Canadian MinerologbtVol. 29, pp. 995-1017 (1991)

ABSTRACT

Fe-Mg-Zn-Al alteration zones at the Linda depositare characterized by metamorphic assemblages thatinclude quartz, muscovite, pyrite, zincian staurolite,kyanite, gahnite, biotite, chlorite and sphalerite. Theseassemblages crystallized under amphibolite-facies condi-tions during syn-D2 peak metamorphism. Analyses ofsulfide. silicate and oxide minerals in low-varianceassemblages were used to balance metamorphic reactionsin the sysrem Si-A1-Fe-Mg-Zn-K-H-F-O-S. Chemo-graphic analysis of assemblages containing quartz'muscovite and pyrite defines a log lS, - log lO2)petrogenetic grid at fixed pressure, temperature, /(H2O)and /(HF). The mineral assemblages can be representedin the subsystem Mg-Zn-AI2. Reactions involving fourMg-Zn-N2 minerals define univariant curves; equi-librium compositions of coexisting minerals vary alongthe curues. Assemblages containing chlorite - kyanite orbiotite - kyanite can be modeled in a Zn-absent subsystemby isopleths of Fe end-member activity in continuoussulfidation-oxidation reactions. Fluorine occurs in sig-nificant amounts in biotite at the Linda deposit. As aresult of Fe-F avoidance, biotite is higher inMgl(Mg+Fe) than coexisting chlorite. The reversal inFe-Mg partitioning modifies reaction topology, andbiotite - kyanite is stable at higher /(S) than chlorite.Sphalerite geobaromelry indicates a maximum pressureduring syn-D2 metamorphism of about 5 kbar' Thecoexistence of Fe-rich staurolite + quartz and margarite+ quanz constrains the temperature of peak metamor-phism to about 550oC, consistent with the chlorite -

biotite - staurolite zone in pelitic metasedimentary rocks.In the metamorphosed altered rocks, the assemblagesstaurolite - anthophyllite and biotite - kyanite, whichwould normally crystallize at higher grade, were stabilizedby Zn and F, respectively, The presence of Zn and F, andthe reversed Fe-Mg panitioning between biotite andchlorite, invalidate the application of P-T grids derivedfor pelitic bulk-rock compositions. Metamorphic isogradsin the Snow Lake region require a reinterpretation that

.Lithoprobe Publication No. 219; Geological Survey of

Canada contribution number 57990.

METAMORPHIC PETROLOGY OF Fe-Zn-Ms-Al ALTERATIONAT THE LINDA VOLCANOGENIC MASSIVE SULFIDE DEPOSIT,

SNOW LAKE, MANITOBA"

EVA ZALESKI AND EDGAR FROESEGeological Suney of Canada, 601 Booth Street, Ottawa, Ontario KIA 088

TERENCE M. GORDONDepartment o/ Geoloey and Geophysics, University of Calgary, Calgary, Alberta T2N IN4

takes into accounl the composition of a hydrothermallyaltered protolith.

Keywords: sulfidation-oxidadon equilibria, .zincianstaurolite, gahnite, Fe-F avoidance, partitioningreversal, Snow Lake, Manitoba.

SovnaelnE

Les zones d'altdration du gisement Linda, au lac Snow(Manitoba), contiennent des assemblages m€tamorphi-ques i quartz, muscovite, pyrite, staurotide zincifere,kyanite, gahnite, biotite, chlorite et sphalerite' Cesassemblages ont cristallis6 dans le facies amphibolite lorsdu paroxysme m6tamorphique syn-Dr. Les r6sultatsd'analyses des sulfures, silicates et oxydes des assemblagesi faible variance ont 6t6 utilises pour balancer les rdactionsm6tamorphiques dans le systime Si-Al-Fe-Mg-Zn-K-H-F-O-S. L'analyse chimiographique des assemblages itquafiz, muscovite et pyrite d6finit une grille p6trogdn6ti-que en termes de log lS) - log /(O2) ?r pression,temp6rature, /(H2O) et IHF) constantes. On peutrepr6senter les assemblages de min6raux dans le sous-sys-teme Mg-Zn-AI2. Les r6actions impliquant quatremin6raux it Mg-Zn-Al2 d6finissent des courbes univa-riantes; les compositions des min6raux ir l'dquilibre varientle long des courbes. Les assemblages ?r chlorite - kyaniteou d biotite - kyanite peuvent €tre reproduits dans unsous-systCme sans Zn, par des courbes d'activite constantedu p6le ferrifbre dans les r6actions continues desulfuration et d'oxydation. Le fluor est present enquantites importantes dans la biotite de ce gisement. Vuel;instabilitd relative des liaisons Fe-F, la biotite est plusriche en Mg/(Mg+Fe) que la chlorite coexistante.L'inversion dans la distribution Fe-Mg modifie latopologie rdactionnelle, et la paire biotite - kyanite eststable ir une valeur de lS) plus 6lev6e que la chlorite.La g6obarom6trie fond6e sur la sphal6rite indique unepression maximale pendant le m6tamorphisme syn-D2d'environ 5 kbar. La coexistence des paires staurotideferrifbre - quartz et margarite - quartz limite latemp6rature maximale du m€tamorphisme d environ5506C, ce qui concorde avec la zone d chlorite - biotite- staurotide dans les roches p6litiques m€tas6dimentaires.

99s

996 THE CANADIAN MINERALOGIST

Dans les roches alt6r6es mdtamorphisdes, les assemblagesstaurotide - anthophyllire et biotite - kyanite, quiapparaissent normalement i un niveau m6tamorphiqueplus 6lev6, onr er6 stabilis6s par le Zn er le F,respectivement. La pr6sence de Zn et F, et I'inversiondans la distribution Fe-Mg entre la biotite et la chlorire,rendent invalides l'application des grilles P-T d6riv6es irpartir des roches ?r compositions p6litiques. Les isogradesmdtamorphiques dans la rdgion du lac Snow exigent uner6interpr6tation qui tient compte de la composition desroches dont la composition a 6td modifi€e par unealteration hydrothermale.

Mots-clds:. 6quilibres de sulfuration et d'oxydation,staurotide zincif0re, gahnire, incompatibilitd Fe-F,coefficient de distribution invers6, lac Snow.Manitoba.

IurnooucrroN

The conceptual framework needed to modelmetamorphic silicate - oxide - sulfide equilibria isstill evolving. In Fe-Mg minerals, progressivesulfidation or oxidation of Fe end-membersproduces progressively more magnesian composi-tions (Bryndzia & Scott 1987, Froese 1911, 1977,Mohr & Newton 1983, Nesbitt 1982, 1986a, b).However, at the Linda deposit, the presence of Znand F in altered rocks results in more complexphase-relations among the Fe-Mg phyllosilicates,Fe-Zn-Mg-bearing staurolite and gahnite, andZn-Fe-bearing sphalerite.

In this contribution, we have taken into accountternary Fe-Zn-Mg solutions in gahnite andstaurolite in the treatment of silicate - oxide -sulfide assemblages. The approach is initiallyempirical; we make use of electron-microprobedata on the composition of minerals in low-variance assemblages to balance reactions, and weuse chemographic analysis to develop petrogeneticgrids as functions of 52 and 02 fugacities, pressureand temperature. Mineral abbreviations are fromKretz (1983), with the following exceptions: Al-silfor aluminosilicate, bi for biotite, ch for chlorite,da for daphnite, gh for gahnite, hpo for hexagonalpyrrhotite, il for ilmenite, mg for magnetite, mpofor monoclinic pyrrhotite, ru for rutile, si forsillimanite, sid for siderophyllite, and xn forxenotime.

RrcroNer_ Gsorocy

The Linda deposit is located in the Snow Lakeregion of the Flin Flon - Snow Lake volcanic belt(Fig. l), in the Proterozoic Trans-Hudson Orogen.The volcanic belt is flanked by the Kisseynew gneissbelt to the north, and to the south, it is overlainby Paleozoic sedimentary rocks. The followingbrief summary of the regional geology of Snow

Lake (Fig. l) is based mainly on Froese & Moore(1980).

Supracrustal rocks have been divided into twogroups, Amisk and Missi. The Amisk Groupencompasses a suite of intercalated mafic and felsicmetavolcanic rocks, and overlying metasedimen-tary rocks, mainly greyrnackes and shales. Amiskmetasedimentary rocks (ornamented in Fig. l) havepelitic bulk-compositions and commonly containmineral assemblages definitive of regionalmetamorphic grade. The Missi Group, which wasdeposited on Amisk Group rocks, consists mainlyof metamorphosed lithic arenites.

The Linda deposit is one of several volcanogenicmassive sulfide deposits in the Snow Lake region.It is subeconomic and contains massive anddisseminated pyritic mineralization with minoramounts of pyrrhotite, sphalerite and chalcopyrite.Economic Cu-Zn deposits in the vicinity includethe mines at Anderson Lake, Stall Lake and Rod.The deposits are hosted by Amisk felsic volcanicrocks and are associaled with extensive zones ofsynvolcanic hydrothermal alteration. The alteredrocks are characterized by assemblages of am-phibolite-facies metamorphic minerals that resultedfrom the regional dynamometamorphic overprint-ing of altered protoliths (Bailes 1987, 1988, Froese& Moore 1980. Walford & Franklin 1982. Zaleski& Halden 1988).

Froese & Moore (1980) proposed three deforma-tional events. D1 produced tight to isoclinal F,folds. The Mcleod Road thrust faulr (Fig. l),although locally showing evidence of late reactiva-tion (Galley et al. 1988), is a probable pre-Dr orsyn-Dt fault. D2 produced northeasterly trendingflexural folds. D3 deformation was restricted to thenorthern part of the Snow Lake region.

Metamorphic assemblages in the Flin Flon -Snow Lake belt show a transition from a low-grademetavolcanic terrane in the south, to the high-gradeKisseynew gneiss belt in the north (Froese &Gasparrini 1975, Froese & Moore 1980). In theSnow Lake region, metamorphic isograds delineatefour metamorphic zones characterized by theassemblages: chlorite - biotite, chlorite - biotite -staurolite, biotite - staurolite - sillimanite andbiotite - sillimanite - almandine (Fig. l). Progradeisograd reactions (listed in Fig. l) were inferredfrom parageneses in Amisk pelitic rnetasedimentaryrocks.

The reaction assemblage for the biotite -sillimanite isograd (chlorite - staurolitealuminosilicate - biotite - muscovite) is present inthe altered rocks at the Anderson Lake and StallLake mines. The parageneses in altered rockscommonly resemble those of metapelites; hence thebiotite - sillimanite isograd was mapped on the

METAMORPHIC PETROLOGY OF THE LINDA DEPOSIT' MANITOBA 997

I? !

6 !l l

! t3iB i . S i . A lm

1. chlorlte + almandlne + muscovlte* blotlte + staurollte + quartz + H2O

2. chlorlte + staurollte + muscovlte + quartz

* blotlte + alumlnoslllcate + HrO

3, staurollte + muscovlte + quartz +

blotlte + alumlnoslllcatg + almandlno + H2O

PARAGENEgES

chlorlte-blotlte-staurollto-muscovlted€lmandln6

chlorlte-blotlte-staurolltE-sllllmanltEimuscovlto

Bi'st'Si V",, sto, ...'

