position est dansla partie sup6rieuredes primitives.Le fractioncoul6es,sontenrichisen Au et Pd, et appauvrisen ft, Rh, Ru et Ir par rapponaux rochesbasaltiques nementde min6raux,par exemplela chromite,ne pourraitexpliquerl'enrichissement de en Au et Pd des 6chantillons pegmatitemafique,et leur appauvrissement pounaientr6sulterdu mouvementpr6coceet de la en Pt. Parcontre,cestendances concentration de saumuresassoci6es avecI'ascensionde bulles d peinequelquesmois aprbsI'extrusion.En moyenne,les dchantillons de basaltetr orthopyroxBne sontappauvrisenPd et Au, et enrichisen ft, Rh, Ru et Ir, parrappo( aux 6chantillons de basalteprimitii maisnousdocumentons s6lectives desdiffdrencesimportantes d'un 6chantillonh l'autre.Cettevariabilit6, qui estd l'origine d'une s6paration de Pt et Pd,pounaitr6sulterde micro-inclusions conlenantla suitede m6tauxnoblesdans par immiscibilit6du liquidesiliceux, I'orthopyroxbne. Des niveaux"rhyolitiques"d6riv6sde la pegmatitemafique,peut-Ctre contiennent lesplusfaiblesconcentrations desm6tauxnobles.Certains6chantillons "rhyolitiques"tdmoignentd'un enrichissements6lectifen h, et doncpeut-Otre d'une capacit6s6lectivede transfertdansune phasefluide magmatique ou, dansle cas d'une originepar immiscibilit6,d'une r6partitionpr6f6rentiellede l'iridium dansle liquide siliceux.Cesr6sultats,et ceux d'6tudescomparables incompatiblede de suitesmafiquesm6sozoiques de I'est desEtats-Unis,montrentquele comportement Pd et Au pourraitcauserun enrichissement de coul6es6paisses subtantielde cesm6tauxnoblesdansles partiessup6rieures et de massifsintrusifs. (Traduitparla Rddaction) Mots-cl6s:basaltethol6iitique,jurassique,est de I'Am6riquedu Nord, "rhyolite", 6l6mentsdu groupedu platine,diff6renciation,phasevolatile,immiscibilit6liquide,Formationdebasaltes de NorthMountain,Nouvelle-Ecosse. *Presentaddress: Dean'sOffice,Facultyof Science,Universityof Windsor,Windsor,OntarioN9B 3P4. t54 T}IE CANADIAN MINERALOGIST INTRoDUcrroN The North Mountain Basalt sequenceborders shoresof the Bay of Fundy in Atlantic Canadaand providesan exampleof continentalflood-basaltvolcanismassociated with Jurassicrifting betweennorthern Africa and eastemNorth America (Wark & Clarke 1980;Fig. 1). The basaltshaveattractedconsiderable attentionas a naturallaboratoryfor igneousprocesses during a major rifting event.Severalauthorsused them to study crustal assimilation(Dostal & Dupuy 1984, Greenoughet al. 1989).Othersused this episodeof volcanismto placeconstraintson the time of rifting as well as the absolute age of the Triassic-Jurassic boundary(Hodych & Dunning 1992). In contrastto other flood-basaltsequences, which can be thousandsof metersin thickness.the North MountainBasaltsequence is merelya few hundred merersthick (White 1989,Papeziket al. 1988). However,two of the lava flows are amongstthe largestknown individualflows on earth,with dimensionsof -0.2 km x 100km x 230 km (Greenough& Papezik1987).As a resultof theirthickness,thesetwo flows underwentsubstantialdifferentiation.Bowen ( I 9 I 7) recognized the utility of studyingtheserocksto help identify igneousprocesses that aremaskedwhere coolingis slower,as in gabbroicintrusions. These basaltspresbrveevidenceof crystal settling,silicate liquid immiscibilityandvolatilemovementduringthe differentiation of the mafic magma(Papeziket al. 1988,Greenough & Dostal1992a,b). I The mechanismscontrollingdifferentiationin the North MountainBasalt sequencemay not be well understood.However,the cooling regimefor lava as a result,the time flows can be closelyconstrained; when variousprocesses of differentiationoccurredis reasonablywell known (Greenough& Dostal 1992b,. that "Rhyolite" bandswithin the basaltsdemonstrate substantialdifferentiationoccurred(Greenough& Dostal1992a).The basaltsthusprovidean oppormniry to study the effectsof magmadifferentiationon elementslike the platinum-group elements(PGE),whose behaviorduring magmaevolutionis not well understood(Crocket1981,Briigmannet al. 1987,Fryer & sf mafic G r e e n o u g h1 9 9 2 ) .S i m i l a ro u t p o u r i n g o that differentiation magmain the U.S. demonstrate mechanismscan enrich specificPGE more than ten-fold(Gofffried& Froelich1988). In this paper,we reporton the PGE concentrations in variousrock typesfrom the North MountainBasalt Thesedata allow us to commenton the sequence. effectsof differentiationon PGEconcentrations. GEoI-ocv AND SAMPLINc The Jurassic(202 ! I Ma: Hodych& Dunning 1992)North MountainBasaltFormationconsistsof a 150-to 2O0-m-thicklower flow separated from an upperflow of similar thicknessby a 50-m-thick middle unit of thin flows. The upperflow is locally overlainby 75 m of thin flows (Papeziket al. 1988, Greenough et al. 1989).The two thick flows aretracefor over 230 km able and isotopicallydistinguishable r u o n r HM o U N T A t N \ c ,!, (t c o E Eo u 0, ot a CAP d,OR t cHlt{Atl|PAs ORAT{D llANAr{ ITLAND Flc. l. Map showingthe distributionof North MountainBasaltaroundthe Bay of Fundy. NorthMountainBasaltalsooccursin theChinampas exploratoryoil well (x). 155 BEHAVIOROF T}IE PGE.NORTHMOUNTAINBASALT along the Bay of Fundy (Fig. 1), but they cannorbe distinguishedon the basisof major- or trace-element g e o c h e m i s t r ya l o n e ( J o n e s& M o s s m a n1 9 8 8 , Greenoughet al. 1989).The rocks consistof high-Ti quartz-nonnative tholeiite(Weigand& Ragland1970) representative of the Early JurassiceasternNorth Americandoleriteprovince (Dostal & Dupuy 1984, Papeziket al. 1988). The lower portionsof thetwo thick flows tendto be massive,columnarjointed, and have accumulated orthopyroxene in the Digby area(Papeziket al. 1988). The percentages of orthopyroxenedecreaseto the TABLE 1. NOBLE METAL CONCD.ITRATIONS IN NORTH MOUNTAIN BASALT $nple# Rk aypeI udrF Lmtiod EF-r(C) Pemlite I E Fary EF-I(D Pegualite I E, Fcry EF-2(C) Fe@aite t E Fery Noble Maals Au 3.9 3.3 5.1 Pd 8.5 t6 lt R t.9 3.5 1.3 Rh 0.r9 0.25 0.21 fu Ld Ld. n.d. 0.06 h 0.