American Mineralogist, Volume 74, pages 981-993, 1989

Two-phase nickeliferous monosulfide solid solution (mss) in megacrysts fromMount Shasta, California: A natural laboratory for nickel-copper sulfides

Wrr,r,ram E. Sroxn, Mrcnl.nr, E. Fr-nnr, NBrr, D. M,q,cRArDepartment of Geology, University of Western Ontario, London, Ontario N6A 5B7, Canada

Ansrntcr

Sulfide spherules in basalt and andesite tephra from a cinder cone on Mount Shasta,California, have been characterized in detail. The nickel-iron-copper sulfides occur inmicrometer-sized subspherical polymineralic aggregates with an unidentified opaque hy-drous iron silicate within olivine and pyroxene megacrysts, along fractures in the mega-crysts, and in altered glass. The sulfide minerals include monosulfide solid solution (mss),pentlandite and chalcopyrite, and minor pyrrhotite, violarite, pyrite, idaite, and Fe-richintermediate solid solution (iss). Mss is generally S-rich (>53 at.o/o S) and occurs as twolargely intergrown lamellar phases: mss(l) wirh 6.2 to 12.0 at.o/o Ni and mss(2) with 13.9to 19.7 at.o/o Ni. Pentlandite coexisting with mss or violarite is Ni-rich, whereas thatcoexisting with pyrrhotite is relatively Fe-rich and is invariably S-rich. The proportion ofsulfide minerals present in an aggregate and within the volume of a handspecimen variesmarkedly, but sulfides in olivine megacrysts have relatively more mss(2) and relativelyless chalcopyrite.

The phase relations for the iron-nickel sulfides bear only a fair correspondence to thoseof low-temperature laboratory studies. In particular, the observed miscibility gap in mssis on the Fe-rich side of the system, and the field of mss(2) does not extend to the Ni-richside. The extent of unmixing of mss is variable even within a single bomb sample, butthis may reflect annealing on a cooling gradient. The subspherical shape of the sulfidemineral aggregates and the consistency of bulk sulfide and olivine compositions with ex-perimental high-temperature partitioning data suggest an origin through sulfide liquidimmiscibility. However, the presence of many sulfide aggregates along fractures and indacitic matrix and the variable bulk sulfide composition contradict an interpretation ofsulfide liquid immiscibility and suggest instead that high-temperature subsolidus process-es, such as degassing, S infiltration along fractures, and Ni diffusion, may have beenimportant.

INrnooucrroN 1975, Meyer and Boctor, 1975; Peterson and Francis,Most nickel-copper sulfide ores consist of pentlandite, 19771' Clarke, 19791, Haggerty et al., 1979; Tsai et al.,

pyrrhotite, and chalcopyrite and may have exsolved at 1979, Boctor and Boyd, 1980; Botkunov et al., 1980;low temperature from high-temperature monosulfide sol- Hunter and Taylor, 1984).id solution (mss; e.g., Naldrett and Kullerud, 1967). In this paper, mss occurring in samples of basalt andKnowledge of the subsolidus phase relations and cooling andesite tephra from a cinder cone on Mount Shasta,history of mss is therefore necessary to understand fully California (Anderson, 1974a, l9l4b) with as much asthe origin of nickel-iron-copper sulfide minerals and of I9.l at.o/o Ni is described in detail, and its origin andnickel-copper sulfide ores. The subsolidus phase relations phase relations are discussed and interpreted. The mss isof mss are known from experimental study in the system unmixed (two-phase) in the form of a lamellar inter-Fe-Ni-S (Naldrett et al., 1967:- Kullerud et al., 1969; Mis- growth and occurs with pentlandite, chalcopyrite, and anra and Fleet, 1973; Craig, 1973) and are supported by unidentified opaque hydrous iron silicate as micrometer-study of the composition and phase relations of quenched sized, subspherical aggregate inclusions within mega-single-phase mss with up to 17.1 at.0/o Ni in samples of crysts of olivine, orthopyroxene, and clinopyroxene. Thebasalt(Desboroughetal., 1968; SkinnerandPeck, 1969; mss from Mount Shasta is the most Ni-rich reported inKanehira eL al., 1973; Marhez, 1976;, Czamanske and nature and, therefore, presents an opportunity for com-Moore, 1977) and nodules in kimberlite, lherzolite, or parison with the phase relations determined experimen-eclogite(DesboroughandCzamanske, l9T3;Frick,,1973; tally for relatively Ni-rich bulk compositions in the Fe-Bishop et al., 1975;' de Waal and Calk, 1975; Haggerty, Ni-S system.

0003-004x/89/09 r 0-098 1 $02.00 98 1

982 STONE ET AL.: TWO-PHASE NICKELIFEROUS MONOSULFIDE SOLID SOLUTION

TABLE 1. Composition of the megacrysts (wt%)

Sample: 5A 5MDescription: ol ol

5Kol core

5Kol margin

5Yopx

5C 5Ccpx core cpx margrn

sio,Tio,AlrosFeO

MnOMgoCaONaroNio

Total

38.8000.01

11.7500.13

48.990 . 1 50.010.18

100.02

88

40.310.010.029.410.130.13

49.80 . 1 50.03o.23

100.22

90

39.2000.037.24o.020.10

53.120.120.01o.24

r 00.08

93

37.810.030.03

13.5300

48.36o.1400 . 1 5

100.30

86

56.750.060.805.730.610 . 1 1

33.311.660.070.06

99.16

50.120 6 02.34

11.870.250.19

14.3618.670.580

oo nl

52.730.161.343.750.660.08

19.8320.280.310.05

99.19

J

88

Fo (%)wo (%)En (%)Fs (%)

4054

o

39421 q

Nofe; ol, olivine; opx, orthopyroxene; cpx, clinopyroxene.

Prrnocn-q.pHy AND MTNERAL cHEMrsrRy

Experimental details

The Mount Shasta region of northern California is un-derlain by Pleistocene and Holocene lava and pyroclasticrocks of basaltic, andesitic, and dacitic composition (An-derson, 197 4a). The samples for this study were obtainedin 197 7 from the Holocene cinder cone labeled S- I 7 (An-derson, 1974a), which is on the northern flank of MountShasta. The samples were in the form of bombs of denseolivine-phyric and olivine- and plagioclase-phyric basaltand basaltic andesite removed from a quarry face in thecinder cone. The section exposed has been divided intolower and upper strata (Anderson, 1974a). Sample I isfrom the east end of the lower strata, sample 5 is from50 feet (about 15 m) west of a vent in the lower strata,and sample 6 is from the vent area. Samples 5A, 58, 5C,5K, 5L, 5M, 5N, 5X, 5Y, and 52 are polished thin sec-tions of sample 5. Several polished thin sections wereprepared from about 10 bomb samples. The sulfide ag-gregates had a very limited distribution and, where moreabundant, were present at less than 0.01 modal percent.More than half the thick sections did not contains sul-fides, and only sample 5 yielded visible sulfides in mostpolished sections prepared from it.

Electron-microprobe analyses of the sulfide and silicateminerals were made with a reor JxA-8600 automated Su-perprobe, with synthetic NiS and FeS, natural chalcopy-rite, and natural silicate minerals as standards. Analyseswere made at 15 kV, l0 n,4,,20 s or 20000 total counts,and a beam diameter of 2 trm. All analyzed sulfide areaswere checked by ros analysis. The counting time for Niin megacryst phases was 50 s, and that for Ni, Cu, and Sin glass was 100 s. The Ni content for glass was calibratedagainst that of a standard basalt glass (UWO2) with 169ppm Ni (cf. Fleet eI al., 1977 , 198 l). Repeated analysesof the standards used for analyses of the silicate mineralsand the sulfide minerals indicate an accuracy relative toabsolute element concentrations of within 0.5 wto/o for

SiOr, FeO, MgO, and CaO; 0.04 wto/o for NiO; and I wto/ofor Fe, Ni, Cu, and S.