'x"r ffip ch .B i

Frc. l. Location (inset), regional metamorphic geology of Snow Lake, and isograd reactions (inset) based on Froese

& Moore (1980). Merapelites belonging tb the Amisli Group are ornamented. Stars mark the locations of assemblages

of lower grade in peliiic rocks north of the biotite - sillimanite isograd. The location of subsurface cross-section

C-C' of rhe Linda deposit (Fig. 2) is indicated by the bar.

basis of the first appearance of aluminosilicate,irrespective of the protolith (Froese & Moore 1980).Assemblages of lower grade (indicated by stars inFig. 1) are present in metapelites to the north ofthe isograd. Chlorile in these assemblages wasregarded as a retrograde mineral by Froese &Moore.

There is a dichotomy in the distribution ofaluminosilicate polymorphs in the region; kyaniteis present at Stall Lake, and both kyanite andsillimanite at Anderson Lake, whereas sillimaniteis present in metapelites. The erosional surface wasinterpreted to expose an isobaric prograde sequencecorresponding to the intersection of the biotite -

aluminosilicate isograd reaction with the kyanite -

sillimanite phase boundary (Froese & Moore 1980).

GEOLOCY OF THE LINPE ANEE

The pyritic mineralization of the Linda depositis hosted by felsic metavolcanic rocks that have

undergone synvolcanic Fe-Mg-Zn-Al alteration(Fig. 2). The alteration is disposed in two discretezones, designated as "proximal" and "distal" withreference to their relative distance from massivesulfide mineralization, and without lithofacialconnotation (Zaleski 1989). The deposit is over-turned, with the alteration zones overlying themassive sulfide bodies.

The distal zone defines a relatively homogeneouszone of Na and Ca depletion; the most commonrock-type is a feldspar-absent staurolite-bandedgneiss. It is enveloped by felsic rocks containingcalc-silicate minerals, mainly epidote. The proximalzone is heterogeneous, encompassing graphitic andmargarite-bearing metasediments, sulfidic mus-covite-rich metasediments, massive sulfide bodiesand altered felsic rocks. There is a single occurrenceof quartz - staurolite - anthophyllite - gahnite rockin an enclave in the largest massive sulfide body(Fig. 2, ll3l). Both the distal and proximal zoneshave an irregular Zn-rich core characterized by

m A l te red rock/ "

. / s tou ro l i t e -gohn i t e

f f i :Tg Mossive & semi-mossiveFn$: f i sutohide minerol izot ion

Ef Grophit ic metqsediments

f j . - I l Mof ic to in termedioter r ^ " ' r v o l c o n i c r o c k t| | Fets ic volconic rocks z

998

muscovite schists with zincian staurolite andgahnite.

In the Linda area, there is evidence of at leasttwo episodes of deformation (Zaleski 1989, Zaleski& Halden l99l), and the planar tectonic fabrics,S1 and 52, are pertinent to this study. S, is definedby metamorphic segregation layering, flattening ofprimary features (e.g., clasts, quartz phenocrysts),mineral schistosity and trails of inclusions. S, ismost commonly a crenulation cleavage, but it alsois defined by the orientarion of (001) planes ofbiotite. Porphyroblastic minerals are mainly syn-kinematic with Dr. The observations indicating thisare: skeletal crystals with habits mimetic after .S,

THE CANADIAN MINERALOGIST

Ftc. 2. Cross-section C-C' of the Linda deposit looking southwest, with samplelocations. Sample ll31 lies in an adjacent cross-section and was projected ontothe C-C'plane. Horizontal and verrical scales are equal.

and S, cleavages but showing some misorientationwith respect to ,S2 in the matrix, and porphyroblastscontaining S, inclusion trails and showing crystal-lographic orientation parallel to Sr. Garnet inaltered rocks shows dimensional orientation andinclusion trails parallel to ,S1 cleavage, indicatingpre-D, and probably syn-D, growth of garnet.

Two generations of kyanite can be distinguished.Syn-D, kyanite shows crystallographic and dimen-sional orientation parallel to S, schistosity, andtransposition and strain related to development ofthe 52 fabric. Syn-Dz kyanite occurs as skeletalcrystals mimetic after ,S, and S, fabrics, aspreviously described for syn-D, porphyroblasts in

METAMORPHIC PETROLOGY OF THE LINDA DEPOSIT, MANITOBA 999

general. Syn-Dr kyanite is not associated withbiotite; syn-D, kyanite is associated with biotite,typically a nearly colorless phlogopite.

The Linda deposit lies in the chlorite - biotite -staurolite zone of regional metamorphism as basedon parageneses in metapelites (Fig. l). At thedeposit, parageneses in altered volcanic rocksinclude muscovite - chlorite - biotite - garnet -

staurolite and muscovite - chlorite - staurolite -

biotite - kyanite; these are the reaction assemblagesfor the biotite - staurolite and biotite - aluminosilicate isograds, respectively. The assemblagestaurolite - anthophyllite, normally stable withinthe biotite - staurolite - sillimanite zone (Froese &Moore 1980), also is present. The relationshipbetween the altered rocks and regional metamor-phism is one of the major questions addressed inthis study.

MINERAL CHEMISTRY

Method

The selection of samples for electron-microprobe

analysis was aimed at representing the variety oflow-variance silicate - oxide - sulfide assemblagesin altered rocks (Fig. 2, Table l). In particular,staurolite-bearing parageneses were represented.Mineral analyses were carried out on a JEOLSuperprobe 733 at the Mineral Sciences Division'Canadian Museum of Nature' Ottawa. Mineralcompositions are reported here for staurolite,gahnite, garnet, sphalerite, biotite, muscovite andchlorite (Table 2).

Mineral compositions were determined bywavelength-dispersion spectrometry (WDS) usingan accelerating potential of 15 kV' a beam currentof 20-25 nA, and a beam diameter of 10-16 pm.A focused beam was used for some very finedisseminated grains of sphalerite. Mineral stand-ards were used, except for Zn in phyllosilicateminerals, for which synthetic BaZnGeOa was usedas standard. For the analysis of phyllosilicates, thefollowing mineral standards were used: chlorite forSi, Al a;d Mg, biotite for Ti, Fe and K, chromitefor Cr, willemite for Mn, diopside for Ca,sanbornite for Ba, albite for Na, fluoriebeckite forF, and tugtupite for Cl. For staurolite' gahnite,

TABLE 1. SUMMARY OF MINERAL ASSEMBLAGES

a. DISTAL ALTERATION ZONE

Saruple Assernblages Accessory Minerals

22-90r

22-957

22-996.5

34-799a

3+799b

qtz, ms] ch, \r, st' pyr sp.

qtz, ms, st, ch, pY.

qtz, ch, ms, st, mgr PY.

qtz, ms, glr, et, py, sp.

qtz, ms, ghr pyr ep.

pll ap, aln, tur, zrn, nrr ill

rnrg, grt, tur, spr ap, alnr mg, il, rut mnzr )<rl'

il, ap, aln.

tur, chl ap, ccp, sif m.

tur, ccp, nr, "hf

b. PROXIMAL AI,TERATION ZONE

Sample Assernblages Accessory lWnerals

22A-t625

22A-1638

22A-r752

228-1723

uc-L377

34C-1453

uc-I747.6

34-1500

24.1L31

qtz, ms, ch, st, grt, bit PY' sP.

qtz, ms, st, ky, PYr Bhr sp.

qtz, ms, PYt kyr ch, et, bi.

qtz, ms, st, py, gh' bit sP.

qtz, ms, *, py, kJ, bi, ch, sp.

qtz, ms, sfi, ghr py.

qtz, ms, \r, st, bir Pn hPo.

qtz, Brr ll5r, py, hpo' bit trre' sP.

qtz, st, edh, py, gh, mg, hPo.

*pof t*o aln, ap, zrnr ccpr ru.

tur, ap, zrn, Por ccPt ru.

pl, tur, ru, aPr ccP.

nrrg, ap, tux, znt, gnr ru' grn, "nf

cal9

pl, tur, ap, zlnt ru, :cr, sil

rttpol nil "hf

ap, zrn, aln, ru, il, xn, mnz.

"p, .npof

"hfccp, ru.

*pof -"g, chf apy, ccp, tur.

pt, mpof ccp, bif apy, sil

I

3Analyzed minerals in boldface.

Retrograde.

2 Inferred from altered pseudomorphs.a Not part of the equilibrium assenrblage.

a

o

roc

ott

ot9q

N

oo

N

{cnN

t9

{t99'

{p

q

a

q

otr&

15

C'-f{

-l'l-l

g

THE CANADIAN MINERALOCIST

gsg r E€Bt$$o E *x39y F.Fpe

I ' I ' P P Pr 0 c r g r o cl t s | 5 ( o O O{ o o o o

c t o o c eo o e e 0| € \ 9 r € O ( 'r l q l e O F

v v v v v

t9| L r t s t s F O O Og o @ r t s | | o oC O - l g l o g oO O O C t - l ( l €

o o o o 0o o o o oo t 9 F ( : or F O O ) O H

v v v v v

g J | Fr t s ( O t s O O O t 9 O O qC F F C ! E H O O H | | {g r q r | t s C - l C l . O l J o 9 ,

o e e e o c o ( ] 0 0o F o 9 0 0 F 9 r c , 9 'o r @ o o t 9 c t o c F g r o

v v v v v v v v v v

o t | |r t s @ F O g e P O O qCrl aD Ctl O F -il 9, Cl rl OC r C l 6 O i l O N g t - 1 t s

o c o o o o c o o oo t s o 9 0 0 0 0 0 t 9N C n 6 O t s \ t a r c r | F g rv v v v v v v v v v

g , | sr t s - i l a 9 O ( 3 e C 6 O | |r F g O O H g t ( ! N C P O !g r g t - l C t 9 @ O O N @

e N o o o o o N o t r 9r u 9 r c t s g r 6 g r p o

v v v v v v v v v v

s | |S o N o o e t s q o oq F O O C g ) t s | | N - l1 9 O g e \ r F - l € ! ( O 9 ,

P P 9 9 P P P t s P t so s t 9 o ( t F F o o oq € | 5 O ( l i l O O O N

o o o o o o o c c o( ' q r O F o o O o e - lS O U H 6 6 O H A q

v v v v v v v v v v

o 5r S O F e e O F O O e ^6 ( € O O F q N { 9 r N( 3 o ) O t 9 q t 6 O - l C r ( !

e o c o c o c e o oc H o o o o o | s o aq t - i t € . 9 N 0 o - . 1 t 9 c r ! | |

€ t s c t g ,@ T E O O P C O O O O H( o C r t 9 O ( : O \ J O @ O Ot 9 t s q J | | € O r l - 1 o ) 6 ( D

o o c o o o o o e o ca o o o p o o o t s o t 9t s t 9 g ) | F 9 r e $ - . l O P N

v v v v v

o g r l |O r $ O H O O O H { O Ot 5 C r O | F C O O N e C O @O € C l - l g O P O C t O t \ t

e o o o 0 0 c ) 0 0 0 0o o F 0 o o c c q r o t 9@ q t ( o O ! O C n O - ! F ! , e

v v v v v v v v v v v

o o so | F ( o H o o o H c D o o| | C F o o o o o o { o oC F C l - l C ! O { S O O F N

c c o e o o c c o o oo o N o ( t e o € N o H| | ' H N - i l o - l q r o o ! 9 r c f ,

v v v v v v v v v v v

\ !A F to-

s

H O H

a a : -t s . o o

q

O O ^

t s t s o l€ @ t s

P O: ^ : - ( o6 ; P( o t s @

t9O O ^

F F \ ,

H O ): ^ : , @E E P- l g t @

o o ^

o 9 I t@ @ t s

P

, F e g O C ,H H 1 9 ( t - i lo N 0 r s oN @ N O N

o o o c oH C t a 9 0 g ,o ( o o c t | 5

vvvvv

s o € ) o 9 tF H H O . r lO | t s r F A C !t s ( o o { €

o c r c o o\ 9 t 9 9 r | . g t

r9F Fo o0 ( '

o o oe e a@ o aCt g, -l

H q c n

ts(o

c(t

@

N| 5 | | t s P O E Oo 0 6 9 r ( ! o ( 'o e € F ( t | | o oe O H t s r t s C g !