@ 0.05 $daphile md ChalcophileElmts Cr n.d. nd" &d. M n.d. && !.dG! 176 138 156 EF-3(C) P€gEalite t E F6ry EF-3(F] Rhyolite I E Fcry EFJ Bsslt I E Fary FP-3 Easlt I Frcepon 3.9 ll 0.98 0.t6 od. ud. o.24 o.42 t.3 0.01 &d. 0.05 2.7 1,5 4.6 o.a D.d. 0.06 4.7 9.0 6.r 0.36 ud. 0.t2 rd" 4 t67 56 24 93 t9 Ld. 253 258 3l 94 1"4(C) Pqedtr I Fntport 5.6 l.? L5 0.10 od. 0.t0 u& !d. r00 &mplelt F?-9(C) r?-9(Fl IUCKr iICIOB MCKIF'GltCKtT MCI(3B rttct€Fc RL d Pg@tite R$yolite Bselt P€gmtite Rbyolile P€g@aiteP€gmrireBbyotire udrtr333333333333 Iedod Fr€€pon tr'rc€pon llcKay IL McI&y tL lttcltay E McI&y g. McIQyH. McIGyIL FP-4(n Rbyotre I Frrcport 1.2 0.r8 0.34 0.6 0.17 t.2 ud. ud. 49 FP-7A(C) P€gEarite I Fre.port 6.6 u 3.2 0,21 s4 0.ll ud. t 7t MCIST Mn2FC Pegaasite Rhyolite tUcl(ayE MctGyH- FP-7A(D Rfrybite I Fr€€gort o.d" 0.6 0.41 0.08 a.d, o.31 !.d. !.d. 109 FP{BG) R$yolite I Fntport 0.51 0.44 0.50 0.05 ed. 0.10 !.d. nd8t MIIVP llcKlo PegmatiteV6idfrr tt{cKayE lllcXayE NobleMelals 3.6 l0 t.z 0.16 Ld. Ld. AU Pd P. Rh R[ It 4.74 0.21 0.28 a.d. ad. ud 1.8 8.1 4.6 g.2S 0.18 a.d. Sid€ropbileed Chdcophile tselrenb u G nd. rd. l{ ll LrL 34 or lgl 170 65 Samplet Rk typd Unitf Iedotf PEZ-Z! OPX ss. t E. Frry NoUs M€trb t6 Alr Pd 7,7 R 14 Rh 0.57 Ru 0.27 Ir 0.40 PA,A4 Blsslt I E. Frry ta 4.2 5.5 0.28 Ld. 0.06 P&-25 PeSmtire I E Frry ta 3.9 0.45 0.05 Ld. !.d. Sidophile md Chalophile EI@UG 368 Cr 85 l0 Ni9048m 91 147 Cu 192 l. J. l0 0.t4 0.21 !.d. 24 Al lU Pn-26 Boslt I E Fary 1.9 4.7 L6 0.16 B.d. 0.06 4l 34 rtt 0.66 0.41 o.29 0.02 0.14 ud. 2.2 7,8 0.70 0.07 !.d. rd. 2t26t634' lnl0s&!t3136 2N t45 PE2-33 OIX Bs. I E Fery r.6 4.3 12 0,61 0.6 t.t2 528 98 8r 2.6 lt 0.73 0.29 od. od" l8:t V&-Z P'&l-t OFX B&e OPx Be 3 I Digby N. Digby N. 1.3 r.2 t.2 0.t3 0.15 0.14 396 107 t3l t.3 t.4 4.2 0.14 0.34 0.15 124 93 50 l.l 1.6 0.53 0.0r o.& 0.04 yt PA-l? Bslt Z Digbl N, L2 0.55 t.4 0.t3 ad. 0.12 1.6 4.6 0.36 0.23 t.d. n.4 L4 9.0 0.93 0.15 a.d. ad. 214 w BCI Bsslt r Fln Is MBt4 Beelt L6 4.4 4.4 0.21 0.16 0.o2 L6 5.6 8.6 0.31 ud. 0.03 80 243 79?3263656 16l t4 3 McKayE 3.1 1.9 0.58 0.ll ud. ud" 215 1,2 6.0 3,3 0.26 Ld. 0.04 36 CD7 CIi-2 OPX Bts. Basolt | Z Capd'& 6imps 3.2 13 7.4 0,88 0.92 2.5 t.1 2,5 0.08 rd. 142 380 t01 l0l l7t 58 PegmadE= mafic pegmatite,Rhyolite = riyolite bandin mafic pegmatite. Sampleswith the samesamplenumberas a pegmatite (e.9. EF-l) are from that p€matite. Basalt = massivebasaltfrom the bonom of thick flows or baslt from the tops of flows. vesicular = vesicularbasaltftom layea in the thick flows. OPX Bas. - basaltbearingan accumulationoforthopyroxene. Units numbersare as follows: I = lower unit thick flov, 2 = Middle unit thin flows, 3 = Upper unit thick flow. = -see = Freeport, !. Ferry F:st Ferry, increasingsamplenumben indicae increasingdepthsin the upper 40 m of the flow. Frcepofi increasingsamplenumbec indicaE decreasingdepthsin Oe upper - 40 rn of the flow, McKay H. = McKay Head, Greenoughand DosEl (1992b)for depthsin the flow. The preciselocationsofP82 samplesappea.rin Fapeziker al. (1988). Digby tt. = Drgby Neck Preciselocationsofsamples ECl, MHl4, aad CD7 arc given in Grecnoughet al. (1989). The stratigraphiclocationofsample CH-2 is given in Greanoughand Papezik(1987). Copper,Ni and Cr concentralionsin Ppm. Noble meEl conceofatio0sin ppb. Mgf, = M/(MC + 0.9 Fe) atomic. N.d. = not detected. 156 THE CANADIAN MINERALOGIST northeast,and tho mineral disappearsin the Parrsboro area(Greenoughet al. 1989).Layering in the upper 50 m or so of the thick flows was describedin detail by Greenough& Dostal (1992a,b). ln summary,the lowest layers consist of approximately25-cm-thick mafic pegmatitethat containsa thin (2 cm) "rhyolite" band in the middle. The layering graduallychangesin charactertoward the top of the flows; the mafic layers becomehighly vesicularandnonpegsaatitic,andlayering disappearsentirely approximately5 m below the flow top. Both the pegmatiticand vesicularlayers are horizontally traceablefor tens of metersand are vertically separatedfrom one another by approximately l30-cm-thick sectionsof basalt.All basaltsexperiencedzeolite-faciesmetamorphism(Aumento 1966), but the Sr isotopicstudyof Jones& Mossman(1988) demonstratesthat carefully selectedsamplesgive an accuraterepresentationof the original geochemistry. Similarities between the layering in the North Mountain Basalt Formation and in the Hawaiian lava lakes (Helz 1987, Helz et al. L989) suggestthat the layers of mafic pegmatiteand vesicularbasalt represent segregationsof liquid siphonedinto thermally induced horizontal cracks formed in tJte downwardgrowing, semirigiduppercrystallinemush(Greenough & Dostal I992b). Thus the layers form from residual interstitial liquid derivedfrom the crystalmush. The "rhyolite" bands consist of a brownish, vesicular aphaniticrock andis called o'rhyolite"becauseit containsin excessof 70VoSiO2.The bandsappeared during a second stage of formation of horizontal cracks,with interstitial liquid in the mafic pegmatites forced into the crack by effervescenceand filter pressing (Greenough& Dostal L992b).In the mafic pegmatites,interstitial, polygonal patchesof microcrystalline Si-rich minerals (chemically equivalentto and traceableinto the 'thyolites'l) and similar areasof glass-like Fe-rich stilpnomelane(enveloping skeletal Fe-Ti oxides) havebulk chemicalcompositions,relative proportionsand total modal percentagessuggesting that they once were Si-rich and Fe-rich glasses, respectively(Greenough& Dostal 1992b). Samplesused in the study come from the three lower recogn:izedstratigraphicunits (Table I, Fig. 1). They have been divided into five rock groups: (OPX) basalts,mafic pegorthopyroxene-enriched matites, "rhyolites'0,vesicular-layerbasalts,and basalts.The single sampleof vesicular-layerbasalt reflectsthe dfficulty in obtainingan unalteredsample. 