Tourmaline as a recorder of pegmatite evolution: Bob ... · during the formation of the pegmatite...

American Mineralogisl, Volume 71, pages 472-500, 1986

Tourmaline as a recorder of pegmatite evolution: Bob Ingersoll pegmatite,Black Hills, South Dakota

Bnlor.nv L. Jor,r,rrr, Jlrrrcs J. Plprru, CHlnr.Bs K. SrrnlnnnInstitute for the Study of Mineral Deposits, South Dakota School of Mines and Technology, Rapid City, South Dakota 57701-3995

Ansrucr

Tourmaline occurs as an accessory mineral in five of six zones in the Li-, B-, Be-, Nb-,Ta-, and Sn-enriched, internally zoned Bob Ingersoll No. I pegrnatite located in the Pre-cambrian core of the southern Black Hills near Keystone, South Dakota. The purpose ofthis investigation is (l) to examine the usefulness of tourmaline as a petrogenetic indicatorin this selting and (2) to apply this indicator to interpret the internal evolution of thepegmatite. Tourmaline occurs in ten distinct types based on texture or habit and coexistingmineral assemblage. Chemical analyses of the tourmaline show compositional diferencesbetween types that may be useful in determining the crystallization sequence and differ-entiation mechanisms. Trends of major-element variations in tourmaline from the countryrock to the core include the following: (l) Mg and Ti decrease abruptly from the countryrock through the border zone to the wall zone; (2) Fe decreases and (Li + Al) increasefrom the wall zone to the core; and the minor elements Mn, Zn, and Ca generally increasetoward the core. Tourmaline compositional trends in this pegmatite, coupled with texturalfeatures and mineral associations, provide evidence of an inward, generally sequentialcrystallization of the border zone, wall zone, intermediate zones, and finally the core, butwith overlap between the inner wall zone and intermediate zones. The pegmatite melt wasprobably saturated at the onset of crystallization; however, fluid exsolution may havetemporarily stopped during the initial crystallization of the third intermediate zone and aperiod of tourmaline instability, but then resumed, along with tourmaline crystallization,during the formation of the pegmatite core.

INrnonucrroN

Tourmaline is a common accessory mineral in peg-matites that surround the Harney Peak Granite at the coreof the Precambrian terrane of the southern Black Hills.This area is a classic example of a pegmatite field sur-rounding a granitic intrusion that may belong to the samemagmatic system (Redden et al., 1982). The pegmatitesconsist predominantly of small, layered, and weakly zoneddikes, but include a number of large, internally zoned,rare-element pegmatite bodies. Pegmatitic rocks in gen-eral have been the subject of a tremendous quantity ofresearch and geologic literature (e.g., Jahns, 1955), andthe Black Hills pegmatites have also been extensivelystudied (Page and others, 1953; Sheridan et al., 1957;Norton etal.,19621' Redden, 1963). A detailed model ofthe crystallization of pegmatite was presented by Jahnsand Burnham (1969) and was recenfly summarized byJahns (1982). Although great progress has been made inpegmatite research, many questions remain regarding theinternal evolution of zoned pegmatite bodies (e.g., eern!,197 5, 1982; Jahns, 1982).

The study of zoned pegmatites by classical petrologicmethods such as whole-rock chemical analysis is ham-pered by the coarse grain size of minerals, by the heter-ogeneity of large sections of the pegmatite, and by poorexposure in three dimensions. In addition, much of the

0003{04x/86/03044472502.00 472

volume of well-exposed zoned pegmatites has been minedor eroded; therefore, bulk compositions may only be es-timated. One approach that may overcome such difficul-ties is the use of minerals with variable compositions asrecorders of petrogenetic information (e.g., Foord, 1976;Shearer et al., 1985).

Tourmaline occurs in many of the pegmatites surround-ing the Harney Peak Granite, in the granite itself, and inthe country rock ofcontact zones surrounding both graniteand pegmatite. The crystal structure of tourmaline is suchthat it allows substitution of a variety of ions with respectto both size and charge. Additionally, tourmaline is thedominant boron-bearing mineral and is the dominant maf-ic silicate in many of the zoned pegmatites of the BlackHills. These properties make tourmaline potentially wellsuited for the detailed study of chemical variation withina pegmatite during its internal evolution. The purpose ofthis study is to characterize the compositional and texturalvariations of tourmaline in a large, internally zoned peg-matite and to evaluate the use of tourmaline in this settingas a recorder of crystallization history. Compositional andtextural trends of tourmaline are then applied toward in-terpretation ofthe internal evolution ofthe Bob Ingersollpegmatite.

The idea of using the chemical variability of tourmalineto decipher petrogenetic details is not unique to this study.

JOLLIFF ET AL.: TOURMALINE AS A RECORDER OF PEGMATITE EVOLUTION

. H o l

S p r i n q s

0 8 t 6 2 4 3 2 k m

0 5 lO 15 20 mil€s

Fig. l. I-ocation of the Bob Ingersoll pegmatite in the Precambrian core of the Black Hills. Harney Peak Granite shown byrandom V pattern.

473

,lN

I

$

\o\\

It has long been known that the color and other opticalproperties of tourmaline are roughly indicative of itschemistry (Jenks, 1935; Carobbi and Pieruccini, 1947;Bradley and Bradley, 1953; El-Hinnawi and Hofmann,I 966; Leckebusch, I 978; Rose et al., I 98 I ) and that theseproperties may be used to indicate the fractionation trendsand relative degree ofevolution ofthe rocks that host thetourmaline. Staatz et al. (1955) established systematiccompositional variation trends in tourmaline, from thewall inward, of the bizonal Brown Derby pegmatite dikesof the Quartz Creek district of Gunnison County, Colo-rado. They found the major variation from the outer wallzone to the core to be an increase of Li concomitant with

a decrease of Fe. Foord (1976, 1977) characterized thecompositional variation of tourmaline in layered, mia-rolitic pegmatite dikes and host rocks of the Himalayapegmatite-aplite dike system in San Diego County, Cal-ifornia. These and other studies (Power, 1968; Neiva,1974; Manning, 19821' Henry and Guidotti, 1985) haveshown tourmaline to exhibit systematic compositionalvariations and to be an effective recorder ofpetrogeneticinformation.

The pegmatite selected for study is the Bob IngersollNo. I dike, located 2mi(3.2 km) northwest of Keystone,in Pennington County, South Dakota, near the northernborder of NWI/4 Sec. 6, T2S, R6E (Fig. l). The Bob

SOUTH

DAKOTA

474

Ingersoll No. I dike is a large, zoned B-, Be-, Nb-, Ta-,Sn-, and Li-enriched pegmatite that contains tourmalineas an accessory mineral in five of the six zones. Thispegmatite was selected because ofthe abundance and vari-ability of the tourmaline, the good surface and under-

LEVEL G

5too

L E V E L E5080

50@LEVEL C

5OTIO LEVEL B

5020

5000LEVEL A

4980

4960

4940

Quo.lz - cleovelo ndite -l ep i do l i t e pegmo t i l e .Core zone.

Quor lz-c leovelondi tepegm o t i teThird intermedioie zone.