9 r $ o o oq r F ( 0 0 t sv v v v v

N| F r t s H F O O €€ o € q r c l o oo o { @ @ o oo o q r o e o H

o o o c oc r o o ) o oO - l C o n 9

v v v v v

o 5| 5 @ o o o o F o ) o c rg t O t ( o - . l H @ q O | | t sO @ O q O O O ! ( t t s - l

No

(0

N15

o@

t ' P€ c{ N6 r F

{ - l-l -.1

5 ( ) 0 0 9 ,O H t s ( t . . 1P N N N g '- l O t O @ €

e o o o oo o o o ot \ , q r t s O r F{ 0 c r t s o @

r t s O O ( 5 C ,

t s P H O - i to c r - . t A 9 rq O A i F

N

F F I ' I ' P P PO O @ r 5 t \ r P oO ( f ! t s - . l g r o oC O F - l t g o o

o o N@ 5 e

t s € t o g J $b ; - p i P 9 P 9 : - , P P 9 9C r | | C t C F A - l o | 5 N ( 3 @ o - l- l 9 t t 9 O I H O | | @ N C r € O C t ,

T F ( 3 C ' O C ,o N e o oa 9 r o ( ' | 5o € O C - l

F H O O O@ c i P e oC - i l ( O C \ ,- l O ( O O t C !

r ( DO t s5 0

o o

{ - !

| 5 e o o | 5 0 r at s e o o H o H6 q o o F o g ,O I | L O - l @ O n

o o o e o c cc c o o e c oN O O O H F F| F | . . l @ . 9 F g t C r r

| 5 o ' c o cF F P C O9 r c a c r g r F( o o t s c t t f ,

o c o o c( : o o e oH H C O ON O @ t 9 @

| t s O r F u ot s { o @ 9 ,

v v v v v

F O@ o =o o : -c r q o

Fo c ^: - : ^ or F A C rH H 9 '

F P E Ot s q t c 9 t sc r | | ( ! &

$

N

oCA

oe{eoo

o

i9

oqt

|ts

N

At9N

Oo|ts

oN(t

t9| 5 l t s B F O 0 C l t s C C O 9 OO O € O | A O O O H H O T I€ o q N a € o o o N @ g r qO ( t t s - I H O | | t g € ! 4 9 t s 9 ,

c ) e o o o o o o e o9 e o o o o o o o oF e H O O H O O O OO C a - l O $ ( I C r ! € ( , O

v v v v v v v v v v

@ N

o o c et s o c ) Pq @ g @

t9

c r o

Nt t s | |o ( 3

N| | | F t s o o F c l

- l Q H C r oo g r 6 0 cg r @ o o o

o ( ) o o e

N e F t 9 ( f@ @ o o o

v v v v v

Nr F r F F t s g O O

O C - l g r q t O OO O H | F - . l O eO O - l \ ) F O t E

a 0 c r c oo o o c oP | | H O O( o t 9 C r o 0 ,

vvvvv

F OO PO Pc r q

o oo 0

: . 9€ cc a | FcN ct

P 9o oo o

1000

+ a

= F

F oi lgllr +i o: } n0'.to t

HHF +EE.x E

;o

* i t

5 [. F eg l lS ! lt g

rTn x

; c* E

a + =

F.t

!ee5'pq

o

METAMORPHIC PETROLOGY OF THE LINDA DEPOSIT' MANITOBA

TABLE 2b. BIOTITE COMPOSITIONS

l00l

Snmple 1625 (5) ' 1752 (2) 1723 (5) 1500 (6) 1377 (5) 1747.5 (6)

SiO: 37.87 (0.35)Tioz 1.02 (0.05)Alz og 19.19 (0.15)FeO 10.75 (0.15)l\dno 0.24 (0.01)lvlsO 17.52 (0.15)BaO 0.00 (0.00)CaO 0.00 (0.00)Na2O 0.41 (0.02)KzO 9.24 (0.43)F 1 .67 (0 .15)HzOt 3.37 (0.08)

Total| 100.5? (0.72)

si 6.452 (A.O22)rYAt 2.54s (0.022)v rA l o .zo9 (0 .011)Ti 0.111 (0.004)Fe 1.2e4 (0.022)1\4n 0.029 (0.001)I\& 3.760 (0.042)site 5.903 (0.044)

38.40 (0.11)1 .00 (0 .01)

18.92 (0.0?)e.73 (0.14)0.20 (0.01)

1e.34 (0.14)o.oo (o.oo)0.02 (0.02)0.38 (0.02)8.57 (0.38)2.83 (0.08)2.88 (0.05)

101.06 (0 .07)

5.460 (0.006)2.540 (0.005)

0 .631 (0 .003)0 .106 (0 .001)1 .15? (0 .014)0.023 (0.001)4.0e8 (0.025)6.016 (0.042)

o.oo0 (0.000)o.oo3 (0.003)0 .105 (0 .005)1 .555 (o .o?o)1 .662 (0 .080)

1 .273 (0 .037)2.727 (0.037)

22.727

37.15 (0.61)1.27 (0.08)

18.7e (0 .12)14.32 (0.150.17 (0 .03

15.61 (0.250.00 (0.000.03 (0.020.28 (0.078.72 (0.221.51 (0.093.39 (0.05

41.68 (0 .15)0 .34 (0 .02)

L7.21 (O.25)1.30 (0.08)0.05 (0.04)

25.25 (0.15)0 .16 (0 .01)0 .01 (0 .02)0.33 (0.02)8.73 (0.47)3.53 (0.13)2.66 (0.05)

ee.76 (0.6e)

5 .766 (0 .016)2.234 (0.016)

0.573 (0.037)0 .035 (0 .002)0 .151 (0 .010)0.006 (o.oo5)5.207 (0.022)5 .971 (0 .018)

o .ooe (0 .001)0.001 (0.002)0 .087 (0 .005)1.541 (0.078)1 .639 (0 .082)

1.545 (0.054)2.455 (0.054)

22.455

3e.82 (0 .06)o .?2 (0 .01)

18.50 (0.089.04 (0.030.28 (0.02

19.75 (0.0E0.00 (0 .000.00 (0 .000.41 (0.018.18 (0.212.18 (0 .153.23 (0.08

0.11 (0 .0215.?2 (0.150.00 (0 .000.00 (0.00o.4o (0.058.e7 (0.271.64 (0 .133.3S (0.07

87.34 (0.24)1.18 (0.06)

18.6e (0 .21)13.79 (0.23)

BaCaNaKsite

FHto

0.000 (0.000)0.000 (0.000)0.114 (0.006)r .6e7 (0.071)1.811 (0.075)

0.75e (o.o6e)3.241 (0.06e)

23.24L

100.62 (0 .52)

5.423 (0.060)2.577 (0.060)

0 .658 (0 .033)0.13e (0.00e)1.74e (0.024)0 .021 (0 .003)3.3e6 (0.070)5.e63 (0.060)

0.000 (0.000)o.oo5 (o.oo3)o.o7e (0.01e)1.624 (0.035)1 .70e (0 .050)

0.6e5 (0.043)3.305 (0.043)

23.305

101.17 (0.26)

5.606 (o.oo5)2.394 (o.oo5)

0.676 (o.oo5)0.076 (0.001)1.064 (0.003)0.033 (0.002)4.143 (o.o2o)5.ee2 (0.018)

0.000 (0.000)0.000 (0.000)0.112 (0.004)1.46e (0.035)1.581 (0.037)

0.e71 (0.070)3.02e (0.070)

23.O29

100.4e (0.63)

5.453 (0.028)2.547 (0.O28)

0 .670 (0 .028)0 .130 (0 .006)1 .684 (0 .025)0 .013 (0 .002)3 .420 (0 .041)5 .e17 (0 .025)

o.ooo (o.ooo)0.000 (0 .000)0 .113 (0 .014)1.671 (0.043)1.784 (0.03E)

0.757 (0.062)3.243 (0.062)

25.243

I Calculated frorn stoichiornetry. I Cl, Cr and Zn not detected.

garnet and kyanite, the following standards wereused: almandine for Si and Fe, titanite for Ti,gehlenite for Al, nichromite for Cr, diopside forMg and Ca, tephroite for Mn, willemite for Zn,and albite for Na. For sphalerite, the followingstandards were used: sphalerite f.or Zn and S,arsenopyrite for Fe, tephroite for Mn, cuprite forCu, and Cd-bearing apatite for Cd. Data wereprocessed with a conventional ZAF correction.

During calibration, counts on standards formajor elements were collected to a minimum loprecision of 0.50V0. Counts on samples werecollected to a lo precision of 0.5090, or formaximum counting times of 25-40 seconds. Theresults of each analysis were recorded with thecumulative lo standard deviation for both standardand sample. Elements for which the standarddeviation of sample counts exceeded 25.5090(relative error) were considered not detected. Cl wasnot detected in any samples. Detection limits for Fwere 0.3990 in chlorite, 0.44V0 in biotite and 0.4690

in muscovite. Detection limits for Zn were 0.160/oin chlorite and less thar' 0.240/o in biotite.

Results of chlorite analyses tend to have slightlyhigh analytical totals; these average 101.26 t'o102.00s/o after the addition of calculated H2O.Biotite analyses also tend to have high sums, from99.76 to l0l,l70/o. This is probably due to asystematic error introduced during calibration onthe standards and may be related to either thechoice of standards or to inaecuracy in the reportedcompositions. Chlorite and biotite standards wereused for major elements to minimize the differencesin matrix between the standards and the mineralsanalyzed. The element(s) in which the systematicerror occurs cannot be definitely identified, andthis will contribute to the uncertainty of theanalyses. Interpretations relying on relative mag-nitude should not be affected.