5amFlescategorizedas basalt originate from the tops of the thick flows, from thin flows, or from the massive (noncumulate)lower portions of some thick flows. ANALYTICAL METHODS Samplesnot usedin previousstudies(samplenumbers prefixed with EF and FP) were powderedand analyzedfor their major elementsusing atomic absorptionspectrometry.Thesedata can be obtained from tle Depository of UnpublishedData, Canada Institute for Scientific and Technical Information" National ResearchCouncil, Ottawa,OntarioKlA 0S2, or by writing J.D.G. Concentrationsof Rb, Sr, Sc, V, Cr, Ni, Cu, Zn and Ga were determinedusing X-ray pellets,with precision fluorescenceon pressed-powder and accuracybetter than t5%ofor all elementsexcept Cu (XIUVo).All other traceelements,exceptthe noble metals, were determinedby inductively coupled plasma- massspectrometryGCP-MS) following the sodium peroxide sinter procedwe of.I-nngeich et al. (1990). hecision and accuracyare better than4Vofor theseelements. Concentationsof the noble metals(Au, Pd, ft, Rh, Ru, and Ir, Table 1) were determinedby ICP-MS on L\Vo aquaregia solutionsafter preconcentrationof 10 g of rock powder by NiS fire assay.Typical reagent-blankconcentrationsarc 0.77, 0.59, 0.57, 0.06,0.05and0.04ppb for Au, Pd, Pt, Rh, Ru andh, respectively.Run-specific blank concentrationswere subtractedfrom the raw data. Average "detection limits" (3o abovebackground)are equivalentto 0.20, 0.15,0.08,0.01,0.10and0.02ppb (in a rock sample), respectively.Jacksonet al. (1990) shouldbe consulted for details of the procedure.Replicateanalysesof the reference material SARM-7 have given (in ppb) 245 (Au), L4sz (Pd),2n (Rh),3923 @),380 (Ru) atd77 (b), with relative standarddeviationsof 7Voor betterfor all elements.All PGE valuesarewithin 107o of the certified concentrations.Replicate analysesof sampleMCKI indicate precision of. lU%oor better for Au, Pd andPt and 307ofor Rh. Data on MCK 10 suggesta precisionof l50vo for h concentrationsnearthe detectionlimit (0.02 ppb). Al1 analytical work describedabovewas doneat The Memorial University of Newfoundland. Analytical methods and major- and trace-element data (other than for the PGE and Au) for all samples that do not begin with EF or FP appearin Greenough & Papezik (1987), Papezik et al. (7988), and Greenoughet al. (1989). Rrsurrs Variation diagrams plotted in terms of Mg# [= Mg/(Mg + 0.9 Fe) atomic] illustrate the geochemical relationshipsamong rock types (Fig. 2); averagecompositionsfor eachrock type appearin Table 2. The OPX-enrichedbasaltstend to have the highestMg# and Cr values,as well as the lowest concentrationsof alkali elements,high field-strength elementsQIFSE)andrare-earthelements(REE)of the entire suite. The basaltsand mafic pegmatites,taken togetherwith the OPX-enrichedbasalts,define correlation lines with concentrationsof alkali elements, HFSE and REZ increasingand that of "compatible" t57 BEHAVIOR OF THE PGE, NORT}I MOUNTAIN BASALT 2.na t ttt Sl O2 ' 2.OO r.5OTi oa AR o 60 r.oo tltlg.g Fi' ii:J'di^.. ltt e.50 KzO2l . o o Yv ara -f{lerr_ r.oo c.50 , i, vL r60 r40 t20 roo .o xJ '!f{^^ zoo r80 atr ttl , 'r o o t ^ -^!D^r. tf lrt o.30 o.25 o.20 o.,o trou o.ro 30 25 Lo ao 80 400 350 300 250 200 r50 roo o.50 20 r5 40 v Y l' tr '.tlt**. o{tf,a^... Nb 20 lo a 500 300 tt 400 olo a 300 tr 200 A roo rlrr 0.20 (-^ . - i .o - r rt li . ?a erl . trooa ^^a. la .rt 250 200 r50 Cu roo 0.40 Mg* Mg* A BASALT I VESICULAF BASALT '. OPX BASALT Orsarrc PEGMATTTE Y RHYOLITE Ftc. 2. Variation diagrams showing selected major element (wt. 7o oxides, volatile free) and trace elements (ppm) plotted against Mg# [Mg# = Mg/(Mg + 0.9 Fe) atomic]. OPX basalt: orthopyroxene-enrichedbasalt; see text for usageof "rhyolite". elementsdecreasingas Mg# decreases. The single vesicular-layer basaltis distinctivein havingthe highest TiO2 and P2O,concantrations observed,and,REE concentrationsapproachingthose of the "rhyolites". The "rhyolite" samplesare characterizedby their high SiO2,K2O, Zr andLa concentrations, and low TiO2 andSr. Copperconcentrations tendto increaseasMg# decreases in the mafic rocks.but thereis considerable scatter; individual samples of "rhyolite" contain the highest and amongst the lowest Cu values. The distinction among rock types is not as apparent in the noble metal data (c/ Figs. 2,3), partly reflecting the lower precision at low concentrations (Table l). N e v e r t h e l e s s ,m e a n s f o r m a n y e l e m e n t s ( 2 3 o f 30 pairs tested) between rock types are statistically different at the 0.10 level of significance or better r58 THE CANADIAN MINERALOGIST a a 4.2 a al Au ' Oc) .- 'r aO oA o t. -'t$^^ 2,25 a lt . qLoo- aV .ttv -l_t I2 I ^t aA a A _ l2 ^oo 80 70 , to a - 4 60 50 Rh o 6 40 30 20 A ^^l ^-it *{.}^rl^t' ara aGe^roo --L!l .^ lo o..990 0.8 .ao o.7 or..f6 0 R u o) . 