Megacrysts

The olivine megacrysts are euhedral, rounded to irreg-ular, single or composite grains that are generally un-zoned and range in composition from Fo* to Fon, (Tablel), which is less magnesian than the analyses of Anderson(1914a; Fono). However, where the grain margin is inter-grown with the matrix, as in sample 5K, it is relativelymore Fe-rich (Foru).

The orthopyroxene megacrysts are unzoned spongymasses of composition WorEnrrFs, (Table l), which ismore magnesian than the analysis of Anderson (1974a1'WouEnruFs,r). The clinopyroxene megacrysts are large,optically zoned and twinned euhedra or composite grains,with or without orthopyroxene. Single clinopyroxenemegacrysts have an Fe-rich core of WornEnorFs, com-position (Table l), which resembles the relatively Fe-richclinoplroxene analysis of Anderson (197 4a; WorrEnorFsrr),in abrupt contact with a broad margin of Worn-ooEnr-rr-Fsu-, composition, which is more magnesian than themore magnesian clinopyroxene analysis of Anderson(1974a1' WorrEnroFs,,). Locally, the clinopyroxene mega-crysts are spongy at the grain margin.

Subspherical to irregular inclusions of the rock matrixoccur within the megacrysts, particularly within the or-thopyroxene megacrysts. They consist of dark reddish-brown glass, sulfide mineral aggregates, and microlites ofplagioclase and clinopyroxene. The glass is rhyodacite withhigh AlrO, and FeO/MgO ratio, low MgO (<2 wto/o) andvolatiles, and very low Ni, Cu, and S contents (Table 2).The glass analyses were made of spots 10 to 50 mm awayfrom the included sulfide mineral aggregates, and the rel-atively high Ni content of the glass in sample 5Y mayreflect contamination from the sulfides. The glass analy-ses have much higher SiO, contents than the analyses ofglass (basalt to andesite) given in Table 2 of Anderson(1974a).

STONE ET AL.: TWO-PHASE NICKELIFEROUS MONOSULFIDE SOLID SOLUTION

TABLE 2. Composition of the glass and the opaque iron silicate (wt%)

983

Opaque iron silicate

Sample: 5YGlass in: opx

5YMX

5M5KCACJ

mx5Sopx

sio,Tio,Alr03FeOCrrO.MnOMgoCaONaroKrOCuS

Ni.Nio

Total

66.871.34

13.008.9300.061.564.832.852.850.010.010.015n.o

100.96

70.531 .58

1 1 . 9 97.480.o20.121.563.331 7 0

2.000.000.000.005n.o.

100.41

70 .151 3 4

1 4 . 1 95.840.01o.020.90J .OO

3.580.700.000.000.0052n.o.

100.39

69.041.41

1 3 . 1 67.4700.040.943.943.730.970.000.000.0040n.d.

100.70

19.620.110.12

68.300.090.070.151 . 1 700 . 1 6n.o.n.o.n.o.o.24

89.79

15.400.100.13

70.320.040.090.070.630.070 . 1 1n.d.n.d.n.d.0.44

86.96

18.870.120.04

68.940.410 . 1 10.050.8200.03n.o.n.o.n.g .0.33

89.39, Represents Ni content scaled to the Ni content ot UWO2 glass.

An unidentified opaque hydrous iron silicate (Table 2)with optical properties similar to goethite and hereafterreferred to as opaque iron silicate comprises as much as40o/o of the subspherical shape of many sulfide mineralaggregates in thin section (Fig. l). The opaque iron sili-cate also occurs as tiny irregular replacements along cracksand irregular bodies at the intersection of cracks in themegacrysts. It probably represents alteration of earlier ironsulfide and, to a lesser extent, silicate host.

Sulfide mineral aggregates

The sulfide minerals present include monosulfide solidsolution (mss), pentlandite, and chalcopyrite and minorpyrrhotite, violarite, intermediate solid solution (iss),idaite, and pyrite. They occur generally with the opaqueiron silicate mainly as 5- to 100-pm subspherical to ir-regular aggregate inclusions within the megacrysts andrarely as single, irregular-shaped grains in the rock matrix(Fig. 1). Locally, sulfide aggregates occur along fracturesin megacrysts and appear to replace the megacryst par-allel to the fracture trend. They occur also in glass-bearinginclusions, particularly within orthopyroxene. Atthoughsulfide aggregates within the megacrysts commonly con-tain the assemblage mss, pentlandite, and chalcopyrite,sulfide aggregates with either two-phase mss or unmixed,nickeliferous mss are more common in olivine mega-crysts, whereas chalcopyrite is more common in pyrox-ene megacrysts.

Many of the sulfide mineral aggregates consist of un-mixed, two-phase mss with or without exsolved pent-landite (Figs. 1, 2), whereas other sulfide aggregates arecomposed of discrete grains of sulfide minerals. Violariteis spatially associated with pentlandite at the edge of anaggregate in sample 5Y, whereas pyrite occurs in aggre-gate form with mss and pentlandite in 5X, with violariteand chalcopyrite in 58, and as a large isolated grain inthe matrix of 5F. Chalcopyrite occurs as a discrete phasenot always near the edge ofan aggregate. Iron-rich inter-mediate solid solution (iss) is present in an isolated oc-

currence as a small grain in an altered, largely pyrrhotite-bearing aggregare within clinopyroxene, whereas idaite ispresent in an isolated occurrence as a very small grainaggregate with pentlandite within orthopyroxene.

The mss occurs as single-grain lamellar intergrowths(sample 5A', Fig. 2A). Individual lamellae are lathlike orflamelike, and their appearance in polished section clear-ly varies markedly with grain orientation. In Figure 2A,the lamellae are fine (l-3 pm thick) and homogeneouslymixed, whereas in transverse section, they are coarser (5-20 pm) and appear to be heterogeneously mixed. Thelamellae represent two unmixed (Fe,Ni), ,S composi-tions. Electron-microprobe spot analysis and backscat-tered-electron images (Fig. 2B) confirm that the light phaseof sample 5,A. in Figure 2A (which is the major or matrixphase) is relatively Ni-rich (15.3 at.o/o Ni) and the darkphase is Fe-rich (ll.l at.o/o Ni) (Fig. 3). We have labeledthese two phases mss(2) and mss(l), respectively. Theyare not directly equivalent to the mss(2) and mss(l) ofeither Misra and Fleet (1973) or Craig (1973). As notedbelow, the mss(2) phase of Misra and Fleet (1973) is muchmore Ni-rich than the present mss(2), but the presentmss(l) would logically represent a more evolved form ofthe Fe-rich solid solution [mss(l)] of Misra and Fleet(1973) and Craig (1973).