Recasting of analyses (except in the case ofstaurolite) to structural formulae followed themethod of Deer et ql. (1966). The proportion of

r002 THE CANADIAN MINERALOGIST

TABLE 2c. CHLORITE COMPOSITIONS

Sample e01 (6)' 1625 (4) 1762 (6) 1377 (5)

SiOzTiOzAlrOgFeOA/tnOIvIgOZrONa2OKzoFrrz ot

25.30 (0 .40) 26.300.07

24.52!4 .73

0.4322.72

0.1 10.020.160.00

L2.25

1 0 1 . 2 6

26.2e (0.22)o.o5 (0.03)

24.34 (0.28)15.23 (0.46)0.38 (0.02)

23.20 (0.2e)o.o0 (0.00)0.00 (0.00)o.o2 (0.04)0.36 (0.27)

12 .11 (0 .10 )

101 .82 (0 .50 )

5.135 (0.033)2.865 (0.033)

26.59 (0.310.05 (0.03

24.46 (o.2L13.01 (0.180.50 (0.01

24.5r (0.340.00 (0.000.00 (0.00o.o3 (0.05)o.3o (0.25)

L2.22 (0.r0)101.54 (0 .74)

5 .157 (0 .028)2 .843 (0 .028)

2.75t (O.025)0.008 (0.004)2.110 (0.035)0.082 (0.002)7.084 (0.041)0.000 (0.000)o.ooo (o.ooo)0.006 (0 .013)

12.042 (o .oo6)

0 . 1 8 5 ( 0 . 1 5 1 )1 5 . 8 1 5 ( 0 . 1 5 1 )

35.815

Total{ 102.00 (0.30)

0 .00016.000

36.000

0.04 (0.0324.44 (0.2820.59 (0.16

0.30 (0.0219.32 (0.210.00 (0.000.00 (0.000.01 (0.000.00 (0.00

12.00 (0.04

0.53)0 . 0 1 )0 .60)0 . 1 8 )0 . 0 1 )0.43)0 . 1 1 )0 . 0 3 )0 . 1 8 )

(0.00)(0 .12 )(0.e0)

s i 5 .055 (0.075) 5.156 (0.060)rYAr 2.945 (o.o?5) 2.844 (0.060)vrAt 2.BLo (o.o2zTi 0.006 (0.004Fe 3.441 (0.035Ltn 0.051 (0.004I\& 5.?52 (0.058Zr 0.000 (0.000Na 0.000 (0.000K 0.002 (0.000site 12.062 (0.044

2 . 8 2 3 ( 0 . 1 1 60.010 (0 .0012.415 (0 .0440.072 (o .oo16.642 (0 .1870 . 0 1 5 ( 0 . 0 1 60 . 0 0 7 ( 0 . 0 1 10.038 (0 .045

12.023 (0 .054

2.741 (0 .0390.007 (0 .0052.488 (0.0740.063 (0 .0036.753 (0 .0870.000 (0 .0000.000 (0 .0000.00 i l (0 .009

12.057 (0 .019

FHto

0.00016.00036.000

0.223 (0 .164)15.777 (0.164)

35.777

t Calculated froru stoichiouletry.* Cl, Cr, Ba and Ca trot detected.

H2O was back-calculated from stoichiometry, afteradjustment for F content. Staurolite formulae werenormalized to (Si + Al) = 25.53, after theproposed scheme of Holdaway et al. (1986, 1988)for staurolite that formed under reducing condi-tions. Oxygen was assumed to total 48 atoms, andexcess anionic charges were balanced by H+. LirOwas not determined, and the weight Vo of HrO wasback-calculated from the charge balance. Theassignment of cations to structural sites, based onHoldaway et al. (1986), includes 0.250 Fd+ at theoctahedral A(3A) sites, (Mn + Fe2+) : 0.250 atthe U(1) site, and Mg, Zn, Ti and the remainingFe at the tetrahedral (Fe) site. The weight percentsof Fe2O3 and adjusted FeO were back-calculated.This method produces oxide totals averaging 97 .52to 99.5890.

In general, the samples achieved equilibrium atthe thin section scale, as inferred from the lowJevelrandom scatter of analytical values. Most analysesare reported as within-sample means (1.e., withinthin section) with 1o standard deviations from the

mean (Table 2). ln most cases, mean values wereused in balancing model reactions. Staurolite andgahnite show minor differences between core andrim compositions. In cases of systematic variation,average rim compositions were used. Garnet showsstrong Fe-Mn zoning, and rim compositions wereused to balance reactions. Sphalerite formulae werenormalized to S = I before calculating balancedreactions.

Staurolite, gahnite and spholerite

Staurolite and gahnite compositions span a widerange of Fe-Zn substitution, with minor variationin Mg (Fie. 3). The Zn content of staurolite is partlya function of bulk-rock composition; the composi-tional gap (between samples 1453 and l72l) isprobably due to a gap in the bulk-rock composi-tions of the sample suite. The Mg content ofstaurolite and the Zn/Fe ratio of gahnite are mainlyfunctions of 52 and 02 fugacities, as inferred fromthe coexisting Fe-Ti oxide and sulfide minerals.

1003

o

O

zuOs

o

a

v

o

o

aa

€

a

q

aoN

€€

O

€

€N

METAMORPHIC PETROLOGY OF THE LINDA DEPOSIT' MANITOBA

o r oD NF O

o 6N N

F O

r o o o c o Fi ! o ( o o No o o o i: : : : i

N . i r o N @N O i O H

( 4 - F o mO A F S i- N O O 6

o N o o a a

F i O €c r i ) i o

! o s i F Q oo o o i o

€ so 6

N N

o d. q oN O

N O

o $6 0O F

i l o! ' o

6 N

r o

6 6D N

F A

c i o

O N

n r {

O H

r o

o m6 6

€ NF Nn $

t s o

n | r i

F O

N €N F

F O

F

co

r

6

r

€q

N

Cd

F

.;N

r

F€qi

6 A

6 NF F

o iN N

F A

F Mo $N O

O €

o i l$ a

o nv ) NF A

o 46 N

F O

F 6 € L € a € O F€ $ F ; o $ o D F

X i d g i J d c ; c i J i

l - F O @ € c l m o @@ S N d d O N N O

d c ; d i . j d i c i d . io t n

F N 6 c ' ) $

o o e o o

i O O L O O N H dF m @ ; < H o n N @

d c ; d ' j d i 6 i r j - iN N

$ N * F b6 A N i i

< @ F 6 F m o 6 6O O N i $ N O i O< i c i s . i v j d - . i F i iN 6

6 N t - L O O n n ir $ N i o n o H $

<t c; di --: tj ci oi <t .;N D . .

a O < r 6 m . + O H mF S n l ; s n i o €d d i . J r j c ; c i d iN r O

- € @ F m <= o a N @ o. j d c ; c . i i ' i

O t - O !H O O

F N Oo r c s

6 D €i o N

o o €N O 6

D O6 aO O

F

O O F N H( o F ! . o oo r l . o - 9

o H o i a

d . + c a F Fi N O O O

t f r a r 6 i s@ n r o o €o $ o o €c ) i H i o

F 6 0 i oo F o r < i oo o o o i

- i D 6 N6 0 0 4 Do o F r q < .o d o i c o

( o 6 6 . $ r f( ) N N i S

o N r a o o( t € F o $o s c o N oo H o i o

€ o Nd O . { ,

F O

N €

ri o o o

F o F n ) dO Q i O H

o n FN O i i

O o o ii o m a

m o € € oN O O N O

o m 6 s Fi O O O Q

T i O N FN @ O d r

O Q i H H

aoN

N O

O O

O , F N F No € F o 6

oN

€ D F a n i F $ to $ o = o F o o o

F c t d i . i a d o i . ' r ' jN 6

N 6 d N N ( o . O S

c ; o d c t c i c t < ; c t

r o o o N €O t s F N Da 6 o a 6o i o o m

N O 6 i r O- a C ! 6 6( > i o o d: : : : :n o F o r oC O O F d ' No a o r $ nd - c t i d

6 C . $ F

o i Q i

O O N Fa < r N o

€ . a o o €O F N d Oo o a o i

o a 4 d NO O N 6 Fo - ! € N 4

FL gs.;

A D6 0- od c ;

+g n

F F

F M

U ) \

o sN Nl - a

F € @

v o oo - -o i o- ' n $

F 3

H o Q. $ 9 €

v O Oo - -6 0 r. { d o; i ' ? 1

F O

O FN N

F C ;

o sS N

6 6D d

F O

<f,> r i

@

N O < t c ) o o oO O < t r N l i 6 oi o F N < l a ) o

ci c.i c; d dt di q;t

o

FN

6

F

4

€

€ O

O F

F O

6 4

F 6

F N

r o

€

a

N

F

F

N A S

o o N9 $ Ao o $N 6

o i r 6 9 @N O O O i

i Q o

@ @ i €

F c ; c t jN 4

d Q D

o 6 @ i l € n F i O: S o i n @ o N FF c i d i i c t c t ' i ' i . ;N 6 i

ss$$s3933

o€d

r€d;

F

E

4

rr

O A

o

i - :3 6

ON

N6r

F

€m

4N

o !

i

!da

1004 THE CANADIAN MINERALOGIST

TABLE 2e. SPIIALERITE COMPOSITIONS

Sanple t625 (2)' 1638 (7) L728 (2) ?esb (2) ?eea Ls77

FeIltrZncdCusTotal

FeI\taZncdCusite

s

7.?6 (0.0e)0.41 (0.03)

57.88 (0.34)0.24 (0.12)o.o7 (0.04)

33.15 (0.2e)

ee.51 (0.45)

0.139 (0.002)0.007 (0.001)0.885 (0.005)0.002 (0.001)0.001 (0.00r)1.035 (0.00e)

1.034 (0.009)

6.36 (0.2e)0.03 (0.05)

60.59 (0.44)0.00 (0.00)0.02 (0.04)

3:t.27 (0.17)

100.27 (0.45)

0.1x4 (0.005)0.001 (0.001)0.927 (0.007)0.000 (o.ooo)0.000 (0.001)1.042 (0.007)

1.038 (0.005)

8.36 (0.10)0.15 (0.03)

57.58 (0.37)0.47 (0.r5)0.20 (0.04)

32.87 (0.38)

ee.63 (0.79)

0.150 (0.002)0.003 (0.001)0.881 (0.006)0.004 (0.001)0.003 (0.001)1.040 (0.010)

1.025 (0.0r2)

7.26 (0.0e) 5.27 2.440.00 (0.00) 0.00 0.00

58.54 (0.34) 60.56 65.851.10 (0.08) 0.00 0.000.21 (0.04) 0.00 0.00

3.if.53 (0.34) U,32 31.75

100.64 (0.09) 100.06 100.04

0.130 (0.002) 0.094 0.0440.000 (0.000) 0.000 0.0000.896 (0.005) 0.926 1.00?0.010 (0.001) 0.000 0.0000.003 (0.001.) 0.000 0.000r.038 (0.008) r .021 1.051

1.046 (0.001) 1.067 0.990

TABLE 2f. GAHMTE COMPOSITIONS

Sample 1638 (5) ' L72s (7) 799a (3) ?eeb (6)

sio2 0.02 (0.03)AIzOg 5e.02 (0.70)FeO 5.01 (0.14)l\{nO 0.15 (0.01)Ivlgo 2.3e (0.0e)ZrO 32.67 (0.2E)

Total{ 99.26 (o.72)

0.03 (0.04)58.62 (0.38)7.27 (A30)0.23 (0.02)2.44 (0.20)

31.r3 (0.63)

99.72 (0.28)

0.001 (0.001)2.023 (0.005)2.024 (0.005)

0.178 (0.007)0.006 (0.001)0.106 (0.00e)0.673 (0.016)0.e63 (0.007)

4.000

0.25 (0.35)58.71 (0.5r)5.38 (0.13)0.03 (0.04)2.04 (0.10)

32.8E (0.13)

9e.2e (0.26)

0.007 (0.0r0)2.033 (0.012)2.040 (0.005)

0.132 (0.004)0.001 (0.001)0.089 (0.005)0.713 (0.003)0.e35 (0.009)

4.000

0.06 (0.0e)58.31 (0.61)5.10 (0.37)0.00 (o.oo)2.12 (0.33)

34.00 (0.e0)

ee.5e (0.3e)

0.002 (0.003)2.025 (0.005)2.027 (0.005)

0.126 (0.008)0.000 (0.000)0.0e3 (0.014)0.740 (0.025)0.e58 (0.008)

4.000

siIY AIsite

0.000 (0.001)2.040 (0.005)2.040 (0.005)

Fe 0.123 (0.003)I\4n 0.004 (0.000)Mc 0.105 (0.004)Zn 0.707 (0.010)site 0.939 (0.00s)

o 4.000

t Ti and Cr not detected.