6 0 t.2 0 o. o. o.40 ,.40 o.20 ,,20 opd a a t4 2 4 o o ,Y.r o.30 ,^), o.ao a o o. A vlLr o.70 0.20 Mgs o. iAll" 0.40 0.60 Mg? A BASALT V RHYOLITE I O uarrc pEGMArrrE OoPx BASALT vEsrcuLAR BASALT FIc. 3. Variation diagrams showing the noble metals (ppb) plotted against Mg#. OPX basalt: orthopyroxene-enrichedbasalt; see text for usageof "rhyolite". (Table 2, footnote). Gold (correlation coefficient, r = *-0.69) and, to a much lesser extent, Pd (r = 4.34, level of significance of correlation = 0.10) seem to increase with decreasingMg# in the mafic rocks; the "rhyolite" samples contain the lowest concentrations of these elements (Fig. 3, Tables 1, 2). In contrast, levels of Pt, Rh, Ru and Ir tend to decreasewith decreasing Mg#. The concentrations of Ru and Ir commonly fall below detection limits except in the OPX-enriched basalts. Average mantle-normalized patterns indicate substantial fractionation of the noble metals after extrusion (Fig. 4, averagedvalues given in Table 2). T\e pattem representativeofbasalt is straight, with a slight negative Ir anomaly. On average, the OPX-enriched basalts show higher concentrationsof Ni, Ir, Rh and Pt than the basalts,but the pattern flattens out between Pt and Au. The pattern for mafic pegmatite displays the opposite relationships, with average concentrations of Ni to Pt below, and Pd and Au above, those in the basalt. The vesicular-layerbasalt sample displays similarities to both the basalts and mafic pegmatites. The pattern representingthe average "rhyolite" shows a positive Ir anomaly; concentrationsof the other noble metals fall below those in the other rock tvpes. DlscusstoN I nterp retation of maj or- elementand trace-elementdata Although the samples come from several flows and outcrops up to 200 km apart, the coherency of major- and trace-elementdata (Fig. 2) point to 1) similar initial composition of the mafic magma, 2) similar igneous geochemical processesat all localities, and 3) minimal effects of alteration. The first point is emphasized by the incompatible trace elements because small initial variations in concenfiation tend to be enhancedduring differentiation. The wide range of values of Mg# (0.70 - 0.20) and SiO2 contents (53 - 75 wt.%o,volatile-free basis) emphasize the extent ofdifferentiation. The processesresponsiblefor differentiation are equivocal, as reviewed below. 159 BEHAVIOR OF TI{E PCE. NORTH MOUNTAIN BASALT Mass-balancecalculations based on major elements indicate that the mafic pegmatites can, to a first approximation, be derived by crystal-liquid differentiation from the basaltic melt (Papezik et al. 1988). However. similar calculations show that the most TABLE2. AVERAGE COMPOSITIONOFROCK 1YPES IN THE NORTE MOT'NTAIN BASALT oPx Baslt Mlnc Pmtlt€ Vgiohr B&lt Riyoltte s5 ALOr Fqo, MnO MgO CaO Na?O KrO PrOr Mei 13.13(0.55) 53.15(0.J4) 54.E6(1.68) 53.66 1.27(0.13) 0,920.r2) 1.78(0.16) 2.& 14.31(0.38) 12.754.45) 12.58(0.80) 11.27(0.50) 9.96(0./, 13.78(0.63) t4.53 0.17(0.02) 0.16(0.02) 0.21(0.05) 0.29 6.47(0.66) 9.52(t.U2) 4.E20.40) 5.14 9.66(0.54) 10.91(0.54) 7 . 9 10 . l J ) 6.06 2.64(0.58) 1.92(0.13) 2.63(0.t7) x.27 0.93(0.29) 0.63(0.08) 1.20(0.28) l.E4 0.15(0.02) 0.09(0.03) 0.22(0.05) 0.31 0.56 0.43 0.68 o.44 7t.32Q.47) 0.68 (0.15) 10.55(0.53) 7.37 (1.42) 0.11 (0.05) 1.73 (0.68) 2.20 p.m 2.92 (0.20) 2.98 (0.40) 0.14 (0.04) o.34 Alkall ed alkalle @trh Delal9 Rb 27 (3' l9 (4) Cs 1.2 (0.3) 0.s8(0.20) sr 190 (ll) 151 (8) Ba 213 (38) t74 (t9) l8l 08) u2 (€r) Hlgh Fleld Strengll! Eloeft rh 3.12 (0.39) u 0.88 (0.34) z' 107 (r8) lrf 3.05 (o.sa (2) Nb 1r Ta 2.1 (l.E) Y 26.3 (4.t 2.32(0.23\ 0.48(0.07) 83 (!3) 2.aQ3v 8 (2) 1.0 (0.2) 2X.4 (X.4) 3.990.01) 5.96 l.l0 r34 Q5) 255 3.66(0.E2) 6.32 l5 (3) t9 3.1 cJ) 1.6 30.7 (6.8) 41.0 ll.4 (1.2) 23.2 (3.2) 18.6 (3.9) 26.1 40.0 Q.V 61.0 5.470.04) 2.6 (4.2) 5.42(0.88) 7.?E 1.49(0.1n 2.gj 5.39(1.00) 0.93(0.16) t.22 6.21(1.14) L26 (0.23) 3.59(0.68) 0.5r (0.10) 3.18(0.54) 0.75(0.9? 0.61 32.9 Q.r) 69.9 (4.V 8.62 (0.40) 33.0 (1.9) 8.22 (0.65) l.9r (0.50) 7.04 (0.69) 1.42 (0.10) 8.67 (O,72, 1.76 (0.It 5.25 (0.43) 0.76 (0.0o 4.59 (0.56) 0.67 (0.10) 38.5 (6.4) 35J (53) lo C0 67 0l) 15 (10) 3 l5l (4n 74 (9) 19 (2) E.97 (3.78) 28 Q4) (7) 6 70 (17) 2 39 (11) 49 101 (14) 163 298 t03 Q2) 602 (6t lt.3 0.4) 3r7 (69) 8.25 O.94) 32 (9) ll.5 o0.9) 43.0 (4.t chemicallyevolvedmafic rocks,the vesicularlayers, showa poorfit; problemsariseowingto their elevated concentrationsof alkalis (Greenough& Dostal 1992b).The calculatedrate of upper-crustgrowth on the flow indicatesthat the first of thesevesicular layers(depth-5 m) formedafteronly about6 montis & Dostal1992a).The vesicular ofcooling (Greenough natureof theselayers and the fact that they formed earlyled Greenough& Dostal(1992b)to proposethat arein part a resultofthe their chemicalcharacteristics early upwardtransportof elementsin vesicleplumes eitherasvolatilecomplexesdissolvedin the magmaor in a separatefluid phase.Vesicleplumesare commonly preservednear the baseof mafic lava flows, includingtheNorthMountainbasaltflows. Mass-balancecalculationsinvolving primary phasessuggestthat crystalfractionationis incapable of producingthe "rhyolite" (Greenough& Dostal 1992a),which may thus resultfrom a Si-rich immisAlong cible liquid derivedfrom the mafic pegmatites. with petrographicdata alreadydiscussed,the mafic similarto rocks pegmatites havebulk compositions known to produceimmiscibleliquids, and the "rhyolite" resemblesknown Si-rich glassproducedby immiscibility(Philpotts1982,Greenough& Dostal Th, Zr, Hf andBa behavedanti 1992b).Nevertheless, experimentsprethetically to what synthetic-system Rare Earii El@enls Pr Nd Sm Eu cd Tb Dy HO Tm Yb Lu r3.9 (1.t 29.7 (3.2) 3.3E rJ.3 (2.0 3.75(0.36) 1.20(0.12) 3.35 0.7r (0.0E) 3.86 o.n 2.28 0.3r 2.32(0.2?) 