Canadian MinerologistVol. 21, pp. 529-5dA (1983)

THE FORMATION OF SUBCALCIC GARNET IN SCHEELITE-BEARING SKARNS

RAINER J. NEWBERRY.8

Deportment of Applied Eorth Sciences, Stonford University, Stanford, Califurnia 94305, U.S.A.

ABSTRACT

Compositions ofgarnet from 18 tungsten-bearing skarnsin the western Cordillera of North America, with additionalexamples from the literature, indicate that solid solutionbetween grandite and plralspite is far more extensive thanpreviously reported; garnet with up to 75 mole 9o andraditecomponent can contain significant almandine - spessartine("subcalcic") component. Detailed microprobe analysesshow that skarn garnets are typically zoned toward moresubcalcic compositions from core to rim and from peripheryof skarn toward the inferred fluid-source. l\mong tungstenskarns, those that possess subcalcic garnet are more abun-dant, they generally possess higher tungsten grades and ton-nages, and they appear to be more typical of occurrencesin the Cordillera. A model for the occurrence of subcalcicgamet in a skarn requires relatively low oxidation-state, highmanganese-ion activity, and low calcium-ion activity inskarn-forming fluids. These conditions are best met inskarns formed at moderate depths, in highly carbonaceoushost-rocks and in zones chemically isolated from calcitemarble. Semiquantitative calculations suggest that at oxida-tion states approximately at or below pyrite- magnetite-pyrrhotite, changes in activity of calcium hydroxide in solu-tion by a factor of 2 can produce the observed compositional variations. Subcalcic garnet generally is characteristicof tungsten skarns because such skarns are tlpically foundassociated with relatively reduced plutons or wallrocks orfound at relatively great depths.

Keywords: garnet, skarn, scheelite, tungsten, mineraldeposits, metasomatism.

SovvAnn

La composition du grenat dans dix-huit skarns ir tung-stCne de la cordillbre Ouest de I'Amdrique du Nord, ainsique dans les exemples pris dans la litt6rature, montre quela solution entre grandite et pyralspite est de beaucoup plusdtendue qu'on ne le croyait auparavant. Un grenat qui con-tient jusqu'd 7590 du p6le andradite (base molaire) peutcontenir une proportion appr€ciable d'une composante"subcalcique" d almandin + spessartine. Des mesuresd6tai[6es i la microsonde montrent que le grenat des skarnsest generalement zon6, avec un enrichissement en tennessubcalciques du coeur vers la bordure d'un cristal et deslimites du skarn vers la source de la phase fluide (6tabliepar inf6rence), Les skarns i tungstdne qui contiennent ungrenat subcalcique sont plus repandus et plus typiques dans

EPresent address: Department of Geology, University ofAlaska, Fairbanks, Alaska 99701, U.S.A.

la cordillbre, et ils possOdent des teneurs et des rdserves entungstene sup6rieures aux exemples sans grenat subcalci-que. Pour trouver un tel grenat dans un skarn, il faut unmilieu relativement reducteur, et une activit6 6lev€e de I'ionMn et r6duite de I'ion Ca dans les fluides responsables. Cesconditions seraient le mieux satisfaites dans les skarns for-mds i profondeur intermddiaire, dans les roches-hdtes enri-chies en matibre organique et dans des zones isol€es chimi-quement des marbres i calcite. Des calculs d'ordre semiquantitatif montrent qu'i un niveau d'oxydation d6fini parl'assemblage pyrite - magndtite - pyrrhotine, ou m6me inf6-rieur d ce niveau, un changement d'un facteur de deux dansI'activit6 de I'hydroxyde de calcium en solution peut expli-quer les variations de composition observ6es. Un grenat sub-calcique est caractdristique des skarns i tungstbne parce quede tels skarns sont typiquement associ6s i des plutons ouroches encaissantes relativement rdduits ou form6s i grandeprofondeur.

(Traduit par la R€daction)

Mots-clds: grenat, skarn, scheelite, tungstbne, gites min6-raux. m€tasomatisme.

INTRoDUcTION

Garnet from tungsten-bearing skarns in Japan(Shimazaki 1977) covers much of the range in com-position between grossul€fite and almandine -spessartine, with reported compositions limited toandradite contents of less than 15 mole q0, andtypically less than 5 mole 90. Based on these data,Shimazaki (1977) concluded that "the lack of Fe3*is essential for the formation of Fd+ and/orMn2+-bearing grossular under low temperature andlow pressure conditions." Similar occurrences ofgarnet with appreciable pyralspite and low andraditecontents are ieported from the tungsten skarns atVictory mine, Nevada (Lee 1962), the MacMillanPass deposit, Yukon Territory (Dick 1970, theCanada Tungsten, Baker and Lened deposits, North-west Territories @ick 1980), the Costabonne deposit,France (Guy 1980), and Dchenitschke, central AsiaNdmec 1967). From similar data, Zharikov (1970)and Shimazaki (1977) concluded that the chemicalenvironment favorable for scheelite deposition islimited to one in which ferrous-iron-bearing grossularis stable. Zharikov (1970) interpreted this to requireacidic skarn-forming fluids, whereas Shimazaki(1977) suggested, instead, highly reduced en-vironments characterized by oxidation states belowthe pyrrhotite - pyrite - magnetite buffer.

529

530 THE CANADIAN MINERALOGIST

In contrast with the above studies, investigationsof the tungsten-bearing skarns at Lost Creek, Mon-tana (Collins 1977),King Island, Tasmania (Kwak& Tan 1981) and the Osgood Mountains, Nevada(Iaylor l97Q showed that only grandite garnet is pres-ent. Reconnaissance studies of garnet compositionsfrom the Pine Creek mine, California (Wrighr 1973)and from the Black Rock mine, California @lliotl97l), on the other hand, indicated the presence ofandraditic garnet with significant almandine + spes-sartine component. These data show that there is abroad range of garnet compositions in tungstenskarns, that a significant almandine-spessartine com-ponent is not restricted to the grossularitic end ofthe grandite series, and that some tungsten skarnslack subcalcic garnet.

This paper summarizes an investigation into theenvironments of tungsten-skarn formation, with par-ticular emphasis on the occurrence and compositionsof subcalcic garnet and its relationship to scheelitedeposition. It is based on a detailed study of tungstenskarns in the Sierra Nevada area of California(Newberry 1980), with additional data and examplesfrom Idaho, Nevada, Montana, Yukon Territory andNorthwest Territories. In this paper, the term sub-calcic garnet identifies a skarn garnet that containsmore than 5 mole 9o almandine + sDessartinecomponent.

ANALYTICAL PRoCEDURES

Approximately 60 polished thin sections represent-ing 14 skarns were analyzed using an ARL AMX

electron microprobe at Queen's University, King-ston, Ontario. Major elements were determined byenergy dispersion using standards whose energy spec-tra had been previously stored on magnetic tape, andemploying an accelerating potential of l5kV, a beamcurrent of 100-120 pA, a sample current of 0.034pA, a beam diameter of 2-4 pm and a counting in-terval of 120 seconds. Raw data for the elements wereinitially corrected on an Ns-880 minicomputer us-ing a multiple least-squares routine. The intensityratios were corrected for matrix effects followingBence & Albee (1968) as well as for drift using a For-tran correction program developed by P.L. Roeder(pers. comm.). Alpha factors used in the iterativecorrection procedure were taken from Albee & Ray(1970). Analytical accuracy was checked by analyz-ing garnet and feldspar standards during eachsession.