Staurolite in the assemblage staurolire - anthophyl-lite - gahnite (ll3l) has a zincian composition.

Figure 3b focuses on the compositions ofcoexisting staurolite and gahnite. The staurolitecompositions include only Fe2+ calculated at thetetrahedral Fe-bearing site, using the structuralformula of Holdaway et ol. (L986, 1988). Reactions(Table 3) involving staurolite were balanced usinp(Fd+ + Fe3+; on all sites.

Figure 3b portrays cation exchange involvingonly tetrahedral sites. The change in tielineorientation correlates with Zn/F&+ in stauroliteand may reflect nonideal mixing at terrahedral sites.Alz/MS in staurolite shows little variation througha wide range of Zn/Fe (see Model mineralreactions), which supports the proposal that Fe andMg are not completely exchangeable in staurolite(Grambling 1983).

TABLE 29. GARNETCOMPOSITIONS

1005

primarily by Mn-Fe substitution. In coarse por-phyroblasts in sample 1625, the propoltion ofspessartine and almandine end-members, respec-tively, is 52.9 and 32.3 mola/o in the core, and 36.3and 44.9 mol9o in the rim. Rim compositions wereassumed to have equilibrated with syn-D2 por-phyroblasts.

Kyanite

Kyanite shows very slight deviation from idealend-member composition, involving minor sub-stitution of Fe3+ for Al to a maximum of 1.290 ofthe site.

Phy llosilicate minerals

Muscovite, margarite, chlorite and biotite arepresent in altered rocks. Margarite has nearlyend-member composition (Table 2a, 1500 mrg).Muscovite has a Na/(Na + K) of 0.08 to 0.26 andcontains a minor proportion of the celadoniteend-member. Biotite compositions are Mg-rich;Mg/(Fq + Mg) ranges from 0.55 to 0.98 and, inmost cases, the mica can be classified as phlogopite.The most magnesian composition occurs ingraphitic metasediment (1500), in which the micaapproaches end-member phlogopite. Chlorite com-positions also are magnesian and range in Mg,z(Fe,+ Mg) from 0.56 to 0.93. They correspond tohigh-Mg ripidolite and sheridanite, using theclassification of Foster (1962),

Zn was detected sporadically in minor amounts(less than or equal to 0.26t/o ZnO) in chlorite (in 3samples) and in biotite (l sample). The values arenot significant with respect to within-samplestandard deviations at the 2o level. There is noevidence of zincian chlorite, such as that describedat Franklin, New Jersey (up to 9.60V0 ZnQ,Frondel & Ito 1975).

Biotite commonly contains F at the hydroxyl site,from 0.86 to 3.77 wt.tlo F for biotite in the proximalalteration zone (Table 2b). Phlogopite in a graphiticmetasediment (1500) contains the most F. Minoramounts of F, close to the detection limit, weredetected in some cases in chlorite. F was notdetected in muscovite that coexists with F-bearingbiotite.

Fluorine and Fe-Mg portitioning

The F content of hydrous minerals generallyshows a strong positive correlation with Mg/(Mg+ Fe) (Guidotti 1984, Gunow et al. 1980, Munoz1984) as a consequence of "Fe-F" avoidance, anelectronic bonding phenomenon in transition me-tats (Valley et al. 1982) that results in the coupled

METAMORPHIC PETROLOGY OF THE LINDA DEPOSIT, MANITOBA

Sample 1625 (3)* t625rim core

SiO: 36.57TiO2 0.00AlzOs 2t .29FeO 19.67I\&O 15.901\&O 3.16CaO 2.16Naz O o.oo

(o.oe)(o.oo)(0 .12 )(0 .17 )(0.08)(0.05)(0 .10 )(o.oo)

e8.80 (0.22)

5.e41 (o.o1o)o .o5e (o .o1o)

2.672 (0.015)2.L87 (0.013)0.766 (o.ooe)0.375 (0.015)o.ooo (0.000)6 .000 (0 .017 )

Total$

sirY Ar

35.950 .17

2A.76L4.4622.84

L.7t2.420.07

98.40

5.9310.069

3.9680.0213.989

1.9953.1864.4204.4280.0226.052

24.000

vr l . t 4 .019 (o.oro)Ti o.o0o (o.ooo)s i te 4 .019 (0 .010)

FeIVInMsCaNasite

o 24.OOA

f Cr not detected.

The range of. Fe/Zn in sphalerite may reflect Stfugacity, but it is at least partly due to retrogradeexchange. Some very fine grains approach end-member ZnS. In samples containing pyrrhotite, thelow-temperature monoclinic polymorph is present,either with or without hexagonal pyrrhotite. Themost Fe-rich sphalerite (1723) contains 14.5 molqoFeS and is associated with pyrite and syn-D2porphyroblastic minerals. Although somesphalerite compositions do not represent equi-librium, in practice, reaction coefficients (Table 3)are insensitive to sphalerite composition within therange allowed by equilibration with S, in the pyritefield.

Garnet

Strong compositional zoning in garnet is defined

1006 THE CANADIAN MINERALOGIST

FIG. 3. Compositions of staurolite and gahnite. (a) Zn-Fer-Mg (molar proportions).Each point represents 1-3 analyses; ornamented tielines link the compositionalfields for each sample. The dividing line separates assemblages coexisting withpyrite (py) and either magnetite (mg) or ilmenite (il), from assemblages coexisting',rith pyrile + rutile (ru) + sphalerite (sp). Pyrrhotite occurs in some sampleson both sides of the line. Fe, is total Fe. (b) Zn-tvFez+ -Mg (molar proportions)in coexisting staurolite and gahnite, with ornamented tie-lines linkingcompositions from within the same area. Only Fe2+ (calculated) at thetetrahedral Fe-bearing site was included; Fe3+ and Fe2* at the Mn-bearing U(l)site were omitted. Samples 1638, 1723 and 799 conlain sphalerite.

substitution, (Fe,OH) = (Mg,F). Munoz (1984)derived F-intercept values [IV(F)bi] to compensatefor the compositional dependency and to allow theexpression of the degree of F enrichment as anindependent parameter. The F-intercept value forbiotite was formulated as: IV(F)61 = l.52Xun -r0.42X^n + 0.20x,id - log Xo/Xon, in which ?1ouis the mole fraction of Mg in the octahedral site,Xunn is the mole fraction of annite [l - (XM" +X;)1, and Xr,o is the mole fraction of siderophyllitel(3 - Si/Al)/1.75(l-XMJl. At fixed temperarure,IV(F)61 is proportional to log /(HrO)f(HF) in thecoexisting fluid [og f(H2O)/f(HF) = 2100/-t +

IV(F)bi; Gunow et al. l980lt thus, smaller valuesof IV(F)6; correlate with greater degrees of Fenrichment.

Figure 4a shows the relationship between F andMgl(Mg + Fe) in biotite at the Linda deposit, andFigure 4b shows the same data with F expressed asF-intercept values. The inner scale (Fig. 4b) showsvalues of log flH2O)/f(HF) in coexisting fluidscalculated at an assumed temperature of 550"C.Biotite from the proximal zone of alterationequilibrated with fluids of relatively low ratio ofH2OIHF fugacity; biotite that coexists with kyaniteinvariably equilibrated with F-rich fluids.

METAMORPHIC PETROLOGY OF THE LINDA DEPOSIT' MANITOBA

TABLE 3. MODEL MINERAT RSACTIONS

1007

a- THE Si-Al-Fe-Mg-Zn-K-H-F-O-S SYSTEM

Sample Reactiou mL

1. 22-90L

2. 22A-16382

3. 22L-1752

4. 228-17232

5. 3L799a2

6. vc-r377

7. UC-r377

8. UC-L377

9. vc-L377

10. yc-1377

11. UC-L377

8.47 k y * 0.L4ch * 2.t4py * 0.27 a p + 0.55 fr 20 + L.27 O 2r+ st * t.46qtz * 2.2852

st*5.4Isp*2. l7oz;+ 8.83sb *7.llqtz * 0.96pv + t-71"H2O + t.7452

qtz -t 2.36ky + 0.74bi * A33py + 452H2O + 0.L7O2+ 0.44ch * 0.84rns + 0.85"4.F + 0.33^52

s l *5 .50sp+2.L2O2= 8.859h I7.6tqtz t lJ5py * L.\4HzO * 1.5852

st*5.42sp*2.4302+ 8.989h * 7.48qtz * 0.85py * L.78EzO * 1.8652

qtz * 3.84ky + L.22bi * 0.32py * 6.76EzO + 0.L6O2=r 0.68ch * 1.33rns + 1.06I/f' * 0.3252

1.66ch * 2.98rns * 1.15sp *0.L7pv +2.34HF +0.2502= s, * 3.90qtz + 2.72bi + I4.56H2O + 0.32^92

8.62ky * 0.13ch * L.40py * 1.16sp + 0.55 H 20 + t.34O2+ 5,, + L.67qtz + L9752

9.35ky + 0.24bi * L.46py * 1.15sp * L.87 E zO * L.36Ozr= 5t + L.47qtz * 0.26ms + 0.23H F + 2.0452

436ch { 1.15sp * 8.23rns + 6.50.4.F * 0.3502+ st * 15.10*s +7.53bi * 0.58pv *7.8lqtz + 4LJ'5H2O

5.87ch* 1.15sp * lL.1lrns + 8.83.4"F' * 0.6952:=1 st + 23.46ky + 10.17bt * l.27py * 10.009t2 + 55.87 H2O

r.79

0.80

2.00

0.75

0.76

2.00

1.29

1.48

t.49

0

oo

b. THE Si-Al-Fe-Mg-Mu-Zn-K-H-F-O-S SYSTEM

Sample Reaction

12' 22A-1625 0'08srt * 3'22ms * 1'38ch * 0'69pv * 1'19sp + 1'99lrF * 1'00o2 1'28= sf * 2.63bi + 4.68qtz + L2.6982O * l.28Sz

L rn,=logfs"/ logf6". ? Reaction solved by linear regression.

mL

Coexisting phyllosilicate minerals at the Lindadeposit show the normal sequence of F/OHpartitioning: muscoviteMagnesium number, Xyu : Mg/(Mg + Fe),normally increases in pelitic rocks in the sequence:biotite < chlorite (Thompson 1976).In aluminouspelites, muscovite has been reported with Xlu,,unearly equal to that of biotite (Novak & Holdaway1981) or with lower Xrn than biotite (Mengel &Rivers 1988). In the F-rich altered rocks of theLinda deposit, the Xlan sequence is muscovite <chlorite < biotite , which thus matches the sequenceof F-OH partitioning. The Fe-Mg partitioningcoefficient between biotite and chlorite, Kp@i-ch),shows a oositive correlation with the F content of

biotite (Fig. 5). As a result of Fe-F avoidance andthe preference of F for biotite, the normallyobserved sequence of Xra* is reversed, and Fe-Mgdistribution is strongly composition-dependent.This critical observation has ramifications bearingon reaction topology.