0.36(0.04) Sldercphlle Sc v cr co Ni Mo cu zn ca and Chalcophlle 41.0 (4.3) 284 08). 122 Q4) s4 (3) 19 07) 6 (3) (50) llt 73 (8) (3) l8 Noble Mebk Au Pd Pr Rh Ru k 2.6 (0.8) 5.1 (2.7) 4.7 (1.8) 0.23 (0.09) 0.17 0.09 (0.08) rt.2 (t.2) 2.88(0.32) 0.90(0.09) 0.58(0.07) 1.96(0.23) 0.31(0.04) Elemonts 46.6 (4.8) 246 (t9t 399 (69) 63 (9) 85 (25) 3Q, tO6 (42) 66 (10) t5 (2) I.8 (0.7) 5.6 (4.5) 8.r (4.7) 0.4?(0.29) 0.46(0.27) 0.4s(0.32) 33.0 y5 3 52 6 20 3.6 (1.3) 9.r {3.4) 1.4 (1.0) 0.16(0.o7) t.2 6.0 3.3 0.26 0.08(0.02) 0.04 l9l (108) 50 (13) t8 6) 0.85 (0.42) 1.0 (t.4) 0.50 (0.30) 0.06 (0.07) 0.16 0.36 (0.4r) l. Maj6 elemeniorid$ in w. 9o !@lculated b 100% volarile fte wilh bhl Fe I F.iq. T@ elemoB in ppm scept noble mehls in ppd ercepl noble meuls in mb. The sdard deviarim is given in b@ke6. N.d, = oot der*red, A bldk = not demined or in lhe @ of &e sbdard daiadm rhm w s2 mpl6. M8, = Mg/(Mg r 0.9 Fe) domic). 2. No, ot spls is $e nunb6 analrzed for the nqjd oladsb, md of the @ clendts and all Oe noble neEls. B@@ a fd t@ elen6B y@ not demined in all wplq n will be 16 for !h* eled6B. A fev l@ el@b (e.9. Cr ed Ni) dd the noble mesls (lable t) v@ not d@Ed i! ema spls th8 the ma b b€sedon fffi d& poinb. 3. Se Tablc ! for cL lyps dginidonr. 4. Udng dE sd4t !8, t@! lor Au, Pd, P|, Rb ad lr i! @h @k $oup (qepr dE r*iobr.latd bsslt) 6 dlf@t fton @ eo$s d tlb 0. l0 t*cl of siglifi@ with tha folloeing 7 q€ptiom od of 30 t&d: the m@s for lr bery@ baglt ed naic pegnatiE, bqelt dd 'rhyoli@', dd mfic Fg@tib ed 'rhl,ollc'i tte n@3 for pd bery@ basi! and OPx€riched baelt, od mafic peg@ib dd OPx$rtchid bastli tha n@s tor Au begq ba3di dd OPx€nrich€d bsslt. lr.l J F z trJ (, t u o.l Y o o.ol E o.001 Ni lr A BASALT Rh Pt Pd Au Cu V RHYOLITE I V E S I C U L A RB A S A L T PEcMATTTE aoPx luarrc BASALT Ftc. 4. Average mantle-normalized concentrations of the noble metals in each of the major rock types in the North Mountain Basalt sequence. Note that the pattem for the vesicular-layer basalt is represented by a single sample. The significance of differences between data points can be evaluated with the standard deviation and T-test data in Table 2. The values used for normalization are from Barnes et a/. ( 1988). OPX basalt: onhopyroxene-enriched basalt; see text for usage of "rhyolite". 160 THE CANADIAN MINERALOGIST dict for silicate-silicateliquid immiscibiliry (Watson 1976,Ryerson& Hess 1978);althoughmuchof the evidencesupportsimmiscibility, it is not conclusive (Greenough & Dostal 1992a). Papeziket al. (1988) concludedon rhe basisof major-element mass-balance calculationsand traceelementmodelingthat the evolvedmafic rocks in the Digby areacan be formed from normal basaltsby the fractionationof mineralspresentin the orthopyroxene-enriched basalts. an importantprocessby itself. Olivine doesnot occur in the North MountainBasalt.Chromitefractionation can help explainthe high Cr valuesin the OPXenrichedbasalts(Fig. 2), but chromiteis not an important phasein this suite. Furthermore,major-element massbalancecoupledwith trace-element modeling showthat augiteremovalalonecan explainthe variation in Cr (Papeziket al. 1988).Anotherproblemwith callingon chromitefractionationis that the level of Pt alsois higher,on average,in the OPX-enriched basalts and lower in the mafic pegmatitesthan in the basalts Thedistribution of noblemetalsin mafic rocks (Table2). Normally,Pt follows the low-melting-point PGEalongwith Au in evolvingmagmas(Barneser a/. Averagechondrite-normalizedpatternsfor the 1988,Greenough & Owen1992). noble metalsin basalts,mafic pegmatitesand OPXThe resultsobtainedfor the North MountainBasalt enrichedbasalts(Fig. 4, Table 2) showthat the high- sequence aresimilarto the resultsof studiesof ft and melting-pointnoblemetals(Ir, Ru andRh), as well as Pd in sheetsof Mesozoichigh-Ti quartz-normative Pt, occurat higherconcentrations in the OPX-enriched t h o l e i i t i c d i a b a s ei n t h e e a s t e r nU n i t e d S t a t e s basaltsthan in the basalts.Pd and Au concentrations (Gottfried& Froelich1988,Talkingtonet al. 1988, are similar or lower in the OPX-enrichedbasaltsthan Gottfriedet al. 1990).Chemicalsimilaritiesbetween in the basalts.Relationshipsbetweenthe mafic peg- the North MountainBasaltandtheserockshavebeen matitesand basaltsare the reverseof thosebetween emphasized by Puffer & Hurtubise(1982)and Dostal the OPX-enrichedbasaltsand basalts.Although the & Greenough(1992).The PGE studiesindicatethat mafic pegmatiteand OPX-enriched basaltpattemsare orthopyroxenecumulatesin the Mesozoicdiabase complementary, crystallizationof OPX-enriched sheetsshowPt enrichmentandPd depletionrelativeto basalt is probably not directly related to mafic undifferentiated diabase.In contrast,ferrogabbrohighpegmatiteformation becauseOPX-enrichedbasalts er in the sheetsandlocally "pegmatitic"in texturecan do not occur in the northeast(e.g.,McKay Head). havePd enrichedup to l0 times(165ppb) that in the Mechanismsproposedto explainchangesin the con- undifferentiated canbe basalt.Platinumconcentrations centrationof the noble metalsin evolvingmagmas higher (up to 2.5 times) or even lower than in the werereviewedby Crocket(1981),Naldrett(1981), contactbasalt.The Pd enrichmentsare accompanied Barnese, al. (1985,1988),Briigmannet al. (1987), by elevatedAu and Cl concentrations. TheseenrichGreenough & Fryer(1990)andNaldrettet al. (1990). mentsdo not appearrelatedto thicknessof the