An additional 43 samples were prepared as grainmounts and analyzed on an 8-channel ARL electronmicroprobe at the University of California, Berkeley.Samples were analyzed using a 15 kV filamentvoltage, a sample current of 0.2-0.3 pA,.beamdiameter of 2 pm, and a counting time of 50 seconds.Standards employed included grains of well-characterized garnet, pyroxene, rhodonite, anorthite,hematite and chromite. Raw data were corrected fordrift; intensity ratios were corrected for matrix ef-fects following Bence & Albee (1968) using a For-tran program developed by M.L. Rivers (pers.comm.). Several garnet samples analyzed at bothlaboratories yielded negligible differences incomposition.

TAELE I. REPRESENTATIVE COMPOSITIONS OF SUBCALCIC GARI{FI

N o . l a l b 2 3 4 5 6 7 8 9 t 0 l l 1 2 t 3 1 4 t 5(core) (rin)

leight g

si02 36.12 36.82 36.67 36.73T iq 0 .19 o .2o 0 .35 0 .16A 1 2 0 3 1 4 . 6 7 1 9 . 6 0 9 . 1 9 1 7 . 3 6Cr203 0 .15 0 .13 O.23 0 .00Fe{F 17.61 16.19 18.68 14,28lifno 12.76 l8.tl8 3.A2 19.39Mso 0.31 0.86 0.20 0.46Cao 16.23 9.48 30.51 11.94

ToIAL .02 101.n 99.6s l@.32

ible jN gamet endffibers

17.8 14 .6 27 .8 14 .429.6 l t ,4 59 .0 20 .O29.7 40.6 8.7 44.721.2 29.6 3.1 1 9.01 . 2 3 . 3 0 . 8 1 . 90 .4 0 .4 0 .7 0 .0

36.37 37.88 36.830.55 0 ,49 0 .126.42 19 .38 4 .560 . 3 6 0 . @ 0 . 1 3

22.71 9.40 24.563,98 p.50 2.260.31 0.39 0.09

24.57 ?3.42 3t.89

99.26 100.50 100.42

l4.l 51.7 12.267.8 13 .4 79 .39 .1 20 .9 5 .26 .3 12 .5 2 .61 . 2 1 . 5 0 . 3' I

. l 0 .1 0 .4

37.60 37.48 38.480.00 0.45 0.39

21.75 18 .81 18 .180.00 0 .12 0 .27

12.A7 13.88 7.O26.60 10 .62 3 .790.35 0.31 0.40

19.19 19 .14 32 .00

98.34 100.81 100.53

54.7 37.5 67.00.4 t6.3 20.' l

15 .0 23 .9 a .24 .5 20 .7 2 .31 . 4 1 . 2 1 . 50 . 0 0 . 4 0 . 8

49.7 77.66.1

'17.A

9 . 9 5 . 132.4 4.1o . 7 0 . 60 .8 0 .8

GrAdSpAlmPyUY

37.20 36.350.00 0 .16

la .m 15.100.00 0.05't6.21

14.941 6 . 3 6 1 3 . 1 7o,29 0 .35

il.07 19.47

100.05 99.59

21.7 24.11 0 . 0 3 1 . 837.2 30 .029.7 12.7l . t 1 . 40 . 0 0 . 1

36.94 38.360 . 1 8 0 . 1 7

21.51 18 .560.25 0 .27

16.72 7 .334.49 2 .350 . 1 9 0 . 1 6

20.19 32.86

t00.48 100.06

37.12 37.560.36 0.37

19.04 20 .130 . l l 0 . 1 46.92 13 .86

0,06 0 .1831.53 22.86

98.55 98.87

7t .3 57 .812.9 4.68 .0 8 .87 .2 27 .70 .2 0 .70 .4 0 .4

r-Total ircn,as Fe0;-Gr-= grgssularite, nd = andradlte, Sp = spessrtine, Alm = almndine, py. pylpe, Uv = uvamvlte[ o s , l , 2 : P i n e c r e e k m i n e , l l S 9 l s t s u b l e v e l d r l f t , l = . 6 n i m l n t r u s i o n , 2 = 6 m f m i t i t r o i i o n ! , - 5 : B l a c k R o c k n i n e , 3 = 7 l lstope,_6700 level, 4 = 833.stope, 6600 l-eyel, 5 = 7ll drift, 6700 level; 6,7: Strawbemy mine, 6 . iein gamet fm outcrop of $7ntne, 7 = gamet near quartz mnzonite dike, K lens, Znd level, f l olne; g: outcrcp of alpine mine, 7 m im lntrusion. subcalcicrir; 9-l l: outcmp of Gamet Dike-mine, 8 =-Long Ridge am, gimet vein in skarn 2 m rnirintrusi6n, 9 = upper mrkings ire;, skamadjacent to pegmtlte dike, l0 = lorer-mrkings-area, near inioskarn contacti lz,l3: Hardpoint mine,'12 = ubil ttp-tt tsi.si, gi*"tvein in skan 3 m fm dike contact, 13 = outcrcp near portal, gamet vein ln sfim; 14: ;utcmp of Tungsten ,tlm nine, nasiiiegamet skam 2 n fm intnsion; 15: Lened prcspect, Cac Valley area, 3 m fm intrusion. Ska; lmati6ns shm on Fiqure l

SUBCALCIC GARNET IN SCHEELITE.BEARING SKARNS 531

FIG. l. Location of tungsten-bearing skarns of the westernNorth American Cordillera discussed in this paper.Skarns with subcalcic garnet indicated by closed sym-bols; skarns lacking subcalcic garnet indicated by open

Calculation of molecular end-members was madefollowing Rickwood (1968) after assignment of fer-ric and ferrous iron. Ferric iron was determined bysubtracting the amount necessary to completetrivalent occupancy from total iron per 24 oxygenatoms; the remaining iron was assumed to be fer-rous. The analysis was rejected as unsatisfactory ifthe cation total (ferrous iron * manganese + cal-cium) differed from 6.00 gter 24 oxygen atoms) bymore than 2t/0. Typical subcalcic garnet composi-tions are listed in Table l.

Gsot-ocv AND GEocHEMISTRYoF SUBCALCIC GanNgt OccunnsNcrs

Geology of subcalcic garnet occurrences in skarnsol the North American Cordillera

Most of the hundreds of tungsten-bearing skarnsin the Western Cordillera of Canada and the UnitedStates ars located at or near contacts betweencalcareous metasedimentary rocks and coarse-gained calcalkaline plutons of Triassic to Cretaceousage. The locations of tungsten-bearing skarns dis-cussed in this paper are shown in Figure l. Of thethirty occurrences shown, all but seven contain sub-calcic garnet. In all cases, subcalcic garnet t quartzforms an assemblage with a consistent parageneticsetting, proximal to the intrusive contact. The de-tailed occurrence of subcalcic garnet and space-timevariations in its composition vary with the type ofhost rock.