Quo i t z - c l eove l ond i l e -omblygoni te pegmof i le .Second infermediole zone

Potossium fe ldspor- quortz-c leovelondi te pegmol i te.Fi rst intermedioie zone

Cleovelondi le - quor lz -muscovi te pegmot i le.Wol l zone.

Quor lz - c leovelondi le -muscovi te pegmol i teEorder zone.

Quor lz-mico or quor lz-bioli te - gornei schist.Country rock.

Pegmot i le contocl , doshedwhere opproximolelyI oco ted.

JOLLIFF ET AL,: TOURMALINE AS A RECORDER OF PEGMATITE EVOLUTION

; -;n,'-

EXPLANATION

I ; w . ; - lL) o-i-( |1 . . , , " : l

1 . . . 0 ' c . . 1l + + + + I

f:a4TT:ll i O-c-A . ' ll , - : . . : " . : 1

F.1I-rIIf . . K-o-c :11."'. '"';. : ': I

E-X- i',5l r / _ ; : \ i r \ - - ' 1 1

l.'-.c-o-M -iil , r) ' : ' ; , ' ; t>-,1

Fa+lJ,;1| )) |

N

SCA LE

20 rO O rO 20 40 6oft

Fig. 2. Horizontal geologic sections of Bob Ingersoll No. I dike showing distribution of pegmatite zones. Geology based onmapping as part of this work and mapping by Hanley (1953) in areas now mined out, backfilled, or flooded.

ground exposures, and because it was mapped and de-scribed during an earlier stage ofits mining history (Hanley,1953; in Page and others). As part ofthe present work,the pegmatite was remapped to include new exposuresand to establish precise locations of samples. Tourmalines

475

SAMPLINGLEVELSi G

i F

- E

- D

Quortz - c leovelo ndite -lep ido l i te pegmot i te .Core zone.

Quor tz -c leove lond i lepegmot i te .Third intermediote zone

Quorlz - c leove lond i ie -omblygon i ie pegmoli ie.Second inlermediole zone.

Potossium feldspor -quorlz -c leove lond i te pegmot i te .Firsl inlermediole zone.

Cleovelondite -quorlz -muscovife pegmoli le.Woll zone.

Quortz - cleovelond i te -muscovite pegmotiteBorder zone. (shownon ly where unusuo l lyth ic k )

Count ry rock

Pegmotite contocl, doshedwhere opproximotely locoled,dotled where projecled.

Bockf i | |

- C

r B

5000

4980

4960

4940

49?O

4900

from 45 sample locations have been analyzed using mi-croprobe, neutron activation, and atomic absorption tech-niques.

Tourmaline occurs in ten distinct associations at theBob Ingersoll No. I dike. Descriptions of the textural typesand an overview ofthe geology ofthe pegmatite dike aregiven in the next section.

Several of the important questions addressed by thisstudy are as follows: (1) Does tourmaline show consistent,systematic compositional variations related to spatial dis-tribution in the pegmatite, e.9., vertical position or zonallocation? (2) Do textural variations of tourmaline corre-spond to compositional variations? (3) Do the composi-


EAST

ELEVATION( fee t )

E X P L A N A T I O N

ffiFi-=l

f .o-e .lr - ' - ' - l

tHl

ffilKrlF,.t;qY.]

E:il

NFSIlt; .i,:::-:,1

Fig.3. CrosssectionA-A'ofBoblngersoltNo. ldike.Geologybelow5040-ft(1536-m)levelbasedonHanley(1953).Distributionof zones in area of above-surface projection based upon exposures north and south of section line.

tional variations indicate the sequence of crystallizationof the pegmatite zones? Is the sequence consistent withmineral assemblages and textures? (4) Do compositionaland textural variations of tourmaline reflect changes inthe crystallization process and the nature of the crystal-lizing medium?

Gnor-ocv oF BoB INcnnsor,r, No. 1 nrrn

The No. I dike was mined extensively b etureen 1922 arl.d 1945 ,principally for lepidolite, beryl, and amblygonite. Since the timeofprevious mapping, several subsurface levels have been flooded,and additional mining has exposed new surfaces. As part ofthisstudy, the No. I dike was remapped by plane-table surveying of

476 JOLLIFF ET AL.: TOURMALINE AS A RECORDER OF PEGMATITE EVOLUTION

"h

c - Q - M K - Q - C Q - C - A Q - C Q - C - LWol l F i rs t Second Th i rd CoreZone t- lnfermediote Zones

lilil-l F"""'-llour,-,n I lo,o,,r. | | | 1o,o,,,. I E:::'-" I l3l:iJ' I A.....o.yFA'JI:,"1 136:11^'., 1 ll3i?i+. I l3i3Ji'. I lBEh,#il li'i.'fJl,. l MinerotsITOURMALINEI ITOURMALINEI ITOURMALINEI IPERTHITE I ITOURMALINEI ITOURMALINEI

Q - C - MBorderZone

successive levels of the pegmatite. The structure of the dike isshown as a series of horizontal sections in Figure 2.

The No. I dike is an inclined and flattened funnel-shapedpegmatite, approximately 25 x 30 m in plan near the top, andis at least 75 m deep (Fig. 3). The long axis ofthe dike trendsnorth-northwest, and the deep root of the dike plunges to thesoutheast.

The pegmatite of the No. I dike conformably intrudes meta-morphic rocks consisting predominantly of biotite-garnet-stau-rolite schist and minor quartzite. The pegmatite-country rockcontact is sharp and easily recognized, but locally irregular inshape. The contact is generally paralleled by schistosity in thecountry rock. The conformable nature of the schistosity aroundthe dike in general, and where the contact bulges or rolls, suggestsa relatively ductile country-rock response as it was shoulderedaside by the intruding pegmatite.

The dike is shown in vertical cross section in Figure 3. Muchofthe upper portion ofthe dike has been rnined; however, thestructure of the central portion as it existed prior to mining canbe broadly inferred from exposures within and around the pe-rimeter of the glory hole.

Tnnes

There are six pegmatite zones in the No. I dike, distinguishableby their texture or mineralogy. These include a border zone, wallzone, three intermediate zones, and the core. Approximate av-erage modes for each of the zones are given in Figure 4. Modesare based on visual estimates and grid counts (10-cm spacing) ofpresent exposures with consideration given to abundances ex-pressed by Hanley (1953).

Fig. 4. Approximate average modal composition of pegmatite zones. Symbols: Q, quartz; C, cleavelandite; M, muscovite; P,perthite; A, amblygonite-montebrasite; L, lepidolite. Abbreviations: Cassit, cassiterite; Amblyg, amblygonite-montebrasite; Col-Tan,columbite-tantalite; Triph, triphylite-lithiophilite.

M o j o r

M o d o I

Minero logy

The border zone is generally very fine grained (see grain sizescheme, Table l) and consists of varying proportions of quartz,albite (or cleavelandite), and muscovite. Microcline, present insmall quantities, is surrounded by reaction rims of muscovite.Several accessory minerals are present, including tourmaline,apatite, and amblygonite. Tourmaline is commonly the coarsestmineral in the border zone, attaining lengths up to 2.5 cm. Tour-maline occurs as slender, tapered crystals that are oriented sub-perpendicular to the contact with country rock, with the taperedend nearest to the contact. Some of the crystals are inclined ornearly parallel to the contact, a feature that may represent flowalignment or may indicate the presence of inclined gradientscaused by the flowage of melt interior to the border zone as itcrystallized. The border zone is nearly continuous around thepegmatite and ranges in thickness from I to l0 cm. The transitionfrom border to wall zone is sharp, commonly occurring over adistance on the order of centimeters, and is marked by an abruptrncrease rn glarn slze.