diabase Nucleationandsegregation of an immisciblesulfide sheets.Similar enrichments(150-200ppb Pd) are liquid (Hertogenet al. 1980,Hamlyn et al. 1985) observedin the York Havendiabasesheet(-550 m: shouldeffectivelyremoveall the noblemetalsduring Gottfried et al. 1990)and Palisadessill (-300 m: magmaevolution;measuredsulfide liquid - mafic Tafkingtonet al. 1988),whereasthe marginally magmapartition coefficients (K) zre on the order of thinnerNorth MountainBasaltflows (-200 m) show ld to 105(Peachet al. 1990,Sioneer al. 1990).Pd maximumPd concentrations of <16 ppb. Maximum and Au concentrationstend to increasebetweenthc Pd enrichmentoccursin narrow zones;thus there is basaltsand mafic pegmatites(bulk Ka < l), arguing t h e p o s s i b i l i t yt h a t o u r s a m p l i n go f t h e N o r t h againstremoval of sulfide liquid as a significant MountainBasaltmissedthe most stronglyenriched process.Similarly, Cu is higher in the mafic peg- layers. matitesandslightly lower in the OPX-enriched basalts Authorsof the paperson the easternU.S. rocks thanin thebasalts(Table2,Fig.4). Thusthe behavior providedno mechanismfor the Pt enrichmentother of Cu indicatesthat not more than 0,47oof fraction- than to associateit with orthopyroxene accumulation. ating phases(for a bulk D < 1) could be sulfide The enrichmentin Pd and Au in the ferrogabbrowas becausethe Cu sulfideliquid - silicateliquid partition attributedto movementby "late- to post-magmatic coefficient is high (-250, Rajamaniand Naldretr fluids and volatilesenrichedin Cl and S" (Gottfried 1978). et al. 1990,Gottfried& Froelich1988). The fractionationof oxidephases(Crocket& Chyi The degreeof Pd enrichmentand Pd/Pt fraction1972, Agiorgitis & Wolf 1978),PcE-bearingalloys ation is not as extremein the North MountainBasalt ( B i r d & B a s s e t 1 9 8 0 , K e a y s 1 9 8 2 )o r o l i v i n e a s i n s o m eo f t h e e a s t e r nU . S . d i a b a s es h e e t s . (Briigmannet al. 1987)hasbeenproposedas a means Nevertheless" the abovemodelfor Pd and Au enrichof removingthe high-melting-pointPGE (i.e., Ru, Ir ment fits well with data availablefor the North and Os) from evolvingmafic magmas.The formation MountainBasalt,exceptthat usingthe basalts,it is of PGE alloys appearsassociated with chromitepre- possibleto show that the 'olate-stage" enrichment (e.9., cipitation Briigmannet al. 1987)andmay nor be probably occurred early after emplacement BEHAVIOR OF T}IE PGE" NORTH MOT'NTAIN BASALT t61 (Greenough& Dostal 1992b).As noted above,rising Thedistribution of the noblemetals plumesof volatilesprobablyalteredthe major-element inthe "rhyolite" bands geochemistry of vesicle-richlayers and trace-element Regardlessof whetherthe "rhyolite" bandsformed and may have had a similar though less extensive effect on the layers of mafic pegmatite(Greenough& by crystalfractionationor silicateliquid immiscibility, Dostal 1992a,b). Throughcomparisonwith Hawaiian Figure4 showsthat the average"rhyolitic" liquid was lava lakes, we know that theselayers formed very stronglydepletedin most of the PGE and Au, but enrichedin Ir. Figure3 illustratesthat the Ir anomaly earlyin thecoolinghistoryof thelavaflow. Although there is disagreementon whether only occursin a few "rhyolite" samples,whereit is The high variabilitycould be relatedto a hydrothermalmovementof the PGEdirectlyresultsin pronounced. ore deposits,mostauthorswould agreethat suchsolu- samplingproblem causedby a "nugget effect". of Cu are similarly variablein these tions are capableof at least local movementof the Concentrations metals(Barnes& Campbell1988,Wood et al. 19921. rocks,but thereis no correlationbetweenCu and lr, As in the easternU.S. diabasesheets(Belkin 1989;, and thus no indicationthat the Ir residesin a sulfide studiesof fluid inclusionsand halogen-bearing phase.Analogousrocks termed"granophyres"from mineralsin intrusionssuch as the Bushveldand easternU.S.diabasesheetsalsocontainextremelylow Stillwatercomplexessuggestthat the PGE can be levelsof Pd andPt (Gottfried& Froelich1988). moved by Cl-rich fluids (Boudreatret al. 1986, ACKNOWLEDGEMET.{TS Ballhaus& Stumpfl 1986,Boudreau1988,Hamey & Merkle 1990).Thermodynamic calculationspredict Dr. J. Hodych participatedin the sampling.Dr. significant solubility of Pd in Cl-rich magmatichydrothermalsolutions(Sassani& Shock 1990). H. Longerich,Dr. S. Jackson,Ms. G. Andrewsand Preliminaryexperimentalresultssuggestevenhigher Mr. W. Gosseacquiredthe analyticaldata.Diagrams solubilitiesof Pd than predictedfrom calculations weredraftedby Ms. M. Halko. Ms. M. Wesnoskiand (Hsuet al. l99l). Althoughthe theoreticalandexperi- Ms. C. Neid preparedthe tables.The constructive mental work supportsthe possibility of moving the criticism of the reviewersand helpful commentsof PGEwith Cl-rich fluids,it cannotasyet quantitatively R.F. Martin and AssociateEditor M. Zientek are explainfractionationof the PGE or assessthe impor- greatly appreciated.This researchwas supportedby tanceand effect of other species(e.g.,S, NH3 and individual operatinggrantsto the authorsfrom the Councilof andEngineeringResearch NaturalSciences othercations)on PGEsolubility. Volatile complexingprovidesa tentativeexplana- Canada. tion for Pd and Au enrichment in the mafic REFERENcES pegmatites, but it doeslittle to explainPt enrichment (relative to basalt)in the averageOPX-enrichedbasalt (Table 2, Fig. 4). One possibleexplanationis that ActoRcrls, G. & Wolp, R. (1978):Aspectsof osmium, rutheniumandiridium contentsin someGreekchromites. orthopyroxenehas an affinity for Pt and the highChem.Geol. ?3, 267-272. melting-pointPGE,muchas olivine seemsto havean affinity for the high-melting-pointPGE alone (Briigmannet al. 1987).Unfortunately,it hasnot been AuMENro,F. (1966):Zeolite minerals,Nova Scotia./tt Geologyof Partsof the Atlantic hovinces. Geol.Assoc. demonstratedthat any silicateshavean affinity for the Can.