Tungsten-bearing skarns exist in host rocks rang-ing from pure marble to calcareous hornfels, andfrom massive units to thinly intercalated or diamic-titic units. These different host-rocks impart char-acteristic morphologies and patterns of zonationupon the skarns. Wallrocks also vary from extremelycarbonaceous to noncarbonaceous or even hematitic,

symbols. Deposits: l) Lost Rivdr, Ak. (Dobson 1982);2) Spruce Hen prospect, Fairbanks district, Ak.; 3) Mac-millan Pass, Yukon-Northwest Territories (Dick 1976);4) Clea, Yukon (Dick 1980); 5) Lened, N.W.T. (Dick1980, this study); 6) Canada Tungsten, N.W.T. (Dick1980); ? Baker, N.W.T. @ick 1980); 8) Emerald, B.C.;9) Colorado Gulch, Mont.; l0) Lost Creek/Brown'sLake, Mont. (Collins 1977); 1l) Tungsten Jim, ld.; 12)Osgood Mountains, Nv. (Taylor 1976); 13) Mill City,Nv.; 14) Victory mine, Nv. (Lee 1962); 15) MonteCristo, Nv. (Sonnevil 1979); 16) Emerson, Nv.; 17)Alpine, Ca.; 18) Garnet Hill, Ca. @rock 1970); 19)Black Rock, Ca.;20) Hilton Creek, Ca.;21) Hardpoint,Ca.;22) Strawberry, Ca. (this study, Nokleberg l98l);23) Scheelore, Ca. (Morgan 1975): U) Pine Creek, Ca.(this study, Brown 1980); 25) Tungsten Hills, Ca.; 26)Garnet Dike, Ca.;27) Consolidated, Ca.; 28) TulareCo., Ca.; 29) Darwin, Ca. (this study, Eastman 1979);30) El Jaralito, Sonora, Mexico (Peabody 1979, Dunn1980).

c>l l\

KEY:

Scieellte skarnswlth subcalcic'garnets

.thls sludy

t othgr workers

Scheellte skarnslacklng

subcalclc garnels

o thls study

underllned deposlts have>.6 mt ore

0)roductlon + reserveo

13" o12

532 THE CANADIAN MINERALOGIST

TABLE 2. CON1MST1NG CHAMCTERISTICS OF REDUCED AND OXIDTZED TUNGSTEN SIGRNS

Redlced skarnsCharacteri stlc

host rccks

associated plutonicrccks

Stronqly reduced

strongly carbon-aceous ( ) 5Ucarbon comn)

ilrenlte or mgiletltebearing

endoskarn nlneralogy pyroxene-plagloclase: quaru

skarn nlneralogyprcgrade assemblages

pyroxene/garnet ratlo 1.0:1 to 2:1pyroxene composition'mle Z hedenberglte 60-90r c l e % J o h a n n s e n l t e 0 - 1 0

garnet composltlonmle Z andradite 10-30

'late garnet composltlonrc]e Z andradite 0-30mle Z spessartine 3-35mle g almandlne 3-40

retrograde assenblagestyp lca l m lnera logy b io t l te , p lag ioc lase

( l quar tz ' ca lc l te ) hornb lende, opaquesanphlbole compositionmle g ferrctrerc'l lte 0-50rcIe g trmlite 0-10mle Z pargasite 50-98

diagnostic opaques pyfrhotite(Bi snuth )

exanples Hardpolnt, Cali lacMillan pass, yTCanada Tungsten, NLrfIungsten Jim, ld.

0xldized SkarnsModerately reduced

mderately carbon- non-carbonaceousaceous (0.5 to 2, or hemtltlccafbon typical )

mgnetlte bearlng mgne'l ltebearing

pyrcxene-plagloclase quartz-epldote +I quaf tz -ep ldo te p lag loc lase

2, I to l t2 l . :2 to 1 :10

50-70 20-4510-25 0- 5

30-75 50-80

blotite, hornblende, epidote, chlorite,ch lo r i te ,opaques ac t lno l i te ,opaques

5-60 30-6010-40 30-705-70 5-20

pyrite-regnetite- pyrite+mgnetitepyrrhotite (btsmutfinlte)

Pine Creek, Ca. osgood Mntns., Nv.Black Rock, Ca. Lost Crcek, l, lont.E l , la ra l i to , Mex lco Tungsten H i l l s , Ca.Colorado Gulch, l,4ont.

80-980 - 30

10-755-40z-35

*reminde|ls diopslde i *rem,l nder i s grossul arj te

indicating that the oxidation state of host rocks canbe highly variable (Newberry 1980).

Compositions of skarn-mineral assemblages, com-positions, and local geological settings indicate thattungsten skarns can be classified into three cate-gories: "strongly reduced", "moderately reduced "and "oxidized". Characteristics of these three typesare listed in Table 2, adapted from Einaudi et a/.(1981). Reduced skarns, characterized by ferrous-iron-rich minerals, are found in carbonaceous hostrocks or associated with reduced plutons (crystallizedat depth or ilmenite-bearing, or both). Oxidizedskarns, characterized by ferric-iron-rich minerals, arefound in noncarbonaceous or hematitic host rocksor associated with oxidized pluton$ (crystallized atshallow depths or magnetite-bearing, or both). Sub-calcic garnet has only been identified in the reddcedskarns; andradite and epidote-quartz ure the para-genetically equivalent assemblages in oxidized skarns.

Most tungsten-bearing skarns show either simpleor complex mineralogical zonation depending on thehost lithology. Skarns formed from relatively pure,thick marble beds, (e.g., the Canada Tungsten mine,N.W.T., and the Pine Creek, Hilton Creek, andGarnet Dike mines, Ca.), are characterized.bymineralogically simple time-equivalent metasomaticzones containing mineral assemblages with increas-ing calcium contents progressing from the intrusive

contact toward marble. In these simple cases, zonescontaining subcalcic garnet a quaftz are at the cen-tre of skarn veins and closesl to intrusive contacts,in contact with and replacing salitic pyroxene - grzm-dite garnet skarn. Zones of subcalcic garnet containvariable amounts of quartz, typically between 5 and80V0, but on average about 30V0. They contain noscheelite (Newberry & Einaudi l98l) and form asharp cut-off in grade adjacent to ore-bearinggarnet-pyroxene skarn.

Veins of subcalcic garnet a quartz commonly ex-tend outward from massive garnet-quartz, replacinggarnet-pyroxene skarns several metres from the con-tact (Fig. 2). Garnet compositions vary systematicallywithin these veins, becoming less subcalcic withdistance into the garnet-pyroxene skarn @ig. 2) andeventually approach the composition of the wallrockgrandite. Grains of subcalcic garnet within thegarnet-quartz zone may be slightly zoned, becom-ing 5 to 20 mole 9o more subcalcic from core to rim(Table l).

Skarns formed from heterogeneous carbonatelithologies do not show the well-developed outcrop-scale zoning characteristics of skarns formed fromthick marble beds, except that a subcalcic garnet -l-quartz zone may be directly adjacent to the pluton.Calc-silicate grains and layers formed by meta-morphic recrystallization axe typically overprinted by

SUBCALCIC GARNET IN SCHEELITE-BEARING SKARNS 533

KEY: f-lenooskarn f garnet-pyroxene skarn

rtz monzonlte[..T] pyro r"ne ho rnf ets

!subcalcic garnet+ quartz

grossularite andradite

metasomatic calc-silicates; individual garnet grainsare typically zoned, in most cases with grossulariticor granditic cores and increasingly subcalcic rims@igs. 3, 4). Subcalcic garnet typically coats the in-side of vugs formed by coalescing grains of granditegarnet (Fig. 5). Subcalcic garnet grains also occuras late, discontinuous veins in garnet-pyroxene skarnand hornfels, and are not present in marble or otherhigh-calcium rocks.Where subcalcic garnet occursadjacent to calcite, the calcite is always a late, vug-filling mineral (Fie. 3). These textures suggest thatsubcalcic garnet did not form in a given spot untilthe calcite originally present was entirely dissolvedby metasomatic fluids.