The wall zone characteristically contains the assemblagecleavelandite-quartz-muscovite. Cleavelandite is generally mostabundant and is typically intergrown with quartz or in virtuallymonomineralic aggregates as much as I x I m in cross section.Muscovite is generally medium to coarse grained and occurs asisolated books as rnuch as 30 cm in long dimension. Slenderprismatic tourmaline is intergrown with the coarse muscovite,and very coarse tounnaline occurs with the large cleavelanditeaggregates. Fine- to medium-grained rocks also occur in the wallzone, but are more common in the footwall exposures. A notabletrend in the major modal mineralogy ofthe wall zone is a decreasein the abundance ofcleavelandite and an increase in the abun-

JOLLIFF ET AL.: TOURMALINE AS A RECORDER OF PEGMATITE EVOLUTION 477

dance of quartz from the outer part of the wall zone toward theinterior. The wall zone, which forms a continuous unit aroundthe pegmatite, ranges in thickness from I to 3 m. It is commonlythickest on the footwall portion of the zone. Depending uponposition in the dike, the wall zone contacts each of the inter-mediate zones, which are discontinuous (Figs. 2, 3). These con-tacts, which are gradational, were mapped close to the appearanceofthe mineral that characterizes the zone (e.9., potassium feld-spar, first intermediate; amblygonite-montebrasite, second in-termediate, etc.).

The first intermediate zone is distinguished by the presence ofcoarse-grained to very coarse gtained potassium feldspar crystalsthat are set in a matrix consisting largely of massive quartz andsmaller amounts of cleavelandite. Muscovite is present in smallquantity particularly where this zone contacts the wall zone. Theabundance ofpotassium feldspar and its grain size decrease to-ward contacts with other zones. In the pinnacle ofthe glory hole,now mined out, which was spatially close to the middle of thezone, perthite was reported to compose 900/o of the rock (Hanley,1953). Single feldspar crystals currently exposed in the mrne areas much as 3 m across in the maximum exposed dimension.

The second intermediate zone is distinguished by the abun-dance of amblygonite-montebrasite (Ambr5Mbsrr). This zone isdiscontinuous and occurs in what are inferred to be irregularlens-shaped units between the first intermediate zone and thecore, third intermediate zone, or wall zone. Amblygonite-mon-tebrasite forms as circular masses as much as 2 m in diameter.This zone appears to be confined to the upper 25 to 30 m ofthedike, although arnblygonite-montebrasite occurs in trace quan-tities in other zones lower in the dike.

The third intermediate zone is composed predominantly ofmedium-grained to very coarse gtained quartz intergrown withcleavelandite that forms as discrete grains as well as in a veinlikemosaic pattern between quartz grains. The general texture ofthiszone is similar to the texture of the quartz and cleavelanditematrix of the first and second intermediate zones. This zone ismore continuous than the first or second intermediate zones andalmost entirely surrounds the core. It is discontinuous along thefootwall in lower parts of the dike where the core abuts the wallzone. In the central and lower parts ofthe dike, where the firstand second intermediate zones are absent. the third intermediatezone directly contacts the wall zone. An accessory mineral as-semblage of tourmaline, beryl, apatite, and cassiterite typicallyoccurs at contacts between the third intermediate zone and thecore.

The core of the pegmatite contains principally quartz, cleave-landite, and lepidolite (3.5 to 6.0 wto/o Li,O in lepidolite). Quartzand cleavelandite are fine to medium grained, but cleavelanditealso occurs in large pod-shaped aggregates. I-epidolite is very finegrained, and in several exposures large aggregates are composedalmost entirely of lepidolite. The lepidolite content appears toincrease from outer contacts ofthe core inward.

Tourmaline occurrences and textures

The various occurences of tourmaline at the Bob Ingersoll No.I dike can be grouped according to their texture or habit, color,and mineral association. (Groups are hereafter referred to astypes.) Each type tends to occur characteristically in a given zone.Ten types have been identified (includingtourmaline in the coun-try rock); their distinguishing characteristics and zonal distri-bution in the pegmatite are summarized in Table l.

Country rock adjacent to contacts with pegmatite is extensivelytourmalinized (Tuzinski, 1983; Shearer et al., 1986). In samplesof country rock from outside the pegmatite, dark-brown tour-

Table l. Tourmaline distribution at the Bob Ingersoll No.dike

0ccu r rence C o l o r * Textu re si ze** Type

Country fock &country focki n c l u s i o n

Pegnatite Zones

Border zone

I a l l z o n e ( I )

( 3 )

( 4 )

F i r s t i n t e r -med ia te zone

Th i rd in te f -medi ate zone

c o r e ( 1 )

P o i k i l o b l a s t i c , v f - fs u b - t o e u h e d r a l ,s h o r t p r i s n a t i c

T a p e r e d , s l e n d e r v f - fpr i snat i c

E u h e d r a l , t a p e r e d ' m - v cp r i s m a t i c t o n a s s l v es p l a y s

S u b h e d r a l , i n f - nal I ot r i onorphi c-g f a n u l a r m a t r i x

E u h e d r a l , i n t e r g r o w n f - nw i t h c o a r s e n u s c o v i t e

A n h e d r a l , c o n p r i s e s Is m a l l v e i n l e t s

S u b h e d r a l , i n t e r g r o r n f - mw i t h q u a r t z , c , l e a v e -I a n d i t e

A n h e d r a l t o e u h e d r a l , ms h o r t p r i s n a t i c , i n t e r -g r o w n w i t h q u a r t z -c l e a v e l a n d i t e - b e r y l -

a p a t i t e

A n h e d r a l t o s u b h e d r a l , f - ni n t e r g r o w n w i t h f i n e -g r a i n e d q u a r t z - c l e a v e -I a n d i t e - l e p i d o l i t e

S u b - t o e u h e d r a l , f - mcomonly splayed, andi n t e r g r o w n w i t h q u a r t z

Dark brownt o b l a c k

Dark green

B l a c k t od a r k b l u e

l l e d i u m B l u e

B l u e - g r e e n

Green

D a f k b l u eto green

Green tol i g h t g r e e n

L i g h t b l u eto grey

P i n k t ored

1aIb

( 2 )

( 2 )

Notes :* Macroscop ic co lo r* * Gra in s ize schene:

v f - very f ine - <6 mf - f i n e - 6 m t o 2 . 5 c mm - med iun - 2 .5 cn to 10 cmc - coarse - 10 cm to 30 cn

- verv coarse - >30

maline (type la) partially replaces biotite, whereas ferromagre-sian components of country-rock inclusions within the pegmatite

are completely tourmalinized (type 1b). Fine-grained, dark-greentourmaline (type 2) occurs in the border zone as euhedral, slender,tapered prisms. These crystals are commonly fractured perpen-dicular to the c axis, and the fractures are filled by the fine-grainedmatrix characteristic of the zone.