The mss grain of sample 5X (Fig. 4) was removed froma polished thick section and examined by X-ray preces-sion photography (Fig. 5). Reflections of two intergrownand crystallographically aligned NiAs-type phases arepresent (cf. Fleet and MacRae, 1969). From the unit-cellparameters and dro, subcell spacing, these phases areidentified as mss(l) and mss(2) (e.g., Naldrettetal., 1967;Misra, 1972): for mss(l), a: 3.41, c : 5.62, dror: 2.040A , and fo r mss (2 ) , a :3 .41 , c : 5 .38 , d ro2 : 1 .983 A(relative to the NiAs-type subcell). The present dro, andcompositional data (Table 3) for mss(l) and mss(2) arein approximate agreement with the data of Naldrett et al.( I 967) for mss quenched from 600 'C. The stronger X-rayreflections are those of mss(2), indicating that it is the

984 STONE ET AL.: TWO-PHASE NICKELIFEROUS MONOSULFIDE SOLID SOLUTION

major (or matrix) phase in the mss intergrowth, which isconsistent with the relative proportions of mss(l) andmss(2) observed in the polished section of this sample.The precession photograph (Fig. 5) shows that apparentlyboth mss( I ) and mss(2) have the 3C superstructure (Fleet,1968, 1971), which is consistent with their S-rich com-position. Also present in the original photograph (but notreproduced in Fig. 5) are reflections ofpentlandite, whichexhibits the orientation relationship with mss of exsolvedpentlandite in natural and synthetic nickeliferous pyrrho-tite (Francis et al., 1976).

The compositional data for the sulfide phases given inTable 3 are an electron-microprobe spot analysis or av-erages of two to four electron-microprobe spot analysesfor each phase. Computed standard deviations for theaveraged composition of the sulfides are within 2.1 wto/ofor elements present in excess of l0 wt0/0, 1.3 wtO/o forelements present in the range of 5 to I 0 wto/o, 0.6 wto/o forelements present in the range of I to 5 wto/0, and 0.17wto/o for elements present in amounts less than 1 wto/0.Microprobe analysis was complicated by the fine grainsize and the (unavoidable) imperfectly polished surfaces.Analyses that were clearly contaminated by adjacentphases or that resulted in very low totals (less than 95wtolo) were not included in the averaged values. However,the small but anomalous contents of Ni in chalcopyriteand iss and ofCu in pyrrhotite, pentlandite, and iss and,to an unknown extent, the excess S over the nominalcontent of pentlandite reported in Table 3 may all rep-resent secondary fluorescence interference. These com-positional anomalies are less pronounced or absent inanalyses of larger grains of other samples.

Most of the mss analyses have a cation total of 0.86-0.90 on the basis of one S, revealing that mss(l) andmss(2) are relatively S-rich (>53 at.o/o S; Table 3, Figs.3,4). The observed range in composition for mss(l) is 6.2to 12.0 at.o/o Ni and for mss(2) is 13.9 to 19.7 at.o/o Ni.The compositional gap between the mss(l) and mss(2) ina particular sulfide mineral aggregate is variable, but isnarrowest in sample 5A; the distinction of mss(l) andmss(2) here may be obscured by contamination from theadjacent lamellae during analysis. Although mss(2) oc-curs in the olivine-, orthopyroxene- and clinopyroxene-hosted aggregates, it is generally more common in theformer, and it always coexists with mss(l). Mss(l), how-ever, commonly occurs without mss(2), particularly inthe clinopyroxene-hosted aggregates.

The one seemingly uncontaminated pyrrhotite analysis

985

(sample l) has a cation total of 0.93, 0.27 at.o/o Ni, anddoes not have detectable Cu (Table 3; Fig. 3). Alterna-tively, the pyrrhotite analyzed in samples 5M and 5E-lis likely contaminated because the former has significantNi (1.2 at.o/o) and coexists with pentlandite, whereas thelatter has a low total and significant Ni (l at.0/o) and co-exists with pentlandite.

Analyses of the large and homogeneous pentlanditegrains (samples 5K-3, 5L,6-1, and 5E-l), where the l ike-lihood of contamination from other phases is minimal,have a metal to S rat io of 1.097 to l . l16, 26.4to 3 l . lat.o/o Ni, and as much as 0.5 at.o/o Cu (Table 3, Figs. 3,4). The metal to S ratio is less than the nominal 1.125,but it is within the range of pentlandite analyses madeelsewhere (Harris and Nickel, 19721- Misra and Fleet,1973). The most Ni-poor grains of pentlandite are thosethat coexist with pyrrhotite (samples 5M and 5E- l) ratherthan with mss or violarite. The composition of pentland-ite within orthopyroxene and clinopyroxene generally re-sembles that of pentlandite within olivine.

The analyses of violarite indicate variable Ni and Scontents and a metal to S ratio that exceeds the nominal0.750 (Table 3, Fig. 3). Similar metal-rich violarite hasbeen described elsewhere (Desborough and Czamanske,1973). The pyrite grain in sample 5F seems devoid of Ni,whereas the pyrite in sample 5X has 0.58 at.o/o Ni andthatin sample 58 has 1.33 at.o/o Ni (Table 3, Figs.3,4).However, the coexistence of the latter with violarite sug-gests that its relatively large Ni content may reflect con-tamlnatron.

The analyses of relatively large grains indicate thatchalcopyrite is nearly stoichiometric, with less than 0.50at.o/o Ni (Table 3). The presence of approximately 2l wto/oCu and nearly twice as much Fe as Cu in the Cu-bearingmineral in sample 1 suggests it may be Fe-rich interme-diate solid solution (iss) (Table 3, e.g., Barton, 1973; Ca-bri, 1973). However, the low weight percent total of theaveraged analysis indicates contamination by adjacentopaque iron silicate. Similarly, the low total and signifi-cant Si content ofthe averaged analysis ofidaite togetherwith its small size suggests contamination from the ad-jacent orthopyroxene during analysis.

Sur-rrnn PHASE RELATIoNS

There is general agreement that mss is a complete solidsolution between Fe,-"S and Ni,

"S to at least 400 "C,

but below 400'C, its phase relations are much debated(Naldrett and Kullerud, 1967; Naldrett et al., 1967; Misra

STONE ET AL.: TWO-PHASE NICKELIFEROUS MONOSULFIDE SOLID SOLUTION

(--

Fig. l. (A) Olivine-hosted sulfide mineral aggregate 5C-4 insample 5C. Mss(l) (dark gray), mss(2) (medium gray), pentland-ite (pn, bright), chalcopyrite (cp, bright), and opaque iron silicate(ois, dark). (B) Olivine-hosted sulfide mineral aggregate in sam-ple 5M. Pyrrhotite (po, medium gray), penrlandite (pn, bright),chromite (chr, dark gray), and opaque iron silicate (ois, dark).(C) Orthopyroxene-hosted sulfide mineral aggregate 5L. Mss(l)

(dark gray), mss(2) (medium gray), chalcopyrite (cp, pale gray),pentlandite (pn, bright), and opaque iron silicate (ois, dark). (D)Clinopyroxene-hosted sulfide mineral aggregate inclusion 5C-lin sample 5C. Mss(l) (darkest), mss(2) (medium gray), chalco-pyrite (cp, medium gray) and pentlandite (pn, brieht). Polishedsections, reflected light, oil; scale bars represent 0.01 mm.

986 STONE ET AL.: TWO-PHASE NICKELIFEROUS MONOSULFIDE SOLID SOLUTION

Fig. 2. (A) Two-phase lamellar intergrowth of mss(l) (dark) and mss(2) (light) in sample 5A with the opaque iron silicate (ois):polished section, oil; scale bar represents 0.01 mm. (B) Backscattered-electron image of the same area as (A); mss(l) is dark and

mss(2) is light.

and Fleet, 1973; Craig, 1973; Barker, 1983). Accordingto Misra and Fleet (1973), the continuity of the mss phasefield between Fe,-,S and Ni,-"S breaks down between400 and 300 "C, at a composition of about 33 at.o/o Ni.