Zonation in garnet qompositions also occurs onthe scale of hundreds of metres (e.9., the Black Rockmine, Fig. 6). The maximum amount of subcalciccomponent present in garnet at Black Rock variesfrom more than 8090 near the quartz monzonite con-tact to less than l07o in skarn exposed about 250metres from the contact. At the Strawberry andHardpoint skarns the maximum subcalcic contents

2m+

Frc.2. Simplified map of the E rib, 1189 4th sublevelworkings, Pine Creek mine, showing distribution ofskarn types and variation in garnet composition withina subcalcic garnet - quartz vein. Note decreasingly sub-calcic composition of garnet with increasing distancefrom the intrusive contact. Retrogmde veins and altera-tion masses, which are structurally superimposed on theearly skarns, are not shown.

are found in garnet adjacent to quartz monzonitestocks and dykes; the sparse data suggest a decreasein the subcalcic component with distance from theintrusive body. At Black Rock, garnet compositionsare not a simple function of distance from the in-trusive body, but rather reflect the latest paths offluid flow controlled by sedimentary layering.

Most known oxidized skarns formed from in-homogeneous carbonate lithologies [e.81., KingIsland, Aust.: Edwards et sl. (1956)1, so that theirzoning sequences are difficult to decipher. In sev-eral cases [Osgood Mountains, Nevada: Taylor(lW6), Tungsten Hills, California: Newberry (1980),Lost Creek, Montana: Collins (1977)1, however, thesequence appears to be: a narrow outer wollastonite-idocrase zone, a thick main zone of andraditic garnett salitic pyroxene, and an inner epidote-quartz zoneof variable width adjacent to the intrusive body.Garnet grains within the garnet t pyroxene zone ofoxidized skarns are typically zoned from a grossu-laritic core to a nearly pure andradite rim (Taylor1976, Collins 1977).

a lmand ine + spessar t ine

12 ' . \a \

a a

534 THE CANADIAN MINERALOGIST

a lmand ine +

grossular i te andradi te

FIc. 3. Sketch from a thin section of skarn from the 831 stope, 6600 level, Black Rock mine, showing compositionalzoning and development of subcalcic rim on grandite garnet. Subcalcic garnet lueas are indentified in thin sectionby pale orange color in transmitted light and anomalous birefringence with crossed polarizers. Note pattern of zon-ing ol increasingly subcalcic garnet compositions toward rim. Late calcite - chlorite veinlets not shown.

almandine + spessari lne

fl F-:-l

g l . lsubcalc lc I lerandt teg m

fi 6n"o"nbershtc

i ffi."r*':- ffilccheerrte

quartz

calcllegrossularlt lc

FIG.4. Sketch of thin section from No, 2 orebody pit, Strawberry tungsten mine. Metamopphic grossularitic core withquartz and diopside inclusions is surrounded by metasomatic grandite with scheelite and hedenbergitic pyroxeneinclusions, in turn rimmed by subcalcic garnet. Vug is filled with late, clear calcite, Garnet compositions becomemore subcalcic toward rirn. Garnet types are indicated by color differences in transmitted light and crossed polarizers.

Iw

SUBCALCIC GARNET IN SCHEELITE.BEARING SKARNS 535

Bulk-composition relationships in subcalcic grrnet

Compositions of subcalcic garnet determined inthis study are plotted in terms of mole 9o grossu-larite, andradite, and almandine + spessartine com-ponents in Figure 7. The plot clearly indicates thewide range in solid solution between grandite andpy'ralspite found in these occurrences. Almandine-spessartine substitution is clearly not limited to veryandradite-poor compositions: a garnet with as muchas 75 mole 9o andradite can contain appreciablealmandine + spessartine. Garnet compositionsrecorded from strongly reduced skarns, similar tothose reported by Shimazaki (1977), form a subgroupcharacterized by low andradite and high grossularitecontents. Garnet compositions recorded frommoderately reduced skarns form a compositionallydistinct group with a higher andradite content. Thecontrast in compositions between the two groups ofreduced skarns is further illustrated in terms ofrelative almandine and total iron components (Fig.8). In the moderately reduced skarns, total-iron con-tents are approximately independent of almandinecontent, whereas the strongly reduced skarns showa linear increase in iron content with increase inalmandine content. The former trend reflects an aD-proximately fixed amount of total iron in garnet,distributed between ferrous and ferric forms, ratherthan a fixed ratio of ferrous to ferric iron.

Differences in relative almandine,/spessartineratios also help to differentiate between the two typesof skarn CIable l). Strongly reduced skarns arecharacterized by high almandine/spessartine ratios(l : I to 4: l), whereas moderately reduced skarns havelow ratios (l:l to l:5, averaging about 1:2). Thesedifferences are addressed in a subsequent section.

Compositional differences between the two groupsare summarized in a plot of the molar ratios an-dradite,/grossularite yersas almandine,/spessartine(Fig. 9). Strongly reduced skarns are characterizedby low andradite,/grossular and variable, but general-ly high almandine,/spessartine ratios, whereasmoderately reduced skarns possess relatively high an-dradite and low almandine. Subcalcic garnet com-positions from Japanese tungsten skarns (Shimazaki1977) show similar trends, but witJr lower overall ironcontents. The strongly antithetic relationship observ-ed between almandine,/spessartine and andra-dite,/grossularite ratios suggests a causal relationshipbetween the two ratios and merits further considera-tion. Clearly, if garnet compositions are uniquely fix-ed by a buffered /(O), then almandine-andraditeratios should not show this strong inverse correla-tion. Other factors besides oxidation state must beimportant.

| , , : " ' , , IKey:

lll garnet-pyroxEne Efl nasnetlte

5tr subcabtcsamet V% pvrtte

E*lcu"'t I calclre

Ftc.5. Sketch of an unusually subcalcic-garnet-rich handspecimen from the 711 stope, 6700 level, Black Rockmine, showing the complexity of the mineral zonationpresent. The skarn protolith consisted of interlayeredmarble and calc-silicate hornfels beds; replacement ofthe marble interbeds from contacts with hornfelsresulted in the formation of garnet-pyroxene skarnfollowed by subcalcic garnet. Vugs were filled withpyrite, pyrrhotite, magnetite, quartz and calcite duringlater retrograde alteration.

DtscusstoN

Oxidation state and stabitity of subcalcic garnet

The oxidation state present during early stages ofskarn formation may be estimated by using ex-perimentally.derived mineral-asssemblage buffers ap-propriate for calcic garnet-pyroxene assemblages ad-justed for mineral composition. Results of such cal-culations (Newberry 1980, Einaudi et ql. l98l)demonstrate the contrast in oxidation state betweenskarns that lack and skarns that possess a subcalcicgarnet. These workers suggested that tungsten skarnsgenerally form at oxidation states below or onlyslightly exceeding the pyrrhotite - magnetite - pyritebuffer and that there is some causal relationship be-

536

A - l

THE CANADIAN MINERALOGIST

m l n el e Y e l

q . o o

) lU

\--

O O

oo o o o o

o oo o o o o o

oo o oo o oo o

KEY:

o o o \ \ o o

o

o

o

o o \f\oo o O . p o

o o \ 0 O o o

1 0 o m s t e r s

Poat- 8karn dlkes. narlmuat alnaBdlno+

# apeaaartlne oonteot

o t earno t

at lhat Dolnt

data presented by Liou (1973) indicates that over thetemperature range of 750 to 600oc for the as-semblage anorthite-spinel-garnet-quartz, the r atioof almandine to andradite component in garnet isa function of /(O). The-restriction of subcalcicgarnet to oxidation states at or.below Ni-NiO inLiou's (1973) experiments, however, is not reflectedin natural assemblages owing to two.differences be-tween the experimental and natural occurrences: (l)variability in the calcium content of natural mineralassemblages and (2) the presence of a substantialmanganese component in the skarn-forming fluids.These effects are considered in the followineparagraphs.