- Mineral.Assoc.Can.,Guidebook,Tl'78. PGE that cannotbe explainedby microinclusionsof sulfides,chromiteor PGEalloys.Studiesof ultramafic Bellrnus, C.G. & Sruupn, E.F. (1986):Sulfide and platrocks show that olivine has relatively low conceninum mineralizationin the Merenskyreef: evidencefrom t r a t i o n so f t h e P G E ( M i t c h e l l & K e a y s l 9 8 l ) . hydroussilicatesand fluid inclusions.Contrib.Mineral. Incorporationof the PGE in microinclusionsin earlyPetrol. 94, 193-204. could formedphasessuchas olivine or orthopyroxene aresomewhaterratic Beruss,S.-J.,BoYD,R., KonNeltussrN, explainwhy PGEconcentrations A., Ntrsson,L.-P., in the OPX-enrichedbasalts.The numberand size of R.B. & RostNs,B. (1988):The use Or.ren,M., PEDERsEN, of mantle normalizationand metal ratios in discrimiinclusionsand hencethe concentrationsof the PGE nating betweenthe effects of partial melting, crystal would be dependent on nucleationkineticsandgrowth fractionationand sulphidesegregationon platinum group ratesof both the inclusionsand host phase.If the elements,gold, nickel and copper: examplesfrom inclusionswere globulesof immisciblesulfide, then Norway.1z Geo-Platinum87 (H.M. Prichard,P.J.Potts, fractionationof Pd and Au might be attributedto J.F.W.Bowles& S.J.Cribb, eds.).Elsevier,New York, scavenging of theseelementsby volatile constituents, N.Y.013-141). which preventedtheir incorporationin the sulfides. We know of no data indicatins the effect of volatiles NeLonr,rr,A.J. & Gonror, M.P. (1985):The on sulfidetiquid - silicateliquid partitioncoefficients origin of the fractionation of platinum group elements in for the PGE. terresrial magmas. Chern. GeoI. 53, 303-323. 162 THE CANADIAN MINERALOGIST BARNES, S.J.& C.cNapsrLU I.H. (1988):Role of laremagmatic fluids in Merensky-typeplatinum deposits:a discussion. Geology16, 488491. RArr,N. & Anuscavecp,P.J.(1990): Fractionationof palladiumand platinumin a Mesozoic diabasesheet,Gettysburgbasin,Pennsylvania: implicafions for mineral exploralion.J. Geocltem.Explor. 37, 75-89. Br,lrtN, H.E. (1989):Fluid inclusionandmineralogicsrudies of a palladium-platinumanomalyzone in the Reeser's Summitdiabase,Pennsylvania: evidencefor hydro- GREENoUGH, J.D. & DosrAL, I. (1992a): Layered rhyolite thermal transport..Ir U.S.G.S.Researchon Mineral bands in a thick North Mountain Basalt flow: the - 1989(K.S. Schindler,ed). U.S.Geol.Surv., Resources products of silicate liquid immiscibility? Mineral. Mag. Circ. 1035,4-6 (abstr.). 56. 309-3I 8. BIno, J.M. & Basssrr, W.A. (1980):Evidenceof a deep mantlehistory in terestrial osmium- iridium - rutheniumalloys.J. Geophys. Res.85, 546I -5470. BoUDREAU, A.E. (1988):Investigations of the Stillwarer Complex.IV. The role of volatilesin the petrogenesis of the J-M reef,Minneapolisadit section.Can.Mineral.26, I 93-208. MArHEz,E.A. & McCelluu, I.S.(1986):Halogen geochemistry of the Stillwaterand BushveldComplexes: evidencefor transportof the platinum-group elementsby Cl-richfluids.J. Petrol.27,967 -986. BoweN,N.L. (1917):Discussion of "Magmaticdifferentiation in effusive rocks". Trons. Am. Inst. Mininx Engineers54,455-457. B n U c u e N NG , . E . , A R N D TN , . T . , H o n u e N N ,A . W . & ToBSCHALL, H.J. (1987):Noblemetalabundances in komatiite suitesfrom Alexo, Ontario, and Gorgona Island,Columbia.Geocltim.Cosmochim. Acta 51,21592t69. CRocKEr, J.H. (1981):Geochemistry of the platinumgroup elements.In PlatinumGroup E,lements: Mineralogy, Geology,Recovery(L.J. Cabri, ed.). Can. Inst. Mining Metall.,Spec.Vol. 23,47-64. .- & CHyl,L.L. ( I 972):Abundances of Pd,Ir, Os and Au in an alpine ultramaficpluton. 24th Int. Geol. CongressL0,202-209. DosrAL,J. & Dupuy, C. (1984):Geochemistry of rhe Nonh Mountainbasalts(NovaScotia,Canada).Chem.Geol.45, 245-261. &- (1992b):Coolinghistoryanddifferen- tiation of a thick North Mountain Basalt flow (Nova Scotia, Canada).Bull. Volcanol. 55, 63-73. & Fnvrn, B.J. (1990):Distributionof Au, Pd, ft, Rh, Ru and Ir in ODP Leg I 15 (IndianOcean)hot-spot basalts:implicationsfor magmaticprocesses. In Proc. OceanDrilling Program,ScientificResults,vol. 115 (R.A. Duncan,J. Backman,L.C. Petersonel c/., eds.). OceanDrilling Program,CollegeStation,Texas(71-84). , J o N e s ,L . M . & M o s s M A N ,D . J . ( 1 9 8 9 ) : Petrochemicaland stratigraphicaspectsof North Mountainbasaltsfrom the north shoreof the Bay of Fundy,Nova Scotia,Canada.Can.J. Earth Sci.26, 2710-27t7. & OwrN, J.V. (1992): Platinumgroup elemenlgeochemistry of continental tholeiites: analysis of the Long Range Dyke swarm, Newfoundland, Canada. Chem. Geol. 98, 203-219. & PApEzrK,V. S. (1987): The Petrology of North Mountain basalts from the wildcat oil well Mobil Gulf Chinampas N-37, Bay of Fundy, Canada. Can. J. Eanh Sci. ?4, 1255-1260. HAMLvN, P.R., KEAys, R.R., C,qr'aenoN, W.E., Cnewpono, A.J. & Welonon, H.M. (1985): Preciousmetals in magnesian low-Ti lavas: implications for metallogenesis and sulfur saturation in primary magmas. Geochim. Cosmochim.Acta 49, 1797- l8l l. HARNEy,D.M.W. & MERKLE,R.K.W. (1990): Pt-Pd minerals from the upper zone of the eastern Bushveld Complex" South Africa. Can. Mineral. 