Black tourmaline occurs in the wall zone near the border zonecontact. This tourmaline (type 3) is generally coarse to very coarsegrained rangrng up to 1.5 m in length, but may also be fine ormedium grained, and may occur within inches ofthe border zone.Away from the border zone, toward the inner contact of the wallzone, the color ofthe tourmaline is dark blue. The very coarsetourmaline is oriented with the c axis perpendicular to, and ta-pered toward, the outer contact ofthe pegmatite. Small branchesof tourmaline splay from the main body of individual crystals,clearly indicating the direction of growth (Fig. 5a). These crystalsare generally euhedral and are not embayed by other minerals,although they can contain fracture fillings.

Fine-grained, medium-blue, subhedral tourmaline forms a sec-ond texturally distinct type in the wall zone (type 4). Such crystalsoccur in interior parts ofthe wall zone as intergrowths with fine-grained aggregates of cleavelandite, qluartz, and muscovite.

A very distinctive bimineralic assemblage of tourmaline epi-taxially intergrown with coarse muscovite also occurs in the wallzone (type 5). The tourmaline is generally euhedral, forming fine-

478 JOLLIFF ET AL,: TOURMALINE AS A RECORDER OF PEGMATITE EVOLUTION

Fig. 5. (A) Type 3 tourmaline from the wall zone. White mineral encasing tourmaline is cleavelandite (sample B7). (B) Type 8tourmaline from third intermediate zone adjacent to contact with core. Note pen for scale (sample Bll). (C) Photomicrograph oftexture resembling exsolution. Dark bands are Fe-rich; light bands are Fe-depleted. Width of photo is 4 mm (reflected light, crossedpolars). (D) Elongate fluid inclusions parallel to c axis, which impart chatoyancy to the light bands shown in C. Length ofcenteredinclusion is approximately 50 pm.

to medium-grained, slender prisms. Some of these are finely col-or-zoned, with dark-blue cores, light-blue and green intermediatezones, and colorless rims.

Green tourmaline occurs as thin (<l cm) fracture fiIlings inthe wall zone (type 6). This type has been found only in cross-cutting relationships with very coarse, type 3 tourmaline. Suchtourmaline veins extend no more than several centimeters intothe surrounding rock from the tourmaline that they transect.

Tourmaline is not an abundant accessory mineral in the firstintermediate zone. Its occurrence is limited to bluish-green, sub-hedral to anhedral crystals that are fine grained or form fine- tomedium-grained aggregates (type 7). Tourmaline aggregates areintergrown with the quartz-cleavelandite matrix of this zone.

In the third intermediate zone, tourmaline is characteristicallymedium to light green (type 8). The typical habit is one of me-dium-grained, stubby prismatic crystals (Fig. 5b). In several cases,euhedral tourmaline forms skeletal crystals, cores of which con-sist of the assemblage quartz, apatite, and beryl. This textureprobably formed by early nucleation and growth of euhedraltourmaline, followed by cocrystallization with quartz, beryl, andapatite, which commonly filled in the open end of the taperedtourmaline crystals (Norton et al., 1962). Green tourmaline of

this zone is also found as anhedral grains around the rims ofcoarse, pillow-shaped aggregates of radially grown cleavelanditelaths.

Tourmaline in the pegmatite core consists of two texturallydistinct varieties. Very light blue, fine-grained, poikilitic subhe-dral tourmaline is intergrown with fine-grained aggregates ofquartz, cleavelandite, and lepidolite. Pink, euhedral tourmaline(elbaite) also occurs in the core, but appears to be associated withsmall quartz segregations, although it also occurs in quartz-cleavelandite-lepidolite aggregates.

ANl,l,vrrcl.r, PRocEDURES

Tourmaline was collected from 45 sample locations, chosento represent the entire range oftextural types present at the peg-matite (see Table 2). For those types showing a broad range ofcompositions after initial analyses, additional samples were takento characterize the full compositional range (types 2, 3, and 5).Sampling is biased toward the wall zone owing to the abundanceof tourmaline in the wall zone versus interior zones and to thevolume of the zones currently available for sampling. Most ofthe volume of zones containing types 7, 8, 9, and l0 tourmalinehas been mined from the pegmatite.

JOLLIFF ET AL,: TOURMALINE AS A RECORDER OF PEGMATITE EVOLUTION 479

Tourmaline grains were analyzed for ten major and minorelements by microprobe techniques using the fully automatedvec electron microprobe at the South Dakota School of Minesand Technology. The wavelength-dispersive system with TAP,PET, and LIF crystals, was utilized for quantitative determina-tions of Si, Ti, Al, Fe, Mg, Mn, Zn, Ca, Na, and K. All analyseswere conducted at an accelerating voltage of 15 kV, a beamcurent of approximately 0.0150 pA, and a counting time of20 000 counts or 20 s. Corrections were made with the empiricalcorrection technique of Bence and Albee (1968). A minimum offive analyses were taken per grain; however, as many as 20 anal-yses were taken to characteiue zoned grains.

Tourmaline separates were prepared for 27 samples, using acombination ofhand picking, magnetic separation, and heavy-liquid separation. Ferrous iron determinations were conductedusing standard KMnOn titration, following HF-H,SO. dissolutionin pressurized Teflon-lined Paar bombs. The separates were alsoused for lithium and trace-element determinations by atomicabsorption spectrometry. Fluorine contents were determined byion selective electrode (Bodkin, 1977). Data for l5 major, minor,and trace elements were obtained by INee using a high efrciency130-cm3 Ge(Li) detector (2590 rwHlr 108 keVT of 50Co), a 4096channel arralyzet and coincidence-noncoincidence Ge(Li)-NaI(Il)counting systems at the Battelle Pacific Northwest Laboratories.The details of the INAA procedure and the systematics of coin-cidence-noncoincidence counting are described by Laul (1979).

Thirteen samples that span the range of compositions wereanalyzed by X-ray powder diffraction to determine unit-cell pa-rameters. The study was conducted with a Norelco difractometermounted on a Philips constant potential generator, equipped witha graphite crystal monochromator, d-compensator, and samplespinner. Experimental conditions included CuKo radiation, 40kV, 20 mA, Yzo per minute scan speed, and an Alr0, internalstandard.

Rnsur,rsResults of chemical analyses of the suite of 27 samples

are presented in Table 3. Cations per formula unit werecalculated based on normalization to six silicons (see alsoFortier and Donnay, 1975; Sahama et al., 1979; Walentaand Dunn, 1979). The reason for this normalization isthat Si is consistently determined to be slightly in excessof six atoms per formula unit when normalized againstoxygen for charge-balance. Such excesses, although usu-ally small (less than 0.05 atoms per formula unit), aredifficult to reconcile in terms of the structure of tourmaline(next section).

The danger of conducting analyses on powders takenfrom whole or multiple crystals that are compositionallyheterogeneous or zoned is well known (e.g., Donnay, 1968;Foord, I 976). The values presented in Table 3 are averageanalyses; however, several of the samples are composi-tionally zoned and are therefore difficult to characterizein terms of Li content, which must be determined fromchemical analyses of powders. Li contents determined fornonzoned tourmaline, however, correlate well (inversely)with Fe contents (correlation coefficient = -0.95). The Fecontents determined by microprobe anlysis were used toapproximate the Li contents of zoned samples.