Calcium octivity and subcalcic garnet in skorns

Subcalcic garnet compositions from moderatelyreduced skarns are characterized by variableFe+ /FE* ratios, independent of total iron content(Fie. 8). Garnet compositions in these skarns becomemore subcalcic with proximity to the fluid source@igs. 2, 6) and with time @igs. 3, 4); subcalcic garnet

Ii:!

- \ ,r _ J

-l

o

H'",0," ffi *ffif,?||

Q quarrz monronrrs

Ftc.6. Simplified geological cross-section through the Black Rock mine, showing distribution of skarn and change ingarnet compositions within the mine, Numbers indicate the maximum subcalcic component in garnet at various points,projected into the line of section. Compositional data presented include values of cell dimension and index of refrac-tion from Elliot (1971). Note systematic increase in maximum subcalcic component of garnet toward the quartzmonzonite contact. Late, high-angle faults (ess than 20 m displacements) have been deleted for clarity.

tween oxidation state and skarn type. For example,biotite compositions indicate that most plutons ofthe eastern and central Sierra Nevada cooled alonga buffered path appoximately 2-3 log units abovethe Ni-NiO buffer @odge et al. 1969, Dodee 1972).This log/(O) - T path corresponds to thelOt -T regime of moderately reduced tungsten skarns(Newberry 1980) found in the central and easternSierra Nevada of California. In the western UfritedStates, oxidized tungsten skarns (Fig. l) occur eastof the deeper batholith terranes (Sierra Nevada,Aconchi, Idaho, Nelson); depth-of-formationestimates for these deposits are generally less thanthose to the west, and oxidation states werepresumably higher (Newberry & Einaudi l98l).Thus, the mineralogy of a typical tungsten skarn inpart simply reflects the imposed oxidation state.

This evidence from natural assemblages is cor-roborated by the experimental studies of Liou $nr,who found that grandite garnet with a significantalmandine component could form only at an oxida-tion state along or below the Ni-NiO buffer.Graphical interpretation after Winchell (1958) of

SUBCALCIC GARNET IN SCHEELITE.BEARING SKARNS

almandine + spessartine

FeaAlrSi"O.,, M%lbsieorz

537

strongly reduced

. Garnol Dlke, Cal l { .

a Hardpoint ,Cal l f .

OTungslen J im, ldaho

r Lened, NWT, Canada

+Jap tnesoTung!ten ekarng(Sh lm lzak l t e77 )

modera te l v reduced

Plne Greek, Ca[ | . o

Black Rock. Cal l l .A

Strawbsrry, Cal l f . o

Hl l ton Creek, Cal l f@

Tulare Co. Cal l f . *

Alp lno, CaH. O

FIc. 7' Compositions of all subcalcic garnet specimens analyzed for this study, expressed in terms of mole go grossularite,andradite, and almandine + spessartine. Filled symbols represent garnet from strongly reduced skarns, open circlesrepresent garnet from moderately reduced skarns. Nole in this figure, and in Figures 8 and 9, the close similaritybetween strongly reduced skarns of this study and the Japanese skarns of Shimazaki (1977).

andradi teClFerSi.O*mole pe rcen t

K*=

is not found as a replacement of high-calciumminerals. These facts suggest that a low calcium-ionactivity is a critical prerequisite to formation of sub-calcic garnet in skarns. For moderately reducedskarns, formation of subcalcic garnet from granditeskarn can be approximated by a reaction that con-serves iron and aluminum:

3Ca3Fe2Si3O12 + 2CarAlrSirO,, + l5H2O+

in garnet in garnet

2Fe3Al2Si3O12 + /"O2+9SiO2in garnet q:uaftz

+ l5Ca(oH)2(aq) (t)

temperature, calcium ion is represented by a calciumhydroxide complex; depending on fluid composition,chloride, fluoride and carbonate complexes wouldpresumably be present as well.) The equilibrium aconstant for the reaction, assuminga(H2O) : a(SiOJ = 1, is:

o$:Eto"no U"aua3 ar?

(2)

If oxygen fugacity, temperature and pressure are ap-proximately constant, as might be expected duringthe formation of a single zoned grain of subcalcicgarnet (e.g., Fig. 4), then the activity of the calcium

A

o

.s

*?oo oo A a

. t? n,

o ^ A o o

e o l o^

" Xpo&

A

i ' oO

' + t + t t r

- l r .

A o ^A

oo

o a a D

- ^ o n ^o A o

o o $ - 9 q o m d

"&";(d4:D' % o ao o Q o ^ S

Y { t

+ '

o

. { *'$ j ' z ro

(owing to complexation of aqueous ions at high

538 THE CANADIAN MINERALOGIST

oN- q

ooolrogoE q

o

6o

o o

EIoEo

o oo oo € 8o o o

i"-\

t o 2 0

mole percent a lmandlne compon€nt

FIc.8. Almatrdine content versus total moles iron per 12 oxygen atoms in analyzedsubcalcic garnet. Garnet compositions from strongly reduced skarns show an ap-proximately linear relation between total iron and almandine content, garnet com-positio$ from moderately reduced skarns show approximately constant total ironcontent, independent of almandine content. Dashed line through points fromJapanese tungsten skarns describes the equation total moles iron : 7 mole qoAd + mole 9o Alm.

,,,A

I

ion in solution, as expressed by aca(o11,, oughtto be directly proportional toi aad3 ao2 / d*,2 anddesignated R. The calculation of the (uantity R iscomplicated by the fact that some of the componentshave strongly nonideal mixing properties (Ganguly1976) and thus, non-unit activity coefficient$ (Ganexr-ly & Kennedy 1974). Activity coefficients for thealmandine-grossular system at 850'C vary from 0.8to 1.8 (Cressey et al.1978) and are not known below850oC, although calculations by Saxena (1973) sug-gest that they may be as high as 3 or 4 at 400"C.Use of regular-solution models for derivation ofpyrope-grossular solid solutions (Newton 1977) hasresulted in highly divergent estimates of interactionparameters; thus a rigorous derivation of activitycoefficients is not employed here. Within the ac-curacy of the calculated values, activity coefficientscan be approximated as simple functions of the ap-plicable mole fractions.

Because almandine and spessartine mix in arelatively ideal manner (Ganguly & Kennedy 1974),as do grossularite and andradite (Ganguly 1976), ac-tivity coefficients can be formulated in terms of thegrandite and pyralspite components. Interpolation

between the calculated values of Cressey et al. (1978)and those of Saxena (1973) yields aclivity coefficientver,rel,r mole-fraction relations that can be approx-imated as simple functions. These are:

'ys= ^tad: I +3(l - X**aJ for I (Xs*ot">0'4,

zL : :-s(t-Xo-o,J for I (Xu.-oir. ) 0.55' and

?a- = 0.75 for 0.55 ( Xs-6x" ) 0.3.

The relationship between garnet composition and theactivity of a garnet end-member is modeled by amultisite mixing model (Wood & Nicholls 1978):

€.8., Qotm : Xp"z+3 X6,l+2 ?a*, thus

Xs;2+r5 XpS+u ̂yudt ^l"r'D- _ (3)

Xp&*6 ̂ ra^2

Values ofR calculated from equation 3 vary from105 to 1.8 x l0-8 within a single zoned grain ofgarnet (e.g., from the Black Rock mine). If calcium-ion activity is approximately constant, then /(Oz)must vary at a single spot by approximately 6 orders

SUBCALCIC GARNET IN SCHEELITE-BEARING SKARNS 539

of magnitude to account for the observed alman-dine,/andradite ratios. Such variation is exceedinglyunlikely. The observed variation in R, however, canbe completely accounted for by variations in the ac-tivity of calcium hydroxide by a factor of only 4.4(equation 2). Such variations, at constantfO) andT, are feasible once high-calcium minerals have beencompletely replaced or dissolved. Calculations byBird & Helgeson (1981), for example, indicate thargrandite is stable over a S-fold range in calcium-ionactivity at 400'C and I kbar.