MetauoRpHtc GRADE oF THE LlNpe Dspostr

The Linda deposit occupies a small volume withrespect to the scale of regional metamorphism. Thealtered rocks in the sample suite represent an evensmaller volume. and it is reasonable to assume thatall the samples underwent the same tec-tonometamorphic history.

1008 THE CANADIAN MINERALOGIST

o.:P r . oI

s

o

0 . 5L

0

(b)2 . 5

0 . 5

l og f s r6 , / f 6p ,

o .7Mg / (Mg + FeJ

5 5 o o C

IL

0 . 5 0 . 9o . 7

M g / [ M g + F e )

Ftc. 4. Biotite compositions showing the relationships between Mgl(Mg + Fe) andF. In both figures, the solid line separates samples of the proximal zone ofalteration from calc-silicate rocks, and the dashed line separates proximal-zonesamples containing kyanite from those without kyanite. (a) F (per 4 OH sires)as a function of Mgl(Mg + Fe). Note the positive correlarion, parricularly forsamples from the proximal zone of alteration. (b) F-intercept value, lV(F)61,determined by the method of Munoz (1984), as a function of Mg,z(Mg + Fe).Values of log f(H2O)/flHF) in coexisting fluids at an assumed remperature of550oC are shown on the right-hand scale of the y axis.

Equilibrium ossemblages

Mineral assemblages considered to representequilibrium parageneses are summarized in Tablel; low-variance assemblages were used in balancingreactions (Table 3). Accessory minerals allowed the

inference of supplementary information, for ex-ample the distribution of rutile and ilmenite asindicators of relative 52 and O, fugacities.

Microstructural relationships between por-phyroblasts and tectonic fabrics show that mostporphyroblastic minerals crystallized during D2.

p rox ima l

a l t e ra t i on

z o n e

' t 500

Akyan i te -bear ing $

r\u 1752

ca l c - s i l i ca ter o c k s

ca l c - s i l i ca terock s

Ko'0.453Fr0.723

R:0.907

oa 05 00 1.0 1.2

F in biotita (per 4 OH-sites)

Frc. 5. Fe-Mg partitioning coefficient between biotite andchlorite, Kp(bi-ch) = (Mg,/Fe)6/(Mg/Fe).1, as afunction of F in biotite. There is a strong positivecorrelation, and the regression line passes through Kp= 1.0 at F = 0,61. Samples with KD < I .0 were notused in the construction of the grid in Figure 8.

1009

This includes the assemblages biotite - kyanite andstaurolite - anthophyllite, which represent the peakmetamorphic grade achieved at the Linda deposit.Garnet and kyanite (without biotite) crystallizedbefore D, and probably during D1. Garnet rimswere assumed to have re-equilibrated during D2.

Chlorite and muscovite define Sr schistosity;biotite and, less commonly, muscovite and chloriteshow textures indicating crystallization or recrys-tallization during Dr. In general, where two texturalgenerations of the same phyllosilicate mineral occurtogether, there was no systematic difference in theircompositions. Apparently, phyllosilicate mineralsshowing textures typical of early crystallizationre-equilibrated at higher grade.

Disseminated pyrite is abundant in all samples.Sphalerite, where present, varies from abundant totrace amounts. In general, even where present insmall amounts, it has a wide distribution as fine

METAMORPHIC PETROLOGY OF THE LINDA DEPOSIT' MANITOBA

I

x

L

E4I

oU'(l)

3 0 0 4 0 0 5 0 0 6 0 0

T e m p e r a t u r e ( o C )

Frc. 6. Schematic pressure - remperature - deformation path followed by the Linda deposit. The low-grade

kyanite-forming ieaction accounis for syn-D1 kyanite without biotite. Syn-D2 minerals crystallized in the.intervalbounded by thi stability limits of stauiolite'+ quartz and margarite + quartz. Reaction 2 was determined for

fixed bulk ire,zMg of intirmediate value (Hoschek 196?) and represents a naffow reaction-band. Sources of reaction

dara: I Kerrick (1968), 2 Hoschek (1967),3 Storre & Nitsche (19?4); stability fields of aluminosilicates: Holdaway(r971).

- D 1 *

1010 THE CANADIAN MINERALOGIST

inclusions in syn-D2 porphyroblasts and dissemi-nated grains in the phyllosilicate matrix.

Constraints on prcssure and temperature

Sphalerite geobarometry at the Linda depositgives a pressure of 5.0 kbar [using the calibrationfunction (standard error *0.30 kbar) of Hutchison& Scott l98ll, using the most Fe-rich sphalerite(Table 2e, 1123, 14.5 mol9o FeS) associated wirhsyn-D2 porphyroblasts. This sphalerite coexists onlywith pyrite, and not with hexagonal pyrrhotite;therefore, 5.0 kbar is a maximum limiting pressure.Sphalerite inclusions in pre-D2 garnet por-phyroblasts have 13.3 to 13.8 mol9o FeS (2analyses), giving a pressure of 5.7 to 6.2 kbar,within 2o error limits of 5 kbar. The matrix of thissample (1625) contains pyrite and monoclinicpyrrhotite. It is possible that the sphaleriteequilibrated with pyrite and hexagonal pyrrhotiteand was subsequently armored by garnet.

Syn-D, kyanite formed at low grade (Fig. 6). Thecoexistence of Fe-Mg staurolite (0.020/o ZnO) andquartz at the Linda deposit gives an approximatelower limit for peak metamorphic grade, and thecoexistence of margarite (Table 2a, 1500mrg) andquartz gives an upper limit. Syn-D2 peak metamor-phism is bracketed by these two assemblages(stippled reactions in Fig. 6) at a temperature ofabout 550'C. The intersection of the two limitingreactions gives a lower pressure limit of about 4.5kbar. The minimum pressure is also constrained bythe kyanite - sillimanite phase boundary.

The role of bulk-rock and fluid compositions

The variety of syn-D2 parageneses at the Lindadeposit is a function of variation in bulk-rockcomposition and metamorphic fluid composition,and not of a gradient in metamorphic grade. Thetemperature of about 550"C is reasonable for thechlorite - biotite - staurolite zone of regionalmetamorphism (Fig. l). Garnet persisted in associa-tion with chlorite and muscovite by virtue of itshigh spessartine content (Table 2g).

In zincian bulk-rock compositions, the stabilityof staurolite is enhanced relative to binary Fe-Mgsilicates. Staurolite is unique among commonsilicate minerals in that it has a tetrahedrallycoordinated Fe-bearing site into which Zn partitions in preference to the octahedrally coordinatedFe-bearing sites of other Fe-Mg silicates (Holdawayet al. 1986, 1988). The incorporation of Zn intostaurolite and gahnite allowed the assemblagestaurolite - anthophyllite - gahnite to crystallize atthe Linda deposit. Gahnite and staurolite in altered

rocks must be treated as ternary Fe-Zn-Mgminerals.

The presence of F in biotite and amphiboleallows the stable persistence of these minerals ingranulite-facies terranes (Petersen er a/. 1982,Yalley et al. 1982).In the case of the Linda deposit,the incorporation of F into biotite allowed thecrystallization ofbiotite + kyanite at agrade lowerthan that which would normally be expected inpelitic rocks. The influence of F on phase equilibriamust be considered in prograde biotite-formingreactions.

SllrcnrE - Oxrns - Sur-proe Eeurr-rgnre

Model mineral reactions

Mineral assemblages considered to have equi-librated during D, peak metamorphism (Table 1)were modeled in the ten-component system Si-Al-Fe-Mg-Zn-K-H-F-O-S (Table 3a) using themineral compositions determined for chlorite,biotite, muscovite, staurolite, gahnite andsphalerite (Table 2). Table 3b involves Mn as anadditional component, and garnet as an additionalphase. The presence of Na in biotite and muscovitewas neglected.

At fixed temperature and pressure, samplescontaining seven mineral phases in the ten-com-ponent system will have a phase-rule variance of 3.If/(HrO) andlHF) are fixed, the variance will bel, and the assemblage defines a reaction. Thereaction can be balanced using the mineralcompositions determined; the slope in log lSr) -log /(O) space is determined by the stoichiometryof the reaction. The slope is tangential to thereaction curve and applies to the particular mineralcompositions from which it was determined.

Sample 34C-1377 contains eight mineral phases;thus, at fixed temperature, pressure, /(HrO) and

"f(HF), the assemblage has a phase-rule variance of0. The assemblage can be used to balance thebundle of univariant reactions radiating from theinvariant point (Table 3a, Reactions 6-9). Reac-tions l0 and 11 conserve 52 and 02, respectively.

The reaction coefficients were determined exact-ly (in an algebraic sense) by formulating the mineralcompositions as an array of linear mass-balanceequations (Spear el al. 1982). The significance ofthe reaction coefficients was tested by the least-squares method on subsets of the minerals, byfitting the model to the composition of one mineralchosen as a dependent variable (Greenwood 1968,Spear el al. 1982). The procedure was iterated,taking each analyzed mineral in turn as thedependent variable. Residual values in excess of thepropagated 20 standard deviations, i,e., t/fD(coef-

ficient x 2o)21, associated with the analyses wereinterpreted to indicate that the coefficients of thereaction are significant. Errors on the independentand dependent variables were taken into accountby summing the squared weighted 2o standarddeviations for each element from each mineralanalysis; we employed the regression coefficient ofthe mineral as a weighting factor.

Figure 7 is a schematic illustration of theprinciple, using Reaction 2 (Table 3a) as anexample. As solved exactly, gahnite decompositionwould occur in a three-phase triangle to sphalerite,staurolite and a small amount of kyanite. As solvedby least squares in an algebraically overdeterminedsystem (omitting kyanite), if the residual vector (R)does not exceed the 2o error envelope, thendecomposition in the three-phase triangle is indis-tinguishable from degenerate decomposition on thesohalerite - staurolite tie line. In the case of residual

l 0 l I

vectors in excess of 2o errors, the coefficients ofall phases would be significant.