Total Fe determined by microprobe analysis is reportedas FeO. Many ferrous iron determinations indicate that

Table 2. Tourmaline sample population

T o u r m a l i n e s a m p l et y p e l o c a t i o n s

G r a i n s a n a l y z e d l , l i c r o p r o b eb y m i c r o p r o b e a n a l y s e s

1 alb23

567I9

1 0

a high percentage ofthe total Fe is ferrous; however, dueto variable oxidation during dissolution, the results arenot reproducible. The good correlation ofFe and Li valuesalso indicates that a high percentage of the total Fe isferrous. The green color of some of the tourmaline (types2,5,6,7, 8) may indicate the presence of Fe3*, in com-bination $/ith Fe2*, as a chromophore (Taylor and Terrell,1967; Wilkins et al., 1969; Moore, 1982). Green colormay, however, also result from Fe2*-Ti4* charge-transferprocesses (Faye et al.,1974).

Results of the unit-cell determinations are listed in Ta-ble 4 and values of c and c are plotted in Figure 6. Thecell parameters correspond closely with the compositionslisted in Table 3. All samples plot close to the schorl-elbaite reference line (Donnay and Barton ,1972). Sampleswith detectable MgO are displaced slightly toward highervalues of c.

Compositions of zoned crystals

Tourmaline crystals of types 1,2,3, and 5 are typicallyzoned (Figs. 7, 8). Crystals of tourmaline in the countryrock (types la and b) contain cores with low values of Fe/(Fe + Mg) (e.g., 0.54 to 0.67) relative to rims (0.68 to0.80). Type la tourmaline, near the pegmatite contact,replaces biotite that has Fe/(Fe + Mg) values of approx-imately 0.73. Farther out from the pegrnatite (20 m), theFe/(Fe + Mg) value of biotite decreases to 0.61 . The Fe/(Fe + Mg) values of tourmaline in the country rock ap-proach, but do not overlap appreciably, the Fe/(Fe + Mg)values of tourmaline in the border zone (Fig. 9).

Individual crystals of the border zone (type 2) containhigh Mg and Fe in their cores relative to their rims (Fig.7a), and Fe(Fe + Mg) values lower in the cores than inthe rims. Crystals with two repetitions of this type ofzoning also occur (Fig. 8). Zn,Mn, and Ca are in generalinversely correlative with Ti, Mg, and Fe. In the wall zone,type 3 tourmaline is zoned with high Fe at the taperedend, nearest the outer contact of the wall zone, and lowFe on the growth surfaces toward the interior of the zone(Figs. 7b, 7c). Crystals oftourmaline intergrown with mus-covite in the wall zone (type 5) are strongly zoned con-centrically about the c axis, but not as much along theirlength. This zonal arrangement and the shape of the crys-tals reflect the favorability of growth in the c direction.

56

71 2

45

2

233

10

t z2359

t553262492T261622

L+o6

o

o

- o- o !

o -3 8 5! E O O

- o 6 L

E " g bO O O - -

, 9 ' o E E oo + o > o

o 6 + o co - E o -

J O E D E

3 E 5 0 P b+ o o - o +L s 6 . O Oo o o + !

@ o c v E +L - O O

+ L O f C

3 3 8 9 9 , ,. . * l l .o * * EO * t * E .+ * * * * c coz

o 6 0 < 66 N F O N F N F O € F

O N O N O O O O - O O 6 - @o v v v

N

h n a 6 6 o @ 6 o N @ o 6 8 8 8 3 8 3 5 ; 8 H b N 3 3 I Sn O 3 @ O n F - l O { { F O O O O O F O O O O F O F O F F

F o a h o o o o N o - - g @ n 6 o o o o o o o N o o o o o

o o o h n - n F 6 N . ? B f 36 n 6 - Q O O O N n r a O O O O - O 6 O n F - r

O E N @ O @ @ O N O - n a O O O F | @ O O N F o O F O r F

O e h o O - O N O - F n O n A - O o O O O N O O O O on n @

o o o - 6 N F F o N 6 < 6 9 6o F o - $ 6 o o n o € 6 o o o 6 @ o < o F ! - F ^ a ^F ! i i l n o n 6 o n c i : 6 ; 5 6 6 i t I 6 6 6 : i r c O o l - 5 F X

O n O O O - O N O F F € 6 n O O O O O O O N O O O O On n @

a n s F n o o s n F F n q 8 8 8 5 8 F 3 9 ! = E 5 l n * ' +

{6. : :?{{ t : .qQqu. 5.5.5. E5. ,1 id.11E 5.R.8. I qq \ o < o o o o o N o o o 6 0 6 a o o F o o o o N o o o o on n F @

@ n o @ r n 6 n n { o - n o o o 6 6 6 F @ n o F 6 - 6 S +

- .? . . . :?11?\?, : i? 8.8.8. 6.8.$.83.8.44 8.qq" : rm O { - O O O O F O O O € 6 n @ O O - O O O O N O O O O On n - 6

o , { ' o - b . o o - * - - - 8 8 8 3 8 8 R F B R t S R B B Ht o h F o N F o o o < o o o o o F o @ o o o N F o @ o 9 +n o n F o o o o F o o o t @ n @ o o e o o o o N o o o o o

6 @ N OE G O O h N F h O @ - 6 N <

+ - O O N 6 - 6 O n < O {

@ N F - r n a - n F o 6 n 8 8 8 8 9 f i 5 S S S t R = R A RN - O N @ m n F n - 6 6 - o O O O o o N 6 O o @ o @ O @ F

O O h O O O O O N O O F n 6 n @ F O O o o o o N O o O O on n o

e n s 6 6 - r n N N F 6 . 8 8 8 $ 3 n 5 9 n R t E P B t E@ o < n o F { o n o a N o o o o n o N n o o n 9 0 F o F o

+ o - o - o o o N o o F $ +

o n o o o - o o o o N o o o o o

tt

- 6 0 @ 0 n o 6 { o - - n ! 8 8 8 S P S S g S f i B 8 S 8 $ 5 Eo 6 o @ F F F o n o o - N a o o o N - 6 o o o s q o F o F n o

! . . .I i O i O N O O O N O O - n g O n O O O - O O O O N O O O O On h o o o

s +L -o o P ld t ! !