This result was in agreement with Shewman and Clark(1970), who reported a miscibility gap at a similar com-position but at a slightly lower temperature (275 + l0'C). At 300'C. Misra and Fleet (1973) observed that the

STONE ET AL.: TWO-PHASE NICKELIFEROUS MONOSULFIDE SOLID SOLUTION 987

mm ,^-' ms( r , ms9tz ,

orl hopyroxene

A.-"-v"!'--- T1?)m33 /----=-'\ o

orthopyroxene

clinopyroxene c||nopyroxene

Fig. 3. Composition ofsulfide phases within the Fe-Ni-S system. Dots are representative analyses, X represents averaged valuesof Table 3, and * denotes chalcopyrite-bearing sulfide mineral assemblages. Experimental phase relations for 230 'C from Misraand Fleet (1973) are for reference purposes only: megacryst host indicated top right. O is the composition of mss(2) from Misra andFleet (1973).

mss field withdraws away from the Ni-S join and sepa-rates into two mss phases, mss(l) and mss(2), with about25 and 33 at.o/o Ni, respectively, which coexist with pent-landite solid solution (about 33.5 at.0/o Ni). At lower tem-peratures, the mss(l) phase field withdraws further to-ward the Fe-S join, to about 17 at.o/o Ni at 230 oC, whereasmss(2) fails to change much from its composition at 300"C. Plrite-pentlandite tie-lines are first established at about280 "C, whereas monoclinic pyrrhotite and pentlanditecoexist only when the mss(l) field has withdrawn to atleast l0 at.o/o Ni.

In contrast to Misra and Fleet (1973), Craig (1973)found that mss is a complete solid solution between Fe,_"Sand Ni,-,S until 282 oC, where millerite crystallizes. At263 + 13'C, a miscibility gap appears in mss at 18.5at.0/o Ni, below which two phases coexist, mss(l) andmss(2). A miscibility gap forms in mss(2) at 225 + 12lJ, near 26.3 at.o/o Ni, which Craig (1973) has suggestedmay correspond to the gap reported by Misra and Fleet(1973) at 300 "C. Below 250 oC, the miscibility gap ex-pands such that between 225 and 200'C, pentlandite and

pyrite coexist. Final decomposition of the mss phases oc-curs below 100-200 "C.

For all the sulflde mineral aggregates from the MountShasta cinder cone, the miscibility gap between mss(l)and mss(2) is in the interval 12.0 and 15.3 at.o/o Ni (15.5to 18.9 wt0/o), and largely within the mss(l) phase field ofMisra and Fleet (1973) and Craig (1973) (Figs. 3, 4). Theactual extent of miscibility varies from sample to sample;1 to 17 .4 at.o/o Ni in sample 5X and I l. l to 15.3 at.o/o Niin sample 5A. Coincidentally, this natural miscibility gapresembles more closely that between nickeliferous pyr-rhotite solid solution and nickeliferous mss found in sam-ples of medium tenor nickel sulfide ore heated under lab-oratory conditions to 500 and 540 "C (McQueen,1979).However, previous experimentation under more con-trolled conditions (Misra and Fleet, 1973; Craig, 1973)failed to detect any evidence of a miscibility gap in mssat 500-540'C within the Fe-Ni-S system.

Compared to previous experimental studies, the mis-cibility gap is in the more Fe-rich part of the Fe-Ni-Ssystem. The large mss(2) phase field postulated by Craig

STONE ET AL.: TWO-PHASE NICKELIFEROUS MONOSULFIDE SOLID SOLUTION

55 45

Fig. 4. Composition of phases in sulfide mineral aggregatesof sample 5X used for precession photograph of Fig. 5 (see Fig.3).

988

(1973) to extend across the Ni-rich side of the Fe-Ni-Ssystem appears to be absent. It is replaced by pyrite-pent-landite, pyrite-pentlandite-violarite, and pentlandite-vio-larite assemblages (Figs. 3, 4), which is consistent withlaboratory experimentation (Misra and Fleet, 1973). Thediscrepancies between the Mount Shasta and laboratoryphase relations may reflect different bulk compositions,the presence ofadditional components, such as Cu, anddisequilibrium in both the laboratory experiments and innature at low temperatures (<300 "C).

We recognize that the products of laboratory annealingexperiments would be reaction-path dependent at theselow temperatures and that the Mount Shasta sulfides mightrepresent reheated material previously equilibrated tonear-surface temperatures. However, the phase relationsshown by the Mount Shasta cinder cone would representextension of the Misra and Fleet (1973, 1974) phase re-lations to lower temperatures. The present miscibility gapmight then be a second, lower-temperature unmixingevent. The persistence of the restricted and somewhatstranded mss(2) composition of Misra and Fleet (1973)is certainly a curious feature, but the present mss(2) fieldappears to behave in a similar manner. The present mss-pentlandite tie-lines are generally consistent with those ofMisra and Fleet (1973), whereas the present S-rich msscompositions are comparatively inconsistent. By extrap-olation of the progressive withdrawal of the mss(l) solvusof Misra and Fleet (1973) to more Fe-rich compositionswith reduction in temperature, the average mss compo-sitions at Mount Shasta unmixed between 180 and 220"C, the mss(l) of samples 5K-3, 5L, 5C-1, 5N, and 5X(Figs. 3, 4) equilibrated at about 140 oC, and the sulfidemineral aggregates with pyrrhotite equilibrated at about90 "c.

Although the pentlandite in the Mount Shasta sulfideaggregates tends to contain more Cu than coexisting mssor pyrrhotite, its Cu content and that of mss does notappear to increase with increased Fe content (e.g., Hill,1983), and the S contents of relatively Ni-rich and Ni-poor pentlandite are similar. Pyrite in nickel-copper sul-fide assemblages has generally been considered to be ofsecondary origin (Bishop et al. 1975; Maclean, 1977),although cooling of mss from high temperature shouldresult in exsolution of pyrite and vaesite with or without

Fig. 5. Representation ofpart ofzero-level X-ray precessronfilm of sulfide aggregate 5X (Fig. 4; Table 3). NiAs-type subcellreflections are streaked due to mosaic spread. 3C superstructurereflections are represented by small dots. MoKa, p : 20,3 d.

violarite from compositions along the S-rich boundary ofmss (Naldrett and Kullerud, 1967 Naldrett et al., 1967).Pyrite in the Mount Shasta samples coexists with mss insample 5X, but does appear to be secondary in samples5F and 5B. Craig (1973) has attributed the difficulty hehad in nucleating pyrite during formation of pyrite-pent-landite assemblages along the S-rich boundary of mss tothe short time for experimentation compared to that innature. By analogy, the lack of pyrite exsolution in thesamples from Mount Shasta, and in sulfide mineral ag-gregates from other localities, may reflect relatively rapidcooling. Consequently, the mss remains relatively S-richmetastably.

Although slable to 461 "C, violarite, like pyrite, is dif-ficult to nucleate in experimental studies (Craig, 1973),and its presence in nature is generally attributed to alter-ation of pentlandite (Craig and Scott, 1974; Misra andFleet, 1974). The spatial association of violarite withpentlandite and the opaque iron silicate in the MountShasta samples is consistent with its formation by alter-ation of pentlandite. The opaque iron silicate most likelyformed by low-temperature hydrothermal alteration ofthe Mount Shasta cinder cone, perhaps contemporane-ously with the formation of secondary pyrite and violar-ite.