The evaluation of the significance of reactioncoefficients is critical in the case of reactionsinvolving the decomposition of gahnite (228-1723,34-799, 22A-1638). The balanced reactions involvekyanite or an Fe-Mg phyllosilicate mineral' withrelatively small coefficients. These coefficients arenot significant within the 2o errors, and thedegenerate reactions were balanced by linearregression. The phase-rule variance of thesereactions is not increased; within error limits,Alz/Mg in gahnite and staurolite has a constantratio and should be treated as one component.Error evaluation shows that the coefficient ofchlorite in Reactions I and 8, and the coefficientsof biotite and muscovite in Reaction 9, aresignificant in balancing the Mg component, andthe coeflicient of garnet in Reaction 12 is

METAMORPHIC PETROLOGY OF THE LINDA DEPOSIT, MANITOBA

M gZ n

gahn i te = s tauro l i te + spha ler i te + kyan i te

Ftc. ?. Schematic illustration of the evaluation of the significance of reactioncoefficients with respect to analytical uncertainty at the 20 level. If the residualvector associated with least-squares solution of the mass balance, sphalerite +staurolite - gahnite = 0, lies within the 2o error envelope, the coefficient forkyanite is not significant (a). If the residual vector exceeds the error envelope,the coefficient for kyanite is significant (b)'

Alz

Res idua lvec tor w i th iner ro renve lope

Res idua Ivec to routs ide er ro renve lope

t0t2 THE CANADIAN MINERALOGIST

uncertainty of about *0.05 (or about 390) to theslopes of these reactions.

Log f(O) - log f(S) petrogenetic grid

The model reactions (Table 3) involve oxidationand sulfidation, and each yields a reaction slope,m, in logfiS) - log/(OJ coordinates. ASchreinemakers analysis of the reactions and slopeswas used in the derivation of a petrogenetic grid inlog /(S) - log l(Or) coordinates, at fixed pressureand temperature (Fig. 8). The grid is projected

significant in balancing the Mn component. Thisalso implies that the slopes of Reactions 8 and 9are different.

Reactions I and 8 involve the same minerals, butwith different compositions and, thus, the coeffi-cients of the balanced reactions are different. Thecoefficients of 02 and 52 in these reactions weretested by varying the mineral compositions withintheir 2o ranges of uncertainty. This procedureshowed that the difference in the slopes, log/(S)/lo5jf(p) (see next section), exceeds theanalytical error. Analytical error contributes an

\

A,\,

po

l og fg ,

Ftc. 8. Log f(O) - log/(S) petrogenetic grid derived by Schreinemakers analysis of the reactions in Table 3. Theinsets are projections onto the Zn-Al2-Mg plane. Abbreviations: bi biotite, ch chlorite, gh gahnite, il ilmenite, kykyanite, ru rutile, sp sphalerite, st staurolite.

m9' l !

4957

, " fui t

N

olttI

System : Si-Al- Fe-Mg-Zn-K-H-F -O-S

Projected through : muscovitequartzpyriteH2O & HF

Fixed Pressure & Temperature

s.,

"99',

pv

METAMORPHIC PETROLOGY OF THE LINDA DEPOSIT, MANITOBA l 0 l 3

through q\artz, pyrite, muscovite of the analyzedcomposition, H2O and HF. It is based only onsamples showing a narrow range of logf(HzO)/f(HF), from 4,0 to 4.3 (Fig. 4), in fluidscoexisting with biotite. In biotite-absent samples,log f(H2O)/f(HF) in the fluid is unknown.

The mineral compositions can be projected ontothe Zn-Al2-Mg plane. The choice of elementalcomponents allows the projection of phase rela-tions involving silicate, oxide and sulfide minerals.Each nondegenerate univariant reaction involvesfour phases in the subsystem Zn-Al2-Mg. Aspreviously discussed, gahnite compositions lie onthe staurolite - sphalerite tie line, and thedecomposition of gahnite is a degenerate reaction.

On Figure 8, the activity of FeO defines a familyof contours with slopes of 2, determined by thestoichiometry of the reaction: (FeO)r1;"u," + 52 =FeS2 + 0.502. The compositions of ternaryFe-Mg-Zn minerals (and sphalerite) vary along theunivariant reactions; these reactions cross isoplethsof the activity of FeO and Fe end-member in eachmineral. The relative position of the sample points(Fig. 8) was constrained by the requirement thatthe analyzed silicate and oxide compositionsbecome progressively more Fe-rich at lowerfugacities of 52, The compositional constraintsdefine a unique solution of the phase relations forthe samples involved.

The sulfidation of chlorite to biotite and kyaniteoccurs in a Zn-absent subsystem and provides alink to petrogenetic grids that are limited to binaryFe-Mg silicates (Nesbitt 1982, 1986a, b). Thereaction slope is fixed at 2, as required by thestoichiometry of sulfidation, and the compositionsof biotite and chlorite are constant. The reactionrepresents the intersection line of two continuousreactions of slope 2, involving the sulfidation ofdaphnite in chlorite [i.e., (da)"n + 52 : qtz + py+ ky + 0.5Ozl and the sulfidation of annite inbiotite [i.e., (ann)6; + ky + 52 = qtz + py + ms+ 0.5021. These continuous reactions defineisopleths of Fe end-member activity, whichdecreases with increasing Sz fugacity in the pyritefield. (Some isopleths of X1,1u are shown in Fig. 8,for example X1,au in phlogopite of sample 1500.)The discontinuous reaction limits the most Mg-richchlorite and the most Fe-rich biotite compositionscoexisting with kyanite. (Note the limiting X1,auvalues represented by sample 1377 in Fig. 8.)

The significance of fluorine

The consequences of the reversed Fe-Mg par-titioning between biotite and chlorite, due to thepresence of F, can be seen in reaction topology(Fig. 8); biotite, as the more magnesian mineral, is

stable at higher 52 fugacity than chlorite. In low-For F-absent systems, the assemblage biotite +kyanite is sulfidized to a more magnesian chlorite(e.g., Fig. l1 in Nesbirt 1986b).

In the petrogenetic grid (Fig. 8), /(HzO) and

/(HF) are assumed to be constant. The range ofcalculated \oe f(H2O)/f(HF) in fluids (Fig. 4) isconsistent with minor variation, but it does notprove constant fugacities. Samples 1752 and 1377contain chlorite - biotite - kyanite (Reactions 3 and6, Table 3), with slightly different Xvu inphyllosilicate minerals in each sample (Fie. 8). Thedivariance of the reaction may be due to slightlydifferent los f(HzO)/f(HF) values, from about 4.0in sample 1752 to 4.2 it 1377.

Sample 1747.5 contains kyanite + biotite' withbiotite compositions much more Fe-rich than thelimiting values in sample 1377 (Fie.4). The meancomposition in sample 1747.5 is 0.76 F (per 4 OHsites), MglFe = 2.0 and corresponding log

^H2O)/f(HF) = 4.2. Figure 5 predicts- -".tol-?l"Mglfe partitioning between chlorite and biotite forless than 0.61 F in biotite (per 4 OH sites), andMglFe below about 2.4. Log f(S), Mg/Fe, F andKo@i-ch) are all positively correlated; this suggestsrhat a reversal in Ko@i-ch) occurs in thepetrogenetic grid (Fie. 8) at relatively low values of

/(S) regardless of constant.f(HzO) and/(HF). Sucha reversal implies the possible reappearance ofbiotite + kyanite at lower log /(SJ (and moreFe-rich mineral compositions) and may account forthe assemblage in sample 1747.5. The chlorite-bear-ing assemblages in samples 957 and 996.5 areinconsistent with biotite + kyanite in 1747.5. Thisproblem is almost certainly related to variable

f(HtO) and /(HF); however, the available data donot allow conclusive resolution of the grid at low

"f(sr.

PRESSURE - TEUPENATURE PETROCENETIC GRID

PON F-RICH PROTOLITH

The reversal in Fe-Mg partitioning betweenbiotite and chlorite has implications for thetopology of P-T reaction grids. In Figure 9, thetopology expected in pelitic rocks has been revisedassuming a sequence of Xr" in which biotite is moremagnesian than chlorite.- This topology is ap-propriate for a F-rich protolith, such as the alteredrocks of the Linda deposit. The reaction of chlorite+ staurolite to biotite + aluminosilicate, which isthe biotite - aluminosilicate isograd reaction atSnow Lake (Fig. 1) and in pelitic rocks in general,does not occur. In the AFM plane, chlorite lies onthe Fe-rich side of the biotite - aluminosilicate tieline: a discontinuous aluminosilicate-producing

It

cn

k

1014 THE CANADIAN MINERALOGIST

System : SiO2-Al2Oa- FeO- MgO-K2O-H2O-HFProjected through: muscovite

quartz

H2O & HFXyn sequence: s t< ch < b i < c rd

p l u s Z n , S , O

Temperature 5500c

Ftc. 9. Pressure - temperature grid for F-rich protolith in which biotite is more magnesian than chlorite. The insetsare projections on to the AFM plane. The dashed line represents Reaction l0 (Table 3); the point labeled "Linda"corresponds to the gahnite-absent invariant point in Figure 8 (labeled "1377").

h

+

A\ \

\67

oJoooo.

reaction would have to involve some moremagnesian mineral, for example cordierite.

At a higher grade in a F-rich protolith, the P-Tgrid predicts chlorite decomposition within athree-phase triangle to biotite, staurolite andaluminosilicate. This four-phase assemblage is thesame as that which defines the biotite - aluminosili-cate isograd. This is a common assemblage at theLinda deposit, but with zincian staurolite present.The reaction that accounts for this assemblage inaltered rocks involves the additional componentsZn, S and O, and the additional mineral phasessphalerite and pyrite. Reactions l0 and ll (Table3a) are balanced using mineral compositions insample 1377 and Reaction l0 is shown as a dashedline in Figure 9; the point (labeled "Linda')corresponds to the invariant point defined by thisassemblage in log lO2) - log l(SJ space in Figure8 (labeled "1377"). In P-T space, the reaction isdivariant as a function of 02 and 52 fugacities. Ifthe activities of HrO or HF also are not constant,the reaction is multivariant.

IMPLICATIONS FoR THE SNow Lare REGION

A temperature of about 550"C is reasonable forthe chlorite - biotite - staurolite zone of regionalmetamorphism in pelitic rocks. The parageneses atthe Linda deposit are consistent with this grade.Figure l0 compares the reactions in F-rich protolithwith reactions in pelitic rocks as a function ofgrade. The left side of the figure is an isobaricsection of the P-T grid in Figure 9. The biotite -aluminosilicate assemblage formed through dif-ferent reactions in rocks of different bulk-composi-tion, and these reactions occurred at differentgrades. In the F-rich protolith, the reactionoccurred at lower temperature, within the stabilityfield of kyanite. In pelitic rocks, the reactionoccurred at higher grade, within the stability fieldof sillimanite. The reaction in pelitic rocks (Fig. 10)is the biotite - sillimanite isograd reaction proposedby Froese & Moore (1980) at Snow Lake (Fig. 1).The distribution of aluminosilicate polymorphsobserved at Snow Lake supports the difference in

METAMORPHIC PETROLOGY OF THE LINDA DEPOSIT' MANITOBA

b i + s i + s t

c h + s t

ky

[ i + f tyr! ; tr py

c h + s P + 0 2

ch + crd

r0 l5

ch

s i

A1203r k v

o:o

Eo

F

F-r ich proto l i th(qtz, ms, H2O, HF)

grade illustrated in Figure 10; biotite - sillimaniteoccurs in pelites, and biotite - kyanite is morecommon in altered rocks. The dashed reactioninvolves the additional components Zn, S and O,and is not definitive of grade if 52 and 02 fugacities(or HF) vary.