- g 6 0 0

F € 6 f i ; > x oN r c * O O O ; * * v i . O

o o N o o o o o r c N F + * * O- - - O q E c o d s O - * - - - O O c c - O O Lo F < L - = N O Z Y J L O o 6 < < F [ = = N J O Z y L F

JOLLTT ET AL.: TOURMALINE AS A RECORDER OF PEGMATITE EVOLUTION

6 r o R 3 ! - @ 9 SO h n F O O O O - O o6 V

o 6@ n N + 6 FF h F O F F O F O O

m N o F o o o o F o o o o NO V V - F

o oo o . S 8 3 ! - 5 9

F + - F o o o o e o o o F {@ F

h. o s N s o F

E @ 3 6 N O O - N - O n

E - O - O O O O - O O N - <v F

h o oN F @ N F

6 O - N O O - h O N n

N - - F O O O O O F O 6 { Fo v

n oF n n F O O @ N h

{ * @ F o N n o n o @ F F

N N - - O O O O O O O - 6 -

@ 6 9 4 N @ O< N - O t N g 6 F e n m

O m O m Q O O N O - O O @ O

N

F O h - 6 6F F F O O N h + N N @ @

q n - N O O O N O F O O h -

. € o 6o F 6 0 n r N N n 6 0 {

c o @ o N o - 6 R N - Q - 6

- @ h N O h n N N + F

o n + N F N F N n N O h O oN N F O N

n o F F F F F @ - n F n 8 8 9 S $ g F 6 g R t $ g t f ir o < - o 6 h o { o F n - o o o @ | @ o o - F F o F o € r

. C . . .O n @ O O F O N O F F 6 6 n 6 O o o o o o N o o o o on n @

n o 6 F F F N 6 6 n N o 8 8 8 f i = n E € 8 s t B € i g< O O N O n F O C E O + { O O O - O @ Q O N O o O O I I F

A O A @ O O F O N - - 9 O n O - O O O O O O N O O O On n v @

o o o a 6 n o c 6 @ - t g B gn @ F - n h Q m n n n n O O O @ F O @ 6 c h - c€ E n O O < n O * O - { n O O O @ t 6 O O - F F O F O F F

. . . tO 6 @ O O F O N O - - 6 @ h O O O O O O O N O O O O O6 h e

o o o F F N N m o o o o { ! a@ h O - O F O S N @ F n O O O 6 6 O < 6 - a - @ 6 9 -n E n F O n N O < O O 6 O O O O 6 | O O O - F F O F O F F

o 6 @ o o - o N o F F 6 @ h 6 O o O O o o N O O o O O6 n @

o o o n o 6 F h N N l F o e e oN @ h < n @ t + { N m m l O O O @ - O O N O n q O c O pn O @ F O - @ O n O Q O n O O O F O N O O - + 9 O F O F

@ O { O O O O O N O O - 6 @ 6 @ o o - o o o o N o o o O on 6 @

€ o + N o n n { N F o F 3 8 8 ; 3 3 $ 8 3 ; S E B R S@ E F @ O 6 @ O O O F O O O O O @ | 6 O O - + @ O O O 9

. c . . . . | . . . . . .6 < 6 0 0 0 0 N o o F n 6 6 @ O - O O O O N O O O o O6 n €

6 oz r o

c E o ! o o o ! o - - - ( o -6 & O 6 0 6 J z t s t t s t s - e

480

6

o

O L

+- o

- +

^ lo l-Jc l

- l

=lo

J

zIou

o

q

r

O

o

I

a

o

)<eo

o

v

o

o

q

I

I

@

so

r;

F

o

o

oo

a@

r6

oF@

No

6

N

o

o

N

o-

o

F

o

o

b8lL O l

8 5 1N l

i3 lL l

i.tO - lo - l

o

O a

5 F

481

oo

o o o

t -o 9

6 " Io +

o oc @o o +

o oo o oL .+ a o Fo o L s

c o oo + +L ^ Co L v o

o o

L + - O

c o ' co 6 + O O

s 6 ' " 3L O O O

o + o o oo x o f !- o L -d d o cE O O > OO L L- e : E 3

r o E - oo i o L + oo * r o - o+ * + + J Lz


^ 9 - 9 5 8 8 . ! R o .o N o N o o o o + o o N o oo v@

N NN F F F O N

F n o o 6 F - @ o

O O n N O O O O F O O n - FN O 6 6 F

oN @ n @ q m

@ F @ N - F O F O h -

6 N O n F O O O - O O m $ Oo -

6 h o< O n N N - O F

N + n - O O O O O O O F NN F V V O

@ < N nF - F * N M F N F N N

6 O N m - O O O O t O F N nF N

N F

< N @ N @O N F - 6 N O N N F N

6 O n N O O O - O O O - <o N vI

o S o o 6 { 6o r @ N o - o 6 - o n N n

o G F F o o o o F o o N n qo 6

6 - 6 NF N N O O F F F N F

n n o - o o o $ p o e+ -

oo e N n @ h @

o @ n o - < 6 - N N

s o € o o o o o l o o 6 N @{ { F N

6 6 0 6o n o N o F - - r @

o O m O O n . . . I. o - o o N o s

O N O - O O O N

F o < o o - - 6 o o 8 8 8 ; P S $ 5 8 3 $ A 36 ! N - O < 6 @ O E 6 6 O O O O F I N I O O F e - m I e Q

. C . . C . . | .

n S * o o o - 3 - F @ n @ - o o o - N o o o o

3 , o S - 6 Fo < N 6 O ? F n N F 6 €

b t : : d " i < i dd< i . i dd n - ' d^ o o o 6 @ 0 o 6 ! o o Q n i n '

6 O F l 6 € h F 6 O O O O N 6 t € F E F h I Oh - N a 6 6 O + E N n @ O O O @ | @ | O O F n O F I r F

. E . . C . . 1 .F c 6 O O O O N F e + € n e O O O O O N O O O O6 6 v O

O O O ? m - m - < q l o - 6 Fm N N F @ @ o o n n @ o 5 6 6 ' 6 o o < N 6 q o N E q d6 E m o o o - m F o @ o € o o o n I o o F o e e - 6 o e +

F O O O - O O - O F O n O n @ - O o O O - N O O O O On 6 v o

o o o - N i N c o 6 @ F N o o o< O o N - a O F n i F O t O O O < - N O N n 6 = { t N - 6F F o o o 9 o N Q F @ N n o o o N o F o N F o € o o o r @

F O @ O O - F O N O - F a O n @ - O O O O O - N O O O O Oh 6 v o

o o o 6 N o - o 6 6 6 6 0 @ q o6 @ h F F O F < h s < Q F O O O 6 N O O - N 6 = N F Q 9 O< - N N n F N - + O < < 6 O O O O O n O - O 6 @ O F O € F

e o o + o o o o N o e - n o n o - o o o O o o N o o o g on n o

@ o o q - N @ @ n € o o o 8 8 8 R F R E S N S i H S B B S{ - F O 9 n F F A O 6 + i O O O O O 6 F O O O T O @ O P F

O O 6 { O O A O N O - - n @ 6 @ O O o O O O - N O O O g Om n @

O O O n n 6 O @ @ F n n - @ N 6F F 6 F F q F + O n n < n O O O n 6 O Q O O O - N O O 9 <@ N o h N < O - < O m 9 € O O O O O 6 O O O O T O F O I F

O O h T O O O O N O F - n € n € O O O O O O - N O O O O On n @

O A O 6 6 n < N F N E F NF < 6 n 6 F o < n 6 6 n 5 5 6 ; e 6 i d 6 . i o o t X $ i t i< o 6 N O 6 O O 6 ! O h i O O O O O - O O P F _ O F I r @

. . t .O O 6 O O O F O N - - F O n Q F O - O O o O 6 O O O o6 n o

o o o F n N 6 6 n Q F o I -@ m - 6 n @ + 6 - N a o o o 6 o - o 6 N q o o I nO ! < O O F + O O E { + O O O O F I F O - O o e O @ I q F

. . 1 .o m 6 o o o o N - - { a n @ o o o o o O N o o o on n o

o o o G N - F - m $ n o n Q @o N o - q * < 6 6 o N - o o o m o o - r < 9 N N - 9 6m E s n O @ t - N O n < 6 O O o o I n O - - O q O F O t

o @ < o o - o N o - - l o n o - o o o o - N o o o O o6 6 v @

o 6o o o € NO O m n N O N ? < { F OL 6 . . .o < 6 o N o o o o n o o o < o> n v v v N -o

Z E Q\ F \o \ a

F = G

6 m @ n < o q o i o c 8 8 8 S I 9 f i 3 E H S 3 ; $n o o < ! 6 s - { o 6 a 6 o o o - | o I o o o € o F o @ F

O O { O O O N O - - n 6 n O - O O O O N O O O O On n v @

3h @ F 6 ^ ^ o o - 6 o o 5 8 8 8 E n H N S S Q = 8 5 E FF - @ 6 E - F O O O @ < N O O O @ O 6 | O O + F O O O r

. . . . C a . . . . . . t . . . .