28.619-628. & GnEeNoucH, J.D. (1992):Petrogenesis of the Early MesozoicNorth MountainBasaltsof Nova Scotia, Hez, R.T. (1987): Differentiation behavior of Kilauea Iki Canada. /n Eastern North American Mesozoic lava lake, Kilauea Volcano, Hawaii: an overview of past Magmatism(J.H. Pufler & P.C. Ragland,eds.).Geol. and current work. In Magmatic Processes, Soc.Am. Spec.Pap.268,149-159. Physicochemical Principles (8.O. Mysen, ed.). GeoFnypn,B,J.& GnmNoucH,J.D. (1992):Evidencefor manrle helerogeneityfrom platinum-group-element abundances in Indian Oceanbasalts.Can.J. Earth Sci.29.23292340. chemicctlSoc., Spec.Publ. L,241-258. K T R S C H E N B A UHM . ,& M A r N E N r o ,J . W . ( 1 9 8 9 ) : Diapiric transfer of melt in Kilauea Iki lava lake, Hawaii: a quick, efficient process of igneous differentiation. Geol. Soc. Am, Bull. L0L.578-594. Go'r'rFRrED, D. & FRoELtcH, A.J. (1988):Variationsof palladium and platinumcontentsand ratios in selectedEarly Mesozoictholeiitic rock associations in the Eastern HERrocEN,J., JANssENs,M.-J. & PALME,H. (1980): Trace UnitedStates.ft Studiesof the Early MesozoicBasinsof elemenls in ocean ridge basalt glasses: implications for the easternUnitedStates(A.J.Froelich& G.R.Robinson, fractionations during mantle evolution and petrogenesis. eds.).U.S.Geol.Sun.,8u11.7776,332-341. Geochim. Cosmochim.Acta M. 2125-2143. BEHAVIOROF THE PGE,NORTHMOUNTAINBASALT H o D y c H , J . P . & D u N N r N c ,G . R . ( 1 9 9 2 ) : D i d t h e impacttriggerend-of-Triassic massextincManicouagan uon''!Geology20,51-54. 163 of immiscibleliquids Purr-ports,A.R. (1982):Compositions in volcanicrocks.Contrib.Mineral.Petrol.80.201-218. Purrnn,J.H. & Hunrustsr,D.O. (1982):Early Jurassicbasalt oi Fundy Basin,Nova Scotia.Geol. Soc.Am., Abstr. H s u , L . C . , L E C H L E RP,. J . & N s r - s o N J, . H . ( 1 9 9 1 ) : Programs14,74. Hydrothermalsolubility of palladiumin chloride solutions from 300oto 700'C: preliminaryexperimental V. & Nallnprr, A.J. (1978):Partitioningof Fe, RAiAMANT, results.Econ.Geol.86,422-427. Co, Ni and Cu betweensulfideliquid and basalticmelts and the compositionof Ni-Cu sulfide deposits.Econ. S.E.,Fnvpn,8.J., GossE,W., Heelev, D.C., JAcKsoN, Geol.73,82-93. LoNoERrcH, H.P. & SrnoNc,D.F. (1990):Determination of the preciousmetalsin geologicalmaterialsby induc(ICP-MS) R y E R s o NF, . J . & H r s s , P . C . ( 1 9 7 8 ) :I m p l i c a t i o n so f tively coupledplasma- massspectrometry liquid-liquid distributioncoefficientsto mineral-liquid with nickel sulphidefire-assaycoliectionand tellurium Acta 42, 921-932. partitioning.Geochim.Cosmochim. coprecipitation. Chem.Geol.83,119-I 32. andsolubilD.C.& Snocr,E.L. (1990):Speciation D.J.(1988): The isotopiccomposi- SasseNt, JoNEs, L.M. & Mossnaeu, ity of palladiumin aqueousmagmatic-hydrothermai tion of strontiumof the Early JurassicNonh Mountain solutions.GeoIog1-18.925-928. basalts,NovaScotia.Can.J. Earth $ci.25,942-944. andiridiumin komatiites and Keevs,R.R.(1982):Palladium problems.lr associated rocks:applicarionto petrogenetic Komatiites(N.T. Arndt & E.G. Nisbet,eds.).Allen & Unwin.London,U.K. (435-457). S r o N E ,W . E . , C R o c K E rJ, . H . & F l e p r , M . E . ( 1 9 9 0 ) : Partitioningof palladium,iridium, platinum,and gold betweensulphideliquid and basaltmelt at 1200'C. Acta 54,2341-2344. Geochim.Cosmochim. G.A., FRyER, B.J.& JncrsoN,S.E. TalrrNcron. R., Gorrrntso, D., Ltpttt,8., Ratr, N., PurnEn, LoNcenrcu,H.P.,JENNER, (1990):Inductivelycoupledplasma- massspectromeffic J. & Sutnlsv,D. ( 1988):Thedistributionof theplatinumgroupelementsin the Palisades sill, New JeneyandNew analysisof geologicsamples:a critical evaluationbased (H.M. Prichard,P.J.Potts, York. /n Geo-Platinum'87 on casestudies. Chem.Geol.83,105-l18. J.F.W.Bowles& S.J.Cribb, eds.).Elsevier,New York, N.Y. (255-256). MrrcHELL,R.H. & KEAys,R.R. (1981):Abundanceand distribution of gold, palladiumand iridium in somespinel discrimiand garnetlherzolites:implicationsfor the natureand WARK,J.M. & Crenrr, D.B. (1980):Geochemical environmentof the North natorsand the palaeotectonic components in origin of preciousmetal-richintergranular Mountainbasalts,Nova Scotia.Ccr. J. Eanh Sci.17, the uppermantle.Geochim.Cosmochim. Acta 45,24251740-1745. 2442. elementdeposits.1n WArsoN,E.B. (1976):Two-liquid partition coefficienls: NelonErr, A.J. (1981):Platinum-group implications.Contrib. experimental dataandgeochemical Platinum-Group Elements:Mineralogy,Geology, M ineral. Petrol. 56, 119- 134. Recovery(L.J. Cabri, ed,.).Can. Inst. Mining Metall., Spec.Vol.23, 191-231. WETcAND, P.W. & Recr-eNo,P.C. (1970):Geochemistryof BRUcMANN, Mesozoicdoleritedikesfrom easternNorth America. G.E.& WrmoN,A.H. (1990):Models 4. Contrib.M ineral. Petrol. 29, 195-21 for the concentration of PGE in laveredintrusions.Can. Mineral.28,389-408. WHnE,R.S. (1989):Igneousoutburstsand massextinctions. Trans.Am. Geophys.Union70(46),1480,1490. P e p p z r r ,V . S . , G R E E N o U cJH. ,D . , C o L w E L LJ, . A . & MALLTNSoN, T.J. (1988):North Mountainbasaltfrom (1992):The Digby, Nova Scotia:modelslor a fissureeruptionfrom Wooo, S.A., MouNrAIN,B.W. & PAN,PUJINc aqueousgeochemistryof platinum,palladiumand gold: stratigraphyand petrochemislry.Can. J. Eanh Sci. 25, recentexperimentalconstraintsand a re-evaluationof 74-83. theoreticalpredictions.Ccn.M ineral. 30, 955-982. Peacu,C.L., Metrez, E.A. & KEAys,R.R. (1990):Sulfide melt - silicate melt distribution coefficientsfor noble metalsand other chalcophileelementsas deducedfrom MORB: implicationsfor partial melting. Geocltim. Receivedluly 14, 1993,revisedntanuscriptaccepted August3, 1994. Cosmochim.Actct54, 3379-3389.