Concentrations of halogens have not beendetermined in micas from the Anderson Lake andStall Lake mines (Fig. l). The micas are likely tocontain significant amounts of F; in this case, thebiotite - aluminosilicate assemblage is not diagnos-tic of metamorphic grade. If the isograd were based

Frc. 10. Comparison of mineral assemblages in AFM projection between F-richprotolith and pelitic protolith, as a function of grade. The left side is an isobaricsection through Figure 9 at 5 kbar, taking into account the kyanite - sillimanitetransition at a grade slightly higher than that of the Linda deposit. Biotite -

aluminosilicate forms through different reactions in the two bulk-rockcompositions. The reactions occur at different grades, in the kyanite field inF-riCh rocks, and in the sillimanite field in pelitic rocks. The biotite-sillimanite-forming reactioS in pelitic protoliths corresponds to the isograd reaction at SnowLake (Reaction 2, Fig. 1).

Pelit ic protolith(qtz, ms, HzO)

solely on parageneses observed in peliticmetasedimentary rocks, it would lie further northand coincide closely with the leading edge of theMcleod Road thrust fault. A possible relationshipbetween the fault and the isograd is beinginvestigated by further work in the Snow Lakeregion.

CouclustoNs

Two major conclusions of general importanceemerge from this studY:

l 0 l 6 THE CANADIAN MINERALOCIST

1' With appropriate restrictions, and in particularwith a projection through pyrite, metamorphicassemblages in mineralized altered rocks can beprojected into the subsystem Zn-Al2-Mg, Differenttopologies are separated by reactions that form agrid on a log/(O) - log/(Sj diagram. In contrastto earlier work, gahnite and staurolite can betreated as ternary Fe-Mg-Zn solutions, which iscritical to the interpretation of their phase relations.2. Despite the superficial similarity in theirparageneses, P-T grids based on pelitic bulk-rockcompositions are not appropriate for altered rocks.Direct comparison of assemblages in rocks ofdifferent bulk-composition will result in an incor-rect interpretation of isograds. The incorporationof F into biotite expands the stability of biotite -aluminosilicate to lower grades. In the case ofF-rich phlogopitic biotite, the biotite is moremagnesian than coexisting chlorite and, conse-quently, the topology of reactions involving theseminerals is modified. Petrogenetic grids ap-propriate for altered rocks can be developed; asstated by Eskola (1915), mineral parageneses "standin definite relations to the chemical composition".

ACKNOWLEDGEMENTS

This study was part of the doctoral research ofE. Zaleski under the supervision of Norman M.Halden, University of Manitoba. FalconbridgeLimited and Minnova Incorporated provided accessto drill core and to unpublished company maps andreports. Minnova provided funding for mineralanalyses, and Falconbridge allowed the use of theiroffice facilities at Snow Lake. Robert Sparkassisted in the field. Thanks are due to T. ScottErcit, Mineral Sciences Division, CanadianMuseum of Nature, for instruction in the acquisi-tion of high-quality mineral analyses. Discussionswith Doreen Ames, Alan Bailes, Dugald Car-michael, Jim Franklin, Alan Galley and RogerSkirrow eventually germinated into ideas. Themanuscript was improved by reviews from RobertF. Martin, Robert Tracy and anonymous reviewers.Funding from the Natural Sciences and EngineeringResearch Council was provided through theirscholarship program and an operating grant toNorman Halden.

REFERENCES

Berres, A.H. (1987): GS-12 Chisel - Morgan Lakesproject. Manitoba Energt and Mines, MineralsDivision, Repoil of Field Activities 1987,70-79.

(1988): GS-8 Chisel - Morgan Lakes project.Manitoba Energy and Mines, Minerals Division,Report of Field Activities 19t8, 53-61.

Bnvrozra, L.T. & Scorr, S.D. (1987): Application ofchlorite - sulfide - oxide equilibria to metamor-phosed massive sulfide ores, Snow Lake,Manitoba. Econ. Geol. E2.963-970.

Drrn, W,A., HowrE, R.A. & ZussvaN, J. (1966): AnIntroduction to the Rock-Forming Minerols.Longman Group Ltd., London.

EsroLe, P. (1915): On the relations between thechemical and mineralogical composition in themetamorphic rocks of the Orijiirvi region. Bull.Comm. G6ol. Finlande I, 108-145 (English sum-mary).

Fosrrn, M.D. (1962): Interpretation of the composi-tion and a classification of the chlorites. U.S. Geol.Surv., Prof. Pop. 414-4.

Fnoesr, E. (1971): The graphical representation ofsulfide - silicate phase equilibria. Econ. Geol. 66,335-341.

(1977): Oxidation and sulphidation reactions..|lr Application of Thermodynamics to Petrologyand Ore Deposits (H.J. Greenwood, ed.). Mineral.Assoc. Can., Short Course Handbook 2,84-98.

& Gaseennrur, E. (1975): Metamorphic zonesin the Snow Lake area, Manitoba. Con, Mineral,t3. 162-167.

& Moonr, J.M. (1980): Metamorphism in theSnow Lake area, Manitoba. Geol, Surv, Con., Pap,78-27.

FnoNoeL, C. & Iro, J. (1975): Zinc-rich chlorites fromFranklin, New Jersey. Neues Jahrb. Mineral. Abh.r8(2), ln-115.

GaLLEv, A., Aues, D.E. & FnaNrLrN, J.M. (1988):Geological setting of gold mineralization, SnowLake, Manitoba. Geol. Surv. Can., Open-File Rep.1?00.

GneunuNc, J.A. (1983): Reversals in Fe-Mg partition-ing between chloritoid and stawolite. Am. Mineral,6E, 373-388.

GnerNwooo, H.J. (1968): Matrix methods and thephase rule in petrology. XXIII International Geol.Congress 6,267-279.

Gurporrr, C.V. (1984): Micas in metamorphic rocks./n Micas (S.W. Bailey, ed.). Rer. Minerol. 13,357-467.

GuNow, A.J., LuorucroN, S. & MuNoz, J.L. (1980):Fluorine in micas from the Henderson molybdenitedeposit, Colorado. Econ. Geol. 75, 1127-1137.

HoLoewey, M.J. (1971): Stability of andalusite and

the aluminum silicate phase diagram. Am. J. Sct.27t .97- t3r .

-, Durnow, B.L. & HrNroN, R.W. (1988):Devonian and Carboniferous metamorphism inwest-central Maine: the muscovite - almandinegeobarometer and the staurolite problem revisited.Am. Mineral. 73,20-47.

& Suonr, P. (1986): A model for thecrystal chemistry of staurolite. Am. Mineral. 71,tt42-tt59.

HoscHEr, G. (1967): Untersuchungen zumStabilittitsbereich von Chloritoid und Staurolith.Contrib. Mineral. Petrol. 14. 123-162.

HurcursoN, M.N. & Scorr, S.D. (1981): Sphaleritegeobarometry in the Cu-Fe-Zn-S system. Econ.Geol. 76, 143-153.

Krnnrcr, D.M. (1968): Experiments on the upperstability limit of pyrophyllite at 1.8 kilobars and3.9 kilobars water pressure. Am. J. Sci, 266,20+214.

Knrrz, R. (1983): Symbols for rock-forming minerals.Am. Mineral. 68, 277 -279.

Mrxcel, F. & Rrvrns, T. (1988): Reversals in Fe-Mgdistribution between muscovite and coexistingminerals in metapelites of the Ramah Group,northern Labrador: bulk compositional influences.Geol. Assoc. Can. - Mineral. Assoc. Csn.,Program Abstr. 13, 484.

Moun, D.W. & NrwroN, R.C. (1983): Kyanite -staurolite metamorphism in sulfidic schists of theAnakeesta Formation, Great Smoky Mountains,North Carolina. Am. J. Sci. 283. 97-134.

MuNoz, J.L. (1984): F-OH and CI-OH exchange inmicas with applications to hydrothermal oredeposits. Ilr Micas (S.W. Bailey, ed.). Rev.Mineral. 13,469-493.

Nesnrrr, B.E. (1982): Metamorphic sulfide - silicateequilibria in the massive sulfide deposits atDucktown, Tennessee. Econ, Geol. 77, 364-378.

(1986a): Oxide - sulfide - silicate equilibriaassociated with metamorphosed ore deposits. I.Theoretical considerations. Econ. Geol. El. 831-840.

(1986b): Oxide - sulfide - silicate equilibria

l 0 l 7

Pelitic and felsic volcanic terrains. Econ. Geol.8l,841-8s6.

Novnr, J.M. & HolDawev, M.J. (1981): Metamorphicpetrology, mineral equilibria, and polymetamor-phism in the Augusta quadrangle, south-centrall/.une. Am. Mineral. 66,5l-69.

PrrunseN, E.U., EssrNp, E.J., PEecon, D.R. &VellEv, J.W. (1982): Fluorine end-member micasand amphiboles. Am. Mineral. 61,538-5M.

Spmn, F.S., RuuBLE, D., I I I & FEnnv' J.M. (1982):Linear algebraic manipulation of N-dimensionalcomposition space. In Characterization ofMetamorphism through Mineral Equilibria (J.M.Ferry, ed.). Rev. Mineral. f0, 53-104.

Sronnr. B. & Ntrscnr, K.-H. (1974): Zur Stabilidtvon Margarit im System CaO-AI2O3-SiO2-H2O.Contrib. Mineral. Petrol. 43, l-24.

Tuovrpsox, A.B. (1976): Mineral reactions in peliticrocks. II. Calculation of some P-T-X(Fe-Mg)phase relations. Am. J. Sci. 276,425-454.

Vellpv, J.W., PrrrnseN, E.U., EssrNr, E.J. &BowMaN, J.R. (1982): Fluorphlogopite andfluortremolite in Adirondack marbles and calcu-lated C-O-H-F fluid compositions. ,4rn. Minerol.67, 545-557.

WnLnorD, P.C. & FnenrLtn, J.M. (1982): TheAnderson Lake mine. Snow Lake, Manitoba. /nPrecambrian Sulphide Deposits (R.W. Hutchinson,C.D. Spence & J.M. Franklin, eds.). Geol. Assoc.Can., Spec. Pap. 25, 481-523.

ZaLesrr, E. (1989): Metamorphism, Structure ondPetrogenesis of the Lindo Volcanogenic MassiveSulphide Deposit, Snow Lake, Monitobo, Canodo.Ph.D. thesis, Univ. Manitoba, Winnipeg,Manitoba.

& Helorr, N.M. (1988): Reconstruction ofsynvolcanic alteration associated with the Lindamassive sulphide deposit, Snow Lake, Manitoba.Geol. Sun. Can., PaP. 88-lC' 73-81.

& - (1991): Dynamometamorphism ofthe Linda volcanogenic massive sulphide depositand its implications to the Snow Lake region,Manitoba, Canada. Geol. Assoc. Can. - Mineral,Assoc. Can., Program Abstr. 15, A144.

Received May 29, 1990, revised manuscript acceptedMarch 5, 1991.

METAMORPHIC PETROLOGY OF THE LINDA DEPOSIT, MANITOBA

associated with metamorphosed ore deposits. II.