In strongly reduced skarns, subcalcic garnet aquartz appears as a replacement of hedenbergiticplroxene t grossularitic garnet + plagioclase. Com-positional relationships indicate that andradite con-tents are roughly constant and independent of alman-dine content (Fig. 8). The reaction can be modeled as:

3 CaFeSirO6 + Ca3Al2Si3Op + 6 HrO +in pyroxene in garnet

Fe3AlrSirO,, +in garnet

+ 6 Ca(oH), (4)

for which, assuming a(HrO) : a(SiOJ = l,

Qarn Qcu(og)r6

K".:----;_=--: (5)ohd'as,

Reduction of ferric iron is not required in this reac-tion, and it is thus independent of /(O). Calcula-tions using the previously described activity-coefficients indicate that within a single deposit, avariation in aqueous calcium-ion activity by a fac-tor of 3 can account for the spectrum ofgrossular,/almandine values observed. Similar treat-ment of the data of Shimazaki (1977) indicates tharvariations in activity of the aqueous calcium ion bya factor of 2.2 could account for his observedvariations.

In summary, it is important to note that subcalcicgarnet in tungsten skarns is not in equilibrium witha calcic clinopyroxene - grandite garnet - scheelircassemblage, but rather replaces that assemblage.When forming the assemblage pyroxene - garnet -scheelite, skarn-forming fluids replacing marble havesufficiently high values of activity ofthe calcium ionto prevent formation of a subcalcic garnet. Con-tinued addition of a high-silica, low-calciummetasomatic fluid to a garnet - pyroxene - scheeliteskarn results in the depletion of calcium in the fluidand the concomitant deposition of subcalcic garnet+ quartz @ig. 10). Further depletion of calcium ionfrom the metasomatic fluid results in the formationof increasingly subcalcic garnet (Fig. l0). This sametrend of decreasing calcium activity toward the fluidsource in oxidized skarns results in the replacementof andraditic garnet by epidote + quaftz, an as-semblage with approximately the same bulk composi-

mole p€rcent almandlne/spessartlne

Frc.9. Ratio of mole 9o almandine to mole l7o spessartineversus ratio of mole 9o andradite to mole go grossu-ladte in analyzed specimens of subcalcic garnet. Notesharp distinction between moderately and stronglyreduced groups and strong inverse correlation betweenthe trilo ratios,

tion as subcalcic garnet t quartz (Fig. l0), but pos-sessing dominantly ferric, rather than ferrous, iron,

Manganese activity and subcolcic garnet in skorns

Manganese also plays a part in the observedalmandine,/andradite relations, as previously sug-gested (Fig. 9). Moderately reduced tungsten skarnsare characterized by a high spessartine,/almandineratio, always greater than 1:l and typically 2:1,whereas strongly reduced skarns are characterizedby a low ratio, less than 1:l and typically l:2. Theratio of spessartine to almandine also increases withincreasing andradite content (Fig. 9). Manganese, ofintermediate size between ferrous iron and calcium,

6 SiO2q\afiz

2.6

ia.o- q

_ a

GaO

al ! | t rnd ln.-apca!al t l t t

540 THE CANADIAN MINERALOGIST

F e O + M g O l M n O ' F r ! Y - ' ! ! ! E

Frc. 10, Compositional triangle showing generalized compositions of mineral zones in tungsten skarns in terms of moletlo CaO, SiOr, and FeOr.5 + Alol.i + FeO + MgO + MnO. Numbers represent distinct mineral zones from mar-ble (1) toward intrusiveiontact (3). Primed numbers refer to oxidized skarns, unprimed to reduced skarns. Zoningfor both skarn types is from CaO-rich to increasingly CaO-poor assemblages moving away from the marble contact.

FeOr.U

appears to play a role in the stabilization ofandradirc-almandine combinations. Studies ofgarnet-stability relations (Novak & Gibbs 1971) in-dicate that increasing the average cation-size in thesite occupied by trivalent cations (i.e., increasing theFe3+ to AF+ ratio) results in increasing destabiliza-tion of small cations in the divalent cation sites (i.e.,decreases the stability of Fd+in the garnet). Thus,in the andradite-rich garnets of moderately reducedskarns, higher manganese contents are apparently re-quired to stabilize a given amount of ferrous ironin the garnet, and increasing the andradite/grossularratio in garnet increases the minimum spess€u-tine,/almandine ratio required for stability. Strong-ly reduced tungsten skarns, characterized by garnetcompositions with a low andradite content, pos$esshighly variable almandine/spessartine ratios that areindependent of andradite content (Fig. 9). These dataand those of Shimazaki (1977) suggest that alman-dine and spessartine contents can vary considerablyin garnet from skarns with less than 15 mole 9o an-dradite. The hyperbolic form of Figure I I is thus ex-plained by a co-operative substitution mechanism insubcalcic garnet. At moderately reduced oxidationstates, a higher spessartine/almandine ratio is re-quired as the andradite/grossular ratio rises, whereasat strongly reduced oxidation states, there is a limitedamount ofthe andradite component that can be ac-commodated in a subcalcic garnet.

scheelite deposition in skarns has been suggested byseveral workers, e.9., Einaudi (1977) and Shimazaki(1977).ln zones of reduced calcium-ion activity, thislow oxidation state is reflected by the presence ofsubcalcic garnet. In fact, the bulk of the world'sscheelite resources are found in subcalcic-garnet-bearing skarns (Fig. ll). In addition, of the sevenlargest known scheelite-bearing skarns (> 6 milliontons of ore), only one, the King Island deposit (Kwak& Tan l98l), does not contain subcalcic garnet.

Because of the retrograde and pressure-dependentsolubility behavior of scheelite, tungsten skarnsgenerally form at higher pressures and temperaturesthan those that characterize copper and base-melalskarns (Newberry & Einaudi l98l). Because theoxidation state during skarn formation is in partrelated to depth (Shimazaki 1980), deeper scheelite-bearing skarns should be characterized by morereduced mineral-assemblages, including subcalcicgarnet. Except where hematit ic host-rockspredominate or at depths transitional betweenscheelite and base metal skarns [e.9., Lost Creek:Collins (1977), Tungsten Hills: Newberry (1980)1,tungsten skarns were apparently buffered to oxida-tion states that would permit the formation of sub-calcic garngt. Thus, although scheelite skarns lack-ing a subcalcic garnet do occur, they are less com-mon and account for only a small fraction of knowntungsten reserves.

Oxidized and reduced tungsten skarns also con-Subcqlcic garnet and scheelite deposits trast in ore grades. Although grades are a function

of tonnage, mining method, geographic location, ex-The relationship between low oxidation state and tent of retrograde 196elilization of tungsten, host

SUBCALCIC GARNET IN SCHEELITE-BEARING SKARNS 541

rocks and other variables, the reduced (subcalcicgarnet-bearing) skarns are characterized, in general,by higher tungsten grades (Fig. 12). This contrast canbe found both between oxidized and reducedtungsten-skarn districts and also within districts thatcontain both types. In the Bishop district of Califor-nia, for example, the oxidized andradite - quartz -epidote skarns of the Tungsten Hills area (Newberry1980) contain grades of 0.2 to 0.5 Vo WO3@ateman et ol. 1950), whereas t}te moderately reduc-ed subcalcic-garnet-bearing Pine Creek and Adam-son mines are characterized by average skarn gradesof 0.66 to l.l9o WO, @ateman 1945). Similarly, inthe Mount Morrison quadrangle, California, skarnsadjacent to the Round Valley Peak granodiorite arein both highly eraphitic and nongraphitic marbles.Reduced skarns, such as the Hardpoint and Pappasdeposits, located in the highly graphitic BloodyMountain Formation, are characterized by assem-blages dominated by hedenbergitic pyroxene, pos-sess subcalcic garnet compositions (Newberry 1980)and contain tungsten grades in excess of l9o WO3(Rinehart & Ross 1964). Oxidized skarns, such as theScheelore mine, developed in the noncarbonaceousMount Baldwin marble, are characterized by epidote- qtJarlz - andraditic grandite t pyrite assemblages,do not contain subcalcic garnet compositions(Morgan 1975) and have tungsten grades of 0.25 to0.490 WO3 (Rinehart & Ross 1964).