PrrnocnNBsrs

Sulfide mineral aggregates similar to but less Ni-richthan those from the cinder cone at Mount Shasta havebeen reported in megacrysts within samples of nodulesfrom kimberlite (Haggerty, 7975; Clarke, 1979;Haggertyet al., 1979; Boctor and Boyd, 1980; Botkunov et al.,1980; Hunter and Taylor, 1984), nodules from lherzoliteand eclogite xenoliths (Bishop et al., 7975; Tsai et al.,1979), and in samples ofbasalt (Skinner and Peck, 1969;

STONE ET AL.: TWO-PHASE NICKELIFEROUS MONOSULFIDE SOLID SOLUTION

TABLE 3. Composition of minerals in the sulfide mineral aggregates

989

FeGu

Sample Phase (at.%) (wt%)

olololololololololololololololololololololololgtsgtsgtsopxopxopxopxopxopxcpxcpxcpxcpxcpxcpxcpxcpxMXcpxcpxcpxcpxcpx

Nofej mss(1) is relatively Ni-poor mss, mss(2) is relatively Ni-rich mss, pn is pentlandite, po is pyrrhotite, vl is violarite, cp is chalcopyrite, py is pyrite,iss is intermediate solid solution, id is idaite, ol is olivine, opx is orthopyroxene, cpx is clinopyroxene, and gls is glass.

'Analyses with as much as or more than 0.50 at.% Si.

5A mss(1)5A mss(2)5M pn5M po5K-1 mss(1)5K-1 mss(2)5K-3 mss(1)5K-3 mss(2)5K-3 pn5N pn5N mss(1)5N mss(2)5X mss(1)5X mss(2)5X pn5x pv5C-4 mss(1)5C-4 mss(2)'5C-4 cp5C-4 pn58 v l '59 cp'58 py-5Y cp5Y pn5Y vl5L cp5L pn5L mss(1)5L mss(2)52 id'52 pn'5C-1 mss(1).5C-1 mss(2)-5C-1 cp'5C-1 pn6-1 mss(1)6-1 pn1 p o1 iss5F pv5E-1 po5E-1 pn5E-4 mss(1)5E-4 pn-5E-4 cp

36.1732.1126.0946.4439.3529.1340.4529.6823.5321.9839.8230.1139.09287723.1032.0032.9926.9524.6421.296.22

24.7031 2024.1320.9218.4625.5522.6238.9128.591 1 .4622.5438.7931.7224.2721.8540.9322.8349.7932.6432.80+c.o4

25.4840.4421 .5925.0s

1 1 . 0 915.2826.411 . 1 87.00

19.686.70

16.8728.6630.757.O2

1 6.127.02

17.3728.510.58

1 1 . 9 818.311 . 1 7

3 1 . 1 340.8001.380.10

30.5225.990.14

29.237.21

17.420.89

28.797.25

13.860.30

28.606.39

29.460.270 . 1 500.97

26.386.20

29.850.06

000.04000.020.010.030.1300000.010.0300.02o.02

23.'t70 .180.16

24.470

24.390.190.18

23.480.4800.03

39.600.350.010.04

23.790.830.020.420

15.0600.020.1300.22

24.00

52.7452.6147.4752.3853 655 1 . 1 352.8453.4247.6947.2653.17. J . I I

53.9053.8448.3667.4154.9954.7251.0248.4052.8250.8267.4151.3848.37c c . J /

50.8447.6753.9053.9748.0548.3253.9554.3851.6448.7252.6647.3051.9552.1667.2053.4747.9353.3748.3350.89

45.6040.5531.8358.8750.9236.3751.7037.6129.1627.O051.5438.6149.7036.4627.8244.2441 .1634.7528.8025.247.77

30.0344.7129.0824.9523.16an 7(

27.6949.8236.6412.8627.3350.3641.0229.6725.8951.9727.1161.3339.9847.36s6.9030.5150.4125.2028.90

14.6920.2933.691.579.52

25.90o

22.4737.3439.719.55

21.739.38

23.1436.1 00.85

15.7424.811.45

37.6553.6402.07o.12

38.3534.340.17

37.629.70

23.451.05

36.799.91

18.880.38

35.708.54

36.720.370.1901.28

33.238.13

36.660.07

0 38.140 38.160.05 33.260 38 .120 39.870.02 36.650.02 38.770.04 38.860.19 33.940 33.330 39.510 39.510 39.350.03 39.170.04 33.450 53.500.02 39.210.03 40.31

30.94 34.08o.24 32.300.16 37.66

34.00 3s.300 55.17

33.60 3s.400.26 32.970.27 39.68

32.10 35.140.67 33.510 39.660.04 39.71

50.76 30.810.49 33.490.02 40.020.06 41.02

33.23 36.061 .13 32.990.04 38.340.57 32.700 38.26

20.99 36.680 54.920.02 38.330.18 32.890 38.200.29 34.43

31.51 33.70

Czamanske and Moore, 1977). These sulfide mineral ag-gregates have analogous mineral assemblages and equiv-alent host-rock associations to ore deposits ofnickel-cop-per sulfides and may have been formed by similarprocesses. Therefore, the petrogenesis of sulfide mineralaggregates is of some importance because of the potentialfor insight into both the S content of upper-mantle peri-dotites and the origin ofnickel-copper ore deposits. Themodels commonly invoked for the genesis of nickel-cop-per sulfides range from the formation of an immisciblesulfide liquid during the early magmatic stage of the hostrocks, either by segregation (e.9., Naldrett, l98l) or as-similation (e.g., Lesher and Groves, 1986), to sulfidemetasomatism during the late-magmatic stage or subso-lidus history under conditions that excluded chemical

equilibration between the silicate minerals (e.g., olivine)and the sulfides (e.g., Fleet, 1979; Fleet and MacRae,I 983) .

Features invoked by proponents of sulfide liquid im-miscibility include the association with mafic and ultra-mafic rocks, a bulk sulfide composition within the bound-aries of mss in the system Cu-Fe-Ni-S (Boctor and Boyd,1980), and the subspherical shape of sulfide mineral ag-gregates. The subspherical shape ofa sulfide mineral ag-gregate, in particular, is generally considered to indicatethat it was a sulfide droplet that remained liquid as thehost solidified, following segregation of the sulfide liquidfrom the parental silicate melt and inclusion within eithercumulate silicate minerals or in the matrix (Skinner andPeck. 1969: Czamanske and Moore. 1971.

990 STONE ET AL.:TWO-PHASE NICKELIFEROUS MONOSULFIDE SOLID SOLUTION

5M

30 40 50

MgO (wt%)

N i O( w l % )

5 0 4 0 5 0MqO (wt%)

Fig. 6. Plot of NiO vs. MgO contents of olivine megacryst hosts and Ni partitioning dala (KB: 33; Fleet et al., 1977; Fleetand MacRae, I 983, I 988). Dots are individual spot analyses; X represents averaged values given in Table l. The lines that originatefrom the lower right corner are isopleths of constant NiS content.

N i o(wf %)

Crystallization of a sulfide liquid droplet is consideredgenerally to follow the crystallization and subsoliduschanges determined by experimental study in the systemsFe-Ni-Cu-S (Craig and Kullerud, 1969; Hill, 1983) andFe-Ni-S and Fe-Cu-S (Naldrett eL al., 1967l, Kullerud etal.,1969; Misra and Fleet, 1973; Craig,1973; Craig andScott, 1974). Cooling to approximately ll00 'C fosterscrystallization of mss, consuming nearly all the Ni in thedroplet, and the residual sulfide liquid is Cu-rich. Pro-gressive cooling below 610'C forces exsolution ofpent-landite from the mss, which eventually converts to pyr-rhotite. Cu within the mss is partitioned into pentlanditeor exsolves as chalcopyrite. Sulfide mineral aggregates withmss preserved and intricate intergrowth textures are con-sidered to be quenched.