X o R a o o o N O o - n i 6 n @ o o - ' o o o N o o o O o

EE Ygo o

ooega :o o o o

6 : f E a a f t ; > x oN r c O O O - v v v O

o o € o o o o r c N F- - - o o c c o o * o - - - - o o c c - o o L@ F < L = - N Q Z Y J T O @ O < < F L = = N J O Z Y L F

s, gglgEn EFgp Fo o o o o - N o o o o o

m o a N N - q 6 n + 9

9 , R 5 5 3 3 3 8 S 8 E G- o o o o o N o o o o o

O O + - < F - n q n N €O E F @ O F 6 - h O 6 € <

6 O < O O - O N O - - 6n n @

o F F Q - 3 6 O N @ F n6 € O O O m n - Q O q N l

6 O h O O O O N O - - nF n o

c o 6 o o o o 6 o@ d o @ o 6 J z t a F

@

d

4

n io

u i

oo

oo

o-

o

-

o oc €o 6

d

r:

I

F


DRAVITE

-- -)3cHonl

'lO '/;/

t5 80 t5.85 t590

o (R)t5.95 t600 t6.o5

Fig' 6. Unit-cell dimensions of tourmaline samples representative of the ten textural types discussed in text. The size of eachbar indicates + ld error in the refined unit-cell dimensions. Dashed reference lines from Donnay and Barton (1972), modified fromEpprecht (1953).

3(o )1 ( o ) 1 ( b ) +-+- -f

t . /

l l tT - /

t| 7 l

I

3$

III

232625223023

3 031

2 7283028

l a C ' 1 1lb B7c2 8 23 ( a ) E I a 23 ( b ) B 7 a3 ( c ) D l

4 815c r l

6 B 7 ( 1 )i c ' 6 b8 8 89 D ' l10 s1

9- + /

Tounvr,clrNE cRysrAL cHEMrsrRy AND STRUCTUREPrevious work on the tourmaline group of minerals is

extensive. The structures of various tourmaline specieshave been solved and refined over the past three decades(Ito and Sadanaga, l95l; Buerger et al., 1962; Barton,1969; Donnay and Barton, 19721' Fortier and Donnay,

Table 4. Unit-cell dimensions of selected tourmaline samplesfrom the Bob Ingersoll No. l. dike

1975; Foit and Rosenberg, 1979; Nuber and Schmetzer,1979; Schmetzer etal., 1979; Nuber and Schmetzer, 198 l;Schmetzer and Bank, 1984). Tourmaline compositions,although quite variable, may be represented by the generalformula XY3Z6(BO3)3Si6O,8(OH,F).. The tourmalinestructure, illustrating site arrangement and coordination,is shown in Figure 10.

One of the main compositional variables of tourmalineis the occupancy ofthe Y-site octahedra. In dravite, theY-site cation is predominantly Mg; in schorl, Fe2*; and inelbaite, Al and Li. The Y site may also contain substantialFe3* as in buergerite (Donnay et al., 1966), Mn (Schmetzerand Bank, 1984), V (Snetsinger, 1966; Foit and Rosen-berg, 1979; Schmetzer et aI., 1979), and Cr in chromiandravite (Dunn, 1977; Nuber and Schmetzer, 1979). TheZ-site octahedra primarily contain Al but may cntain Fe3+(Buerger et al., 1962; Frondel et al., 1966; Hermon et al.,1972; Walenta and Dunn, 1979), Cr'* (Nubor andSchmetzer, 1979; Rumantseva, 1983), V3* (Foit and Ro-senberg, 1979), or Mg (Dunn, 1977; Schmetzer et al.,1979). In most varieties oftourmaline, Na is the dominantcation in the X site; however, appreciable Ca may bepresent as in uvite and liddicoatite (Dunn el al., 1977a,

c ( A )a( A)S a m p l e

Iype nunber C o l o rNunber of

v ( A3) peaks used

Brownish-b lack 15 .924(4)B r o w n i s h - b l a c k 1 5 . 9 3 2 ( 3 )Brownish-green 15 .884(2)E l u i s h - b l a c k 1 5 . 9 4 8 ( 2 )B l u e 1 5 . 9 2 8 ( 3 )B l u e 1 5 . 8 9 6 ( 3 )

B l u e 1 5 . 8 7 5 ( 3 )B l u e 1 5 . 8 i 8 ( 3 )Green 15 .876( 3 )B l u e 1 5 . 8 8 7 ( 3 )G r e e n 1 5 . 8 7 7 ( 3 )B l u i s h - g r a y 1 5 . 8 4 6 ( 3 )P i n k 1 5 . 8 2 6 ( 3 )

7 . 1 3 5 ( 3 ) r 5 6 6 . 1 ( 8 )7 . 1 3 5 ( 2 ) r s 6 8 . 4 ( 6 )7 . 1 1 4 ( 2 ) r 5 5 4 . 6 ( 4 )7 . 1 3 9 ( r ) 1 s 7 2 . 6 ( 3 )7 . r 2 9 ( 2 ) 1 5 6 6 . 6 ( 5 )7 . 1 1 4 ( 2 ) 1 s 5 6 . 7 ( 6 )

7 . 1 r 2 ( 2 ) 1 5 s 2 . 1 ( s )7 . 1 1 2 ( 2 ) i s 5 2 . 8 ( s )7 . 1 1 1 ( 2 ) 1 5 5 2 . 1 ( 6 )7 . 1 r 1 ( 2 ) 1 5 5 4 . 4 ( 6 )7 , 1 1 4 ( 2 ) 1 5 5 3 . 1 ( s )7 . 1 0 3 ( 2 ) r 5 4 4 . 5 ( 5 )7 . 0 e 3 ( 2 ) 1 5 3 8 . 5 ( s )


I O X r t+

F e : . 9 7 F e : . 8 7

483

M 9 , . 2 8 M g : . 3 5

-i---5-1 - c ori!

5 mm-{

Pogmotile - countr yroc f t con toc t : 3 f l

F-

l n l e r i o? o fPegmot i te

->

- c o x l t

F- 0.5 cm ---{

O.60 co lo r lessb lue r g reen

lue O.3O'0.500.70

0.900.70o.500.30

Fig. 7. Examples of compositionally zoned tourmaline crystals. Numerical values are Fe atoms per formula unit. Contours shown

in B and C are schematic to illustrate inferred gowth of crystals. Lines in D are boundaries between color-zones. (A) Tourmaline

crystal of border zone (type 2) with high-Fe, high-Mg core. Longitudinal section shows variation of core composition along length(sample B2). (B) Longitudinal section of type 3 tourmaline crystal, showing spatial relation to pegmatite<ountry rock contact andgowth direction toward interior of pegmatite (sample B7). (C) Type 3 tourmaline from exposure near top of dike (sample Gl l).(D) Type 5 tourmaline, with Fe-rich core and concentric zonation around the c axis. No variation occurs along the length of crystalshown in (2). Kl) Sample E2; (2) sample B16.l

( t )

1977b). Also, K and Mg may occur in minor amounts inthe X site, or it may be partially vacant (Rosenberg andFoit, 1979; Werding and Schreyer, 1984). Tourmaline end-member compositions pertinent to this study appear inTable 5. B is generally assumed to fill the triangular-co-ordinated sites (e.g., Fortier and Donnay, 1975; Ekam-baram et al., I 98 I ; Schmetzer and Bank, I 984). Excess Bmay possibly substitute for Si (Barton, 1969); however,appreciable substitution of B for Si has not been docu-mented.

Complete solid solution occurs between dravite andschorl, and nearly complete solid solution exists between

- 2 mm-l

Inferrod composi t ionol contourso f O .25 Fe i nh r vo lg

1o.93 - c o x i !

schorl and elbaite (Foit and Rosenberg, 1977; Donnayand Barton, 1972). Compositions intermediate to thedravite and elbaite end members are not common al-though limited substitution does occur wherein Mg, Li,Al, and Fe occur together (Deer et al., 1962; El-Hinnawiand Hofmann, 19661' Chaudhry and Howie, 1976). Theremay be a miscibility gap between dravite and elbaite, oras suggested by Foit and Rosenberg (1979), the absenceof intermediate compositions may be due to the fraction-ation of Li and Mg in natural geologic processes and frac-tionation by the tourmaline structure itself due to thesequence ofcation field strengths.

- c o x l S -

I cm ---J

484

Fig. 8. Detailed compositional variation ofzoned tourmalinecrystal in border zone (type 2). Pleochroism as shown. Each pointrepresents an average oftwo to four microprobe analyses in eachzone.


M g

Fe

C o

M n

7 n

. 8 0 9 0 l . o

Fe / (Fe t Mg) ( o toms per fo rmulo un i l )

Fig. 9. Histograms showing range of Fe/(Fe + Mg) values forthe different types of tourmaline.

DrscussroN

Tourmaline as a recorder

The first objective of this project was to determinewhether the compositional variations of tourmaline at theBob Ingersoll No. I pegmatite are systematic and are cor-relative with textures and mineral associations. As indi-cated by the values in Table 3, the compositional char-acteristics (major and minor element) ofeach textural typeof tourmaline are different. Microprobe analyses of ad-ditional samples show similar compositional character fortextural types represented by only one or two samples inTable 3 (see Figs. 15, 16). Smooth compositional varia-tions within single tapered crystals, whose orientationsindicate the direction of growth, clearly demonstrate thesystematic nature of tourmaline compositional variationas the pegmatite-melt system crystallized (analogous tothat ofFoord,1976). In addition, there is a general vari-ation from the border zone to the core of the pegmatiteinvolving Fe depletion and Li enrichment. Such variationis in accordance with the general fractionation behaviorof these elements in magmatic systems and is in goodagreement with previous studies of tourmaline in peg-matitic systems (Staatz et al., 1955; Foord, 1976).

The trace-element characteristics of the tourmaline areneither as consistent nor as systematic in variation as themajor- and minor-element characteristics, with the ex-ceptions of Sn, Co, and Sc. The Co and Sc vary similarlyto Mg, whereas Sn generally increases with Li. There areseveral reasons for the nonsystematic concentrations of

.70

WALL ZONE, F IRST INTERMEDIATE( T Y P E S 3 , 4 , 5 , 6 , 7 , 9 , t O )


Table 5. Selected tourmaline end-member composition

485

the other trace elements. First, several of the elements areessential constituents ofaccessory minerals that occur inthe pegmatite, such as Nb, Ta, and U in Nb-Ta oxides,including microlite. Others are strongly partitioned intomica and feldspar (Rb, Cs) and apatite (REE). For thesetrace elements, the abundance and distribution ofthe min-eral in which they concentrate may complicate their dis-tribution in tourmaline. The second problem results fromthe commonly poikilitic nature of the tourmaline. Eventhe purest crystals commonly contain microscopic solidinclusions, fracture fillings, and unidentified solids inabundant fluid inclusions (cf., Slivko, 1965).

The analysis of trace elements using bulk powders alsoprecludes the determination of variations that probablyoccur in finely zoned crystals, based on the variations inmajor and minor elements. From the results shown inTable 3, it is apparent that compositional variations amongthe major and minor elements are the most useful forestablishing trends. Certain trace elements may also beuseful, but each element must be evaluated within thecontext of the mineralogy of the deposit.

Compositional variation trends

As a first approximation of compositional trends, tour-maline samples may be listed in the following order: coun-try rock-border zone-wall zone-first intermediate zone-third intermediate zone<ore. The general trends of vari-ation of the Y-site cations are shown in Figure ll. Themajor variation is a decrease in Fe2* and an increase inLi + Al(Y), reflecting the dominant substitution 2Fe2* +Li + Al in the Y site. Mg and Ti decrease sharply fromthe country rock to the border zone and then decrease tovery low levels within the inner zones of the pegmatite.Zn and Mn show a general increase toward the core ofthe pegmatite, but with intermediate maxima and mini-ma. Ca (X site) shows an increase toward the core, whichis especially evident at low Fe concentrations (Fig. l7).The compositional trends of Y-site cations are shown interms of major end-member compositions on a temarydravite-schorl-elbaite composition diagram (Fig. 12).

The general compositional trends are related to the po-sition of the tourmaline inward from the country-rockcontact, as illustrated in Figure 13, which shows the samegeneral trends as Figure I l, but also shows the trends ofMq Zn, and Sn. Mn and Sn increase toward the core,while Zn initially increases toward the core but reaches amaximum in the inner wall zone near the core, then de-creases. An increase in Sn concentration has been inter-preted as an indicator of increasing fractionation of themelt (Power, 1968). The trends shown in Figure l3 clearlyreflect an inward fractionation trend, and although theintermediate zones are not represented in this traverse(Fig. 13), the fractionation trends that may be expectedfrom the wall to the core are here established. Further-more, tourmaline samples from the border and wall zonesfrom widely separated locations in the pegmatite are com-positionally similar, which may suggest that initial differ-

Genera l fo rmula x Y3 (B0g )a s i 60 l s ( 0H ,F )a

D r a v i t e

S c h o r l

E l b a i t e

T s i l a i s i t e *

L i d d i c o a t i t e

U v i t e

B u e r g e r i t e

Na M93 A l O (B0S)r

Na r " | t A l o (Bos) r

N a ( L i , A l ) s A l 6 ( B o r ) t

N a ( M n , A l ) : A l o ( B 0 s ) a

c a ( L i , A l ) s A ' l