The relationship between skarn type and tungstengrade might be explained in terms of different P, Tor X(CO) of formation; however, differences ingrade within skarns adjacent to the same pluton aresu'ggestive of differences due to local state of oxida-tion. A proposed explanation is based on the obser-vation that oxidized and reduced skarns are charac-terized by contrasting garnet/pyroxene ratios (Iable2). In particular, abundant pyroxene characterizesreduced skarns, and abundant garngt characterizesoxidized skarns @inaudi et al. l98l). If a skarn-forming system deposits hedenbergitic pyroxene [ow-ing in part to lower /(Orl that contains only 25mole Vo CaO, then more calcium is available in thefluid to deposit scheelite than if the fluid depositsandraditic garnet, which contains 37.5 mole 9o CaO.In other words, in skarns formed from relatively puremarble, the higher the pyroxene/garnet ratio, thehigher the average tungsten grade.

CoNcLUSIoNS

Garnet compositions intermediate between gran-dite and pyralspite are a characteristic feature of mosttungsten-bearing skarns of western North America.Only those tungsten skarns formed from hematite-bearing host rocks or at relatively shallow depths do

o.25 0 .50 0 .7s

Mll l lon tons WOs

Frc. ll. Approximate known tonnages (production r-reserves, expressed in tons contained WO3) of the threeclasses of tungsten skarns discussed in this paper. Ton-nage data from Einaudi el a/. (1981). Note that the bulkof known WO3 tonnage is from reduced (subcalcic-garnet-bearing) skarns.

not possess subcalcic garnet, Garnet compositions,however, are far more variable and andradite-richthan indicated by previous studies. In fact, the sub-calcic garnet of skarns can simultaneously containmore than 30 mole 9o andradite and 45 mole Voalmandine + spessartine component, and a garnetwith as much as 65 mole 9o andradite may containan appreciable subcalcic component. The occurrenceof such garnet compositions requires (l) a moderatelyreduced environment of formation, approximatelyat or below the pyrrhotite - magnetite - pyrite buf-fer, and (2) a relatively low activity of the calciumion in solution. Exceedingly variable almandine/an-dradite ratios occur in garnet compositions from agiven skarn, in part owing to variable activity of thecalcium ion and also to variable activity of themanganese ion. The presence of manganese is ap-parently required to stabilize the simultaneouspresence of appreciable almandine and andraditecomponents. In general, higher relative andraditecontents in garnet require higher spessartine/alman-dine ratios for stability.

Tungsten-bearing skarns typically possess sub-calcic garnet because they are characterized by

c

3sql qto ! -= oo oO Eg 5o

ool!txo

3ott: o: . o6 et tEo o> L

Eg6g

9 depos l ts

I I depos l tg

542 THE CANADIAN MINERALOGIST

a.Costabonne, Frar lce- gcheeloro, CA

Brownb Lake, Mont.Tungstsn Hll ls, GA

. Osgood Mntns., NVMll l Ctty, NV

Klng lsland, Tasmanla

Btack Rock, CA-E l Ja .a l l lo , Mer lco

Hll ton Creek, CA

0ol!Ixo

ltoo=ooE

oqt

oIto

tto(,Jttotriotro

o

MacMll lan Pass, Y-NWT.Garnet D lke , CA

Enerald, BCTungsten Jlm, lD -

Salau, France .Canada Tungsten, NWT

Plne Creek,Strawberry,

a l rbanks d lo t r . . AKSang Dong , Ko rea

development of this research. Early versions of thismanuscript were reviewed by J.M. Schmidt, D.C.Dobson, M.T. Einaudi and J. Trammell. Themanuscript has also benefitted frorn extensivereviews by R.F. Martin, L.T. Trembath and twoanonymous reviewers. Financial support from NSFgrant EAR 78-19588 and from a granr by the UnionCarbide Corporation is gratefully acknowledged.

I{EFERENCES

At-ses, A.L. & Rav, L. (1970): Correction factors forelectron-probe microanalysis of silicates, oxides,carbonates, phosphates, and sulfates. Anal. Chem.42, 1408-7414.

BarsuaN, P.C. (1945): Pine Creek and Adamsontungsten mines, Inyo County, California. Colif, J.Mines Geol, 41, 231-249.

Emcrson, M.P. & Pnoc.ron, P.D. (1950):Geology and tungsten deposits of the Tungsten Hills,Inyo County, California. Calif. J. Mines Geol.46,23-42.

CAC A

o.5 i:o i.b 2'.o 2'.a

Reported Average Skarn Grade (% WOs)

Flc. l2-, Reported average tungsten grades from skarn ore yersas skarn type. Notehigher average grades associated with reduced rather than with oxidizid skarns.Data from: Einaudi ef a/. (1981), Byers (1952), Bateman (1945), Rineharr & Ross(1964), Burchard (1977), Krauskopf (1953) and peabody (1979).

moderate to low oxidation states, owing to wallrocks,an intrusion-buffered oxidation state and a depth offormation distinctly different from other skarn types.Because such skarns possess moderate to high py-roxene,/garnet ratios, the presence of subcalcic garnetin a skarn is also linked to higher average tungstengrades. Thus, the presence of subcalcic garnet inskarns is not only of petrological and mineralogicalinterest, but also constitutes a potentially useful in-dicator of tungsten reserves.

AcKNowLEDGEMENTS

The author is grateful to the Union Carbide Corp.,Union Carbide Canada, Canada Tunsten, TeledyneTungsten, and the Anaconda Company for permis-sion to map and sample their deposits. Microprobeanaly$es were performed at Queen,s University withthe kind assistance of L.A. Dick and at Universitvof California, Berkeley, with the assistance of M.L.Rivers. Some of the analyses at Berkeley were per-formed by D.C. Dobson. Discussions with Drs. M.T.Einaudi, L.A. Dick and L.D. Meinert assisted in the

SUBCALCIC GARNET IN SCHEELITE-BEARING SKARNS 543

BrNce, A.E. & At-snr, A.L. (1968): Empirical correc-tion factors for the electron microanalysis of silicatesand oxides. J. Geol. 76,382-403.

Brno, D.K. & Hprcrsor, H.C. (1981): Chemical in-teraction of aqueous solutions with epidote-feldsparmineral assemblages in geologic systems. II.Equilibrium constraints in metamorphic/geothermalprocesses. Amer. J. Sci. ?Al, 576-614.

Bnocr, K.J. (1910) Mineralogy of the Garnet HillSkarn{ Calaveras County, Califurnia. Ph.D. thesis,Stanfqfd Univ., Stanford, Ca.

BnowN, P.E. (1980)l A Petrologic and Stable IsotopicStudy of Skarn Formotion ond Minerolization at thePine Creek, California, Tungsten Mine. Ph.D.thesis, Univ. Michigan, Ann Arbor, Mich.