The residual, Cu-rich liquid separates from the mss at850 "C (Craig and Kullerud, 1969) and crystallizes as iss.At approximately 550 'C, the iss recrystallizes to chal-copyrite and Fe-rich iss (Barton, 1973 Cabri,1973; Boc-tor and Boyd, 1980; Haggerty et al., 1979), which withlow-temperature re-equilibration, recrystallizes to chal-copyrite and pyrrhotite (Macl-ear', 1977). Ni in issexsolves as pentlandite.

N i o(wt%)

Fleet et al. (1977), Fleet (1979), and Fleet and MacRae(1983) noted that the composition of olivine associatedwith nickel-copper sulfides is incompatible with equili-bration under magmatic conditions. For nickel-coppersulfide deposits that have not been metamorphosed be-yond low-grade conditions, the NiO contents of olivineare too high, and, correspondingly, the NiS contents ofthe sulfides are too low, relative to laboratory partitiondata. Metamorphosed deposits exhibit a progressive ap-proach to equilibrium values with increase in metamor-phic grade (Fleet and MacRae, 1983). Recent experimen-tation (Fleet and MacRae,1987,1988) has investigatedthe exchange of Ni/Fe between olivine and monosulfide-oxide liquid at 1300-1395 'C under controlled atmo-sphere conditions. The distribution coefficient (Kor) is es-sentially constant at about 30 to 35 in the temperaturerange 900 to 1400'C for olivine of composition Fon, toFoo, monosulfide composition with up to 70 mo10/o NiS,and a wide range of fo, andf,r. Ko, exhibits only a weakdecrease from 35 to 29 with increase inf, from the iron-wi.istite (IW) buffer to close to the fayalite-magnetite-sil-ica phase (FMQ) buffer. In comparison, calculated Ko,values for komatiite-hosted nickel-copper sulfides range

5K

STONE ET AL.:TWO-PHASE NICKELIFEROUS MONOSULFIDE SOLID SOLUTION 991

from 2 to 5. The olivines in these komatiites have far toomuch NiO to have been equilibrated \Mith the nickel-cop-per sulfides in the early magmatic stage. Alternatively,equilibration of olivine and sulfide compositions mightnot be expected if the sulfides had been formed by meta-somatic replacement during either the late magmatic stageor in the subsolidus history. One other feature of nickel-copper sulfides inconsistent with the early magmatic (im-miscible sulfide liquid) models is the erratic variation inbulk sulfide composition, between orebodies, within or-ebodies, and within individual handspecimens (Fleet,1978, 1979). There is no correlation of bulk sulfide com-position with magmatic fractionation index. The markedvariation in the composition of the nickel-copper sulfideblebs in handspecimen samples of disseminated ore fromthe Frood mine, Sudbury, Ontario is clearly inconsistentwith an origin through sulfide liquid immiscibility (Fleet,l 978) .

Features of the sulfide mineral aggregates from MountShasta, such as their subspherical shape and bulk com-positions, which plot in the mss field of the system Fe-Ni-S, appear consistent with formation by liquid immis-cibility. Furthermore, the NiO contents of many of theolivine megacrysts are more-or-less consistent with high-temperature equilibration with the included sulfide min-eral aggregates (Fig. 6). A sulfide aggregate from the samecinder cone as the present study, when homogenized byheating, gave a bulk sulfide composition of 40 mo10/o NiS,49.3 at.o/o S (Anderson, l9'74b). This is consistent withestimates of the maximum NiS content of the presentolivine megacryst-hosted sulfides, which is probablygreater than the average mss composition (26-30 molo/oNiS) and less than the coexisting pentlandite composition(Figs. 3, 4).

Thus the sulfide aggregates may have been original in-clusions in the megacryst crystals: segregated sulfide blebstrapped during crystal growth either from the parentalpicritic melt (Anderson, 1974a) or in a peridotitic sourceregion (the megacrysts being xenocrysts). One problemwith this reconstruction is the large variability in thechemical composition and mineral assemblage of theMount Shasta sulfide aggregates. Variable Ni/Cu ratio innickel sulfide ores and sulfide mineral aggregates has beenattributed to fractionation of immiscible sulfide liquid(Craig and Kullerud, 1969;,Czamanske and Moore, 1977;Maclean, 1977). Accordingly, the relatively greater Ni/Cu ratio ofthe olivine-hosted aggregates could reflect thepresence, during olivine crystallization, of a Ni-rich sul-fide liquid that fractionated to a more Cu-rich liquid asthe coexisting silicate liquid crystallized fractionated or-thopyroxene and clinopyroxene. However, if the olivine,orthopyroxene, and clinopyroxene megacrysts were co-magmatic and coeval, this would not be a very satisfac-tory explanation. Andersot (1974a) suggested that onlyolivine and orthopyroxene fractionated until the olivinecomposition reached l6 wto/o FeO, whereupon clinopy-roxene fractionated. Extrapolation to the composition ofthe liquids that coexisted with the megacrysts by using

distribution coemcients for equilibrium low-pressure ol-ivine-liquid partitioning of TiO, (0.02), orthopyroxene-liquid partitioning of TiO, (0. l), and clinopyroxene-liq-uid partitioning of TiO, (0.3) (Pearce and Norry, 1979;Grove and Bryan, 1983) and by using equilibrium low-pressure olivine-liquid partitioning of FeO/MgO (0.3),orthopyroxene-liquid partitioning of FeO /MgO (0.27), andclinopyroxene-liquid partitioning of FeO/MgO (0.29)(Roeder and Emslie, 1970; Green et al., 1979', Stolper,1980) suggests that the olivine, orthopyroxene, andmagnesian clinopyroxene coexisted with magma of sim-ilar FeO/MgO and TiO, contents and are comagmaticand indeed megacrysts, whereas the Fe-rich clinopyrox-ene cores coexisted with liquid higher in FeO/MgO andTiOr. These results reflect the mixing of picritic and an-desitic magma during petrogenesis of the Mount Shastarocks (Anderson, 1974a), a process more characteristic ofhigh-level crustal magmatism than magmatism in themantle.

As a second magmatic explanation, the basaltic andandesitic magmas may have become S-saturated throughassimilation immediately before eruption (e.g., Lesher andGroves, 1986). Infiltration of sulfide liquid along frac-tures in the megacrysts could then have resulted in for-mation of the sulfides and equilibration. Alternatively, alate-stage S-enriched fluid phase might be a more effec-tive melasomatizing agent.

Several features ofthe sulfide mineral aggregates fromMount Shasta, such as the presence of nickeliferous ag-gregates in dacitic glass-bearing inclusions within themegacrysts, alignment of many aggregates along cracks inthe magacrysts, and wide variation in bulk sulfide com-position (e.g., Ni/Fe and Ni/Cu ratios, Figs. 3, 4) are con-sistent with formation by subsolidus processes. We note,in passing, that the subspherical shape of many of thesulfide aggregates should not be regarded as unambiguousevidence for a precursor sulfide liquid droplet because itcould be a manifestation of high-temperature subsolidusreplacement processes (below). Also bulk sulfide com-position within the Fe-Ni-Cu system is defined by whole-rock composition. High-temperature rock-fluid processeswould certainly result in bulk sulfide composition iden-tical to that expected from sulfide liquid immiscibility.

The megacryst-bearing basalt and andesite tephra couldhave been sulfidized by hot gases during formation of thecinder cone. Isolated pyrite grains in the matrix indicatethat S-bearing fluids were present later in the cooling his-tory ofthe volcanic rocks (prior to the cinder cone activ-ity) by degassing during or following eruption. In thisprocess, S saturation results from quenching ofthe erupt-ed lava, and S-rich gas is expelled from the quenchedglass (Moore and Fabbi, l97l). Condensation of the S-richgas in fractures and vesicles forms sulfide liquid dropletsthat interact with accumulated silicate minerals at hightemperature. Ni and Fe scavenged from olivine react withthe sulfide to form mss. Degassing has been consideredfor the formation of some sulfide mineral aggregates,globules, and spherules in samples of MORB (Moore and

992

Calk, 1971; Frick, 1973; Mathez and Yeats, 1976;Macl,ean, 1977). The presence of nickeliferous sulfideswithin dacitic glass inclusions in the margins of the or-thopyroxene megacrysts certainly indicates mobility ofchalcophile metals during the subsolidus history of theserocks. Independent evidence of the mobility of chalco-phile metals under high-temperature subsolidus condi-tions comes from the presence of small amounts of Ni,Fe, and Cu in sublimate over some lavas of Hawaii(Naughton et al., 197 4). Whether the sulfide mineral ag-gregates from Mount Shasta formed by sulfide liquid im-miscibility or subsolidus processes is as yet unresolved.However, the possibility of their formation in responseto degassing deserves further consideration.

The foregoing discussion emphasizes the ambiguousnature ofthe petrographic and geochemical features usu-ally associated with sulfide liquid immiscibility. The as-sociation of nickel-copper sulfides with mafic and ultra-mafic rocks, a bulk sulfide composition within theboundaries of mss, intricate intergrowth textures, andsubspherical shape do not collectively constitute uniqueevidence of quenched immiscible sulfide droplets. Theyare equally consistent with formation of the sulfides byhigh-temperalzre subsolidus processes. As long as equi-Iibrium is attained (or approached), the bulk sulfide com-position is still defined by ambient "f",, .f",, temperature,and bulk compositions.

AcxNowlpncMENTS

We thank R. L. Barnett and D. M. Kingston for their assistance withthe electron-microprobe analysis and John Forth for the preparation ofthin sections. Critical comments by S. D. Scott and an anonymous re-viewer were very helpful. The research was funded by a Natural Sciencesand Engineering Research Council of Canada operating glant to M.E.F.

RrrnnrNcrs crrnoAnderson, A.T. (1974a) Evidence for a picntic, volatile-rich magma be-

neath Mount Shasta, California. Joumal of Petrology, 15, 243-267 .-(1974b) Chlorine, sulfur, and water in magmas and oceans. Geo-

logical Society of America Bulletin, 85, 1 483-1 492Barker, W.W. (1983) The Fe-Ni-S system + (Co,Cu), CSIRO Report no.

FP 26, 8 l p.Barton, P.B. (1973) Solid solution in the system Cu-Fe-S. Part L The

Cu-S and Cu-Fe-S joins. Economic Geology, 68, 433465.Bishop, F.C., Smith, J.V., and Dawson, J.B. (1975) Pentlandite-magnetite

intergrowth in De Beers spinel lherzolite. Review of sulphides in nod-ules. Physics and Chemistry ofthe Earth, 9,323-328.

Boctor, N.2., and Boyd, F R (l 980) Ilmenite nodules and associated sul-phides in kimberlite from Yakutia, U.S.S.R. Carnege Institution ofWashington Year Book, 1979, 302-304.

Botkunov, A.I., Goranin, V.K., Kudryavtseva, G.P., and Perminova, A.(1980) Mineral inclusions in olivine and zircon from the kimberlitepipe "Mir." Doklady Akademii Nauk SSSR, 251, 1233-1236.

Cabri, L.J. (1973) New data on phase relations in the Cu-Fe-S system.Economic Geology, 68, 443454.

Clarke, D.B. (l 979) Synthesis of nickeloan djerfisherites and the origin ofpotassic sulphides at the Frank Smith Mine. In F.R. Boyd and H.O.A.Meyer, Eds., Proceedings of the second international kimberlite con-ference. The mantle sample: Inclusions in kimberlites and other vol-canics, vol. 2, p. 300-308. American Geophysical Union, Washington,D.C.

Craig, J R. (1973) Pyrite-pentlandite assemblages and other low temper-

STONE ET AL.:TWO-PHASE NICKELIFEROUS MONOSULFIDE SOLID SOLUTION

ature relations in the Fe-Ni-S system. American Journal of Science,273A,496-510.

Craig, J.R., and Kullerud, G. (1969) Phase relations in the Cu-Fe-Ni-Ssystem and their application to magmatic ore deposits. Economic Ge-ology, Monograph 4, 344-358.

Craig, J.R., and Scott, S.D. (1 974) Sulfide phase equilibria. In P.H Ribbe,Ed., Sulfide mineralogy Mineralogical Society of America Reviews inMineralogy, 1, CS-l-CS-1 10.

Czamanske, G.K., and Moore, J.G. (1977) Composition and phase chem-istry of sulfide globules in basalt from the Mid-Atlantic Ridge rift valleynear 37'N lat. Geologrcal Society ofAmerica Bulletin, 88, 587-599.

Desborough, G.A., and Czamanske, G K. (1973) Sulfides in eclogite nod-ules from a kimberlite pipe, South Africa, with comments on Yiolaritestability American Mineralogist, 58, 195-202.

Desborough, G.A., Anderson, A.T , and Wright, T.L. (1968) Mineralogyofsulfides from certain Hawaiian basalts. Economic Geology, 63,636-644.

de Waal, S.A., and Calk, L.C (1975) The sulfides in the garnet pyroxenite

xenoliths from Salt Lake Crater, Oahu. Journal ofPetrology, 16, 13+I 53 .

Fleet, M.E. (1968) The superstructures of two synthetic pyrrhotites. Ca-nadian Journal ofEarth Sciences, 5, I 183-l 185.

-(1971) The crystal structure of a pyrrhotite (Fe?SJ. Acta Crystal-lographica, 27, 1864-1867 .

-(1978) Origin of disseminated copper-nickel sulfide ore at Frood,Sudbury, Ontario. Economic Geology, 7 2, 1449-1456.

- (1979) Partitioning of Fe, Co, Ni, and Cu between sulfide liquidand basaltic melts and the composition of Ni-Cu sulfide deposits-Adiscussion. Economic Geology, 7 4, 1517 -1519.

Fleet, M.E., and MacRae, N.D. (1969) Two-phase hexagonal pyrrhotites.

Canadian Mineralogist, 9, 699-7 05-(1983) Partition of Ni between olivine and sulfide and its appli

cation to Ni-Cu sulfide deposits. Contributions to Mineralogy and Pe-tro logy,83,75-81.

-(1987) Partition of Ni between olivine and sulfide: The efect oftemperature, /o, and f,,. Contributions to Mineralogy and Petrology,95,336-342.

- (1988) Partition ofNi between olivine and sulfide: Equilibria withsulfide-oxide liquids. Contributions to Mineralogy and Petrology, 100,462469.

Fleet, M.E., MacRae, N.D., and Hertzberg, C.T. (1977) Partition of nickelbetween olivine and sulfide: A test for immiscible sulfide liquid. Con-tnbutions to Mineralogy and Petrology, 32, l9l-197.

Fleet, M.E., MacRae, N.D, and Osbome, M.D. (1981) The partition ofnickel between olivine, magma and immiscible sulfide liquid. ChemicalGeology, 32, ll9-127.

Francis, C.A., Fleet, M.8., Misra, K., and Craig, J.R.. (1976) Orientationofexsolved pentlandite in natural and synthetic nickeliferous pyrrho-

tite. American Mineralogist, 61, 9 13-920.Frick, C. (1973) The sulfides in griquaite and gamet-peridotite xenoliths

in kimberlite. Contributions to Mineralogy and Petrology, 39, 1-16.Green, D.H., Hibberson, W.O., and Jaques, A L. (1979) Petrogenesis of

mid-ocean ridge basalts. In M.W. McElhinny, Ed., The Earth: Its ori-gin, structure and evolution, p. 265-269. Academic Press, London.

Grove, T.L., and Bryan, W.B. (1983) Fractionation of pyroxene-phyricMORB at low pressure: An experimental study. Contributions to Min-eralogy and Petrology, 84, 293-309.

Haegerty, S.E. (1975) The chemistry and genesis ofopaque minerals inkimberlites. Physics and Chemistry of the Earth, 9, 295-307.

Haggerty, S.E., Hardie, R.B., and McMahon, R.M. (1979) The mineralchemistry of ilmenite nodule associations from the Monastery dia-treme. In F.R. Boyd and H.O.A. Meyer, Eds., Proceedings of the sec-ond international kimberlite conference, vol. 2, p. 249-256. The man-tle sample: Inclusions in kimberlites and other volcanics. AmericanGeophysical Union, Washinglon, D.C.

Harris, D.C., and Nickel, E.H. (1972) Pentlandite compositions and as-sociations in some natural minerals Canadian Mineralogist, 11, 861-878.

Hill, R.E.T. (1983) Experimental study of phase relations at 600'in aportion of the Fe-Ni-Cu-S system and its application to natural sul-

phide assemblages. In D.L Buchanan and M.J. Jones, Eds., Sulphidedeposits in mafic and ultramafic rocks, p. 14-21. The Institution ofMining and Metallurgy, London

Hunter, R.H., and Taylor, L.A. (1984) Magma-mixing in the low velocityzone: Krmberlite megacrysts from Fayette County, Pennsylvania.American Mineralogist, 66, 16-29.

Kanehira, K., Yui, S., Sakai, H., and Sasaki, A (1973) Sulfide globulesand sulfur isotope ratios in abyssal tholeiite from the Mid-AtlanticRidge near 30'N latitude Geochemical Journal, 7, 89-96.

Kullerud, G., Yund, R.A., and Moh, G.H. (1969) Phase relations in theCu-Fe-S, Cu-Ni-S, and Fe-Ni-S system. Economic Geology, Mono-graph 4,323-343.

Lesher, C.M., and Groves, D.I (1986) Controls on formation of koma-tiite-associated nickel-copper sulfide deposits In G. Friedrich et al.,Eds., Geology and metallogeny ofcopper deposits,p.43-62. Springer-Verlag, Heidelberg

Maclean, W H. (1977) Sulfides in Leg 37 drill core from the Mid-AtlanticRidge. Canadian Joumal ofEarth Sciences, 14,674-683.

Mathez, E.A. (1976) Sulfur solubility and magmatic sulfides in submarinebasalt glass Journal of Geophysical Research, 81, 4269427 5

Mathez, E.A, and Yeats, R.S. (1976) Magmatic sulfides in basalt glassfrom DSDP hole 3l9A and site 320, Nazca plare. In R.S. Hart et al.,Eds., Initial reports ofthe Deep Sea Drilling Project, vol. 14, p. 363-373. U.S. Government Printing Office, Washington, D.C.

McQueen, KG. (1979) Experimental heating and diffusion effects in Fe-Ni sulfide ore from Redross, Western Australia. Economic Geology,74,140-148.

Meyer, H O A., and Boctor, N.Z. (1975) Sulfide-oxide minerals in eclogitefrom Stockdale kimberlite, Kansas. Contributions to Mineralogy andPetrology, 52,57-68.

Misra, K.C. (1972) Phase relations in the Fe-Ni-S system. Ph.D. thesis,London, Ontario, University of Western Ontario, 213 p.

Misra, K C , and Fleet, M E. (1973) The chemical compositions of syn-thetic and natual pentlandite assemblages. Economic Geology, 68, 5 I 8-539.

-(1974) Chemical composition and stability of violarite. EconomicGeology, 69,391403-

Moore, J.G., and Calk, L. (1971) Sulfide spherules in vesicles ofdredgedpillow basalt American Mineralogist, 56, 47 6488.

993

Moore, J.G., and Fabbi, B.P. (1971) An estimate ofjuvenile sulfur contentofbasalt Contributions to Mineralogy and Petrology, 33, ll8-127.

Naldrett, A.J. (1981) Nickel sulfide deposits: Classification, composition,and genesis Economic Geology, 75'h Anniversary Volume, 628-685.

Naldrett, A.J., and Kullerud, G. (1967) A study of the Strathcona mineand its bearing on the origin of the nickel-copper ores of the Sudburydistrict, Ontario. Journal of Petrology, 8, 453-53 l.

Naldrett, A.J., Craig, J., and Kullerud, G. (1967) The central portion ofthe Fe-Ni-S system and its bearing on pentlandite exsolution in iron-nickel sulfide ores. Economic Geology, 62,826-84'l.

Naughton, J.J., kwis, V.A, Hamrnond, D., and Nishimoto, D. (1974)The chemistry of sublimates collected directly from lava fountains atKilauea Volcano, Hawaii. Geochimica et Cosmochimica Acta, 38, 167 4-r690

Pearce, J.A., and Norry, M.J. (1979) Petrogenic implications of Ti, Zr, Y,and Nb variations in volcanic rocks. Contributions to Mineralogy andPetrology, 69,3347.

Peterson, R, and Francis, D (1977) The origin of sulfide inclusions inpyroxene megacrysts. American Mineralogist, 62, 1049-105 l.

Roeder, P.L., and Emslie, R.F. (1970) OlivineJiquid equilibrium. Con-tributions to Mineralogy and Petrology, 29,275-289.

Shewman, R W, and Clark, L.A. (1970) Pentlandite phase relations inthe Fe-Ni-S system and notes on the monosulflde solid solution. Ca-nadian Journal of Earth Sciences. 7 . 67-85.

Skinner, B.J., and Peck, D.L. (1969) An immiscible sulfide melt fromHawaii. Economic Geology, Monograph 4, 310-322.

Stolper, E.M. (1980) A phase diagam for mid-ocean ridge basalts: Im-plications for petrogenesis. Contributions to Mineralogy and Petrology,t 4 . t J - z t .

Tsai, H., Shieh, Y., and Meyer, H.O.A. (1979) Mineralogy and S34/S3,ratios of sulfides associated with kimberlite, xenoliths and diamonds.In FR Boyd and H O.A. Meyer, Eds., Proceedings of the second in-ternational kimberlite conference. The mantle sample: Inclusions inkimberlites and other volcanics, vol. 2, p. 87-103. American Geo-physical Union, Washington, D.C.

MeNuscnrrr RrcHvED DecrrraseR. 5, 1988Mer.ruscnrrr AcCEPTED Mnv 8, 1989

STONE ET AL.: TWO-PHASE NICKELIFEROUS MONOSULFIDE SOLID SOLUTION