Buncnanp, U. (1977): Genesis of the King Island(Tasmania) scheelite mine. .In Time- and Strata-Bound Ore Deposits (D.D. Klemm & H.J.Schneider, eds.). Springer-Verlag, New York.

Bysns, F.M., Jn. (1957): Tungsten deposits in the Fair-banks districr, Alaska. U,S, Geol. Sum. Bull. l0?A-1.

Corurs, B.I. (1977): Formation of scheelite-bearingand scheelite-barren skarns at Lost.Creek, PioneerMountains, Montana. Econ. Geol. 72, 1505-1523.

Cnsssev, G., Scxyro, R. & Woop, B.J. (1978):Thermodynamic properties of almandine-grossulargarnet solid solutions. Contr. Mineral. Petrologt 6f1,397-404.

Drcr, L.A. (1976): Mptamorphism and Metasomatismat the MacMillan pxs Tungsten Deposit, Yukon andDistrict ol Ma$enzie, Canada. M.Sc. thesis,Queen's Univ., Kingston, Ont.

(1980): z4 Comparative Study of the Geology,Mineralogy, ond Conditions of Formotion of Con-tact Metasomatic Mineral Deposits in the NortheastCanadian Cordillera. Ph.D. thesis, Queen's Univ.,Kingston, Ont.

DonsoN, D.C. (1982): Geology and alteration of theLost River tin-tungsten-fluorine deposit, Alaska.Econ. Geol. 77, 1033-1052.

Doocr, F.C.W. (1972): Yariation of ferrous-ferricratios in the central Sierra Nevada batholith, U.S.A.Int. Geol. Congr. (24th) 10, 12-19.

-, Srvurn, V.C. & Mrvs,'R.E. (1969): Biotitesfrom granitic rocks of the central Sierra Nevadabatholith, California. J, Petrology 10, 250-271.

Dur.n, D.P. (1980):, Petrology oJ the San AntonioScheelite Skarn, Baviacora, Sonora, Mexico. M.S.thesis, Arizona State Univ., Tempe, Az.

Easrvrax, H.S. (1979): Skqrn Genesis and Sphalerite-Pyrrhotite-Pyrite Relationships in the Darwin Mine,Inyo County Califurnia. Ph.D. thesis, StanfordUniv., Stanford, Ca.

Eowenos, A.G,, Barnn, G. & Canow, K.J. (1956):Metamorphism and metasomatism at King Islandscheelite mine. ./. Geol. Soc. Aust. 3.55-100.

Errauot, M.T. (1977): Petrogenesis of the copper-bearing skarn at the Mason Valley mine, Yeringtondistrict, Nevada. Econ. Geol. 72,769-795.

-, Mrrurnr, L.D. & NrweEnnv, R.J. (1981):Skarn deposits . Econ. Geol. 75th Ann. vol., 317-391.

Eruor, J.E. (1971): The Mineralogy and Geochemistryof the Tungsten Deposits of the Black Rock Area,Mono County, California. Ph.D. thesis, StanfordUniv., Stanford, Ca.

GaNcurv, l. (1976): The energetics of natural garnetsolid solution. II. Midng of the calcium silicate end-members. Contr, Mineral, Petrology 55, 8l-90.

& KnNr.nov, G.C. (1974): The energetics ofnatural garnet solid solution. I. Mixing of thealumino-silicate end-members. Contr. Mineral.Petrology 48, 137-148.

Guv, B. (1980): Etude gdologique et pdtrologique dugisement de Costabonne. Bur. Rech. Gdol. Minidres,Mdm. 99,237-250.

Kneusxopr, K.B. (1953): Tunsgten deposits of Madera,Fresno, and Tulare counties, California. Cald. Dep.Nat. Resources Spec. Rep.35.

Kwer, T.A.P. & TnN, T.H. (1981): The geochemistryof zoning in skarn minerals at the King Island(Dolphin) mine. Econ. Geol. 16, 468-497.

Lrp, D.E. (1962): Grossularite-spessartite garnet fromthe Victory mine, Gabbs, Nevada. Amer. Minerol.47. t4't-15t.

Lrou, J.G. (1973): Synthesis and stability relations ofepidote, Ca2Al2FeSi3Orz(OH). J. Petrology 14,381-413.

Moncer, B.A. (1975): Mineralogy and origin of skarnsin the Mount Morrison pendant, Sierra Nevada,California. Amer, J. Sci. 275, ll9-t42.

Nir"rnc, D. (1967\: The miscibility of the pyralspiteand grandite molecules in garnets. Mineral. Mag,36,389-402.

Nrwspnnv, R.J. (1980): The Geology and Chemistry ofSkarn Formation and Tungsten Deposition in theCentral Sierro Nevada, California. Ph.D. thesis'Stanford Univ., Stanford, Ca.

& Enieuot, M.T. (1981): Tectonic andgeochemical setting of tungsten skarn deposits in theCordillera. Ariz. Geol. Soc. Digest 14,99-112.

Nnwros, R.C. (1977): Thermochemistry of garnet andaluminous pyroxenes in the CMAS system. .fnThermodynamics in Geology (D.G. Fraser, ed.). D.Reidel, Boston.

Normsrnc, W.J. (1981): Geologic setting, petrology,and geochemistry of zoned tungsten-bearing skarnsat the Strawberry mine, central Sierra Nevada,California. Econ. Geol. 76. lll-133.

s44 THE CANADIAN MINERALOCIST

Novar, c.A. & Grsns, G.V. (1971): The crystalchemistry of the silicate garnets, Am.er. Minemi. St.79t-825.

prasooy, c.n. (1979): Georogy and petrorogy of aTungsten Skarn: El Jaralito, BaviacoTa, Sonora,Mexico. M.S. thesis, Stanford Univ., Stanford, Ca.

Rrcrwooo, P.C. (1968): On recbsting analyses ofgarnet into end-member molecules, Contr. Mineral.Petrology lE, 175-198.

RrNenanr, C.D. & Ross, D.C. (1964): Geology andmineral deposits of the Mount Morrison quadrangle,Sierra Nevada, California. U.S. Geol. Surv. pioi.Pap.3E5.

Saxare, S.K. (1973): Thermodynamics of Rock-Forming Crystalline Solutions. Springei-Verlag,New York.

Sorwpvrr, R.A. (1979): Evolution of Skarn at MonteCristo, Nevada. M.S. thesis, Stanford Univ., Stan-ford, Ca.

Tayron, B.E. (1976): Origin and Significance of C-O-H Fluids in the Formation of Ca-Fe-Si Skarn,Osgood Mountains, Humbolt County, Nevada.Ph.D. thesis, Stanford Univ., Stanford, Ca.

WrxcseLL, H. (1958): The composition and physicalproperties of garnet. Amer. Mineral.43, 595-600.

Wooo, B.J. & Nrcsorls, J. (1978): The thermo-dynamic properties of reciprocal solid solutions.Contr. Mineral. Petrology 66, 389-400.

Wnrcnr, W.A. (1973): Skarn Formation at Pine CreekMine, Bishop, Colifurnia. Ph.D. thesis, Univ. Catif.,Berkeley, Ca.

Received June 2, 1982, revised manuscript accepted Feb-ruary 3, 1983.

Znanmov, V.A. (1970): Skarns. Int. Geol. Rev. 12,Snrpreznn, H. (1977): Grossular-spessartine- 541-SSg,619-647,i\0-l lS.

zumanqrne garnets tiom some Japanese scheeliteskarns. Can. Mineral. 15. 74-80.

(1980): Characteristics of skarn deposits andrelated acid magmatism in Japan. Econ. Geot. lli.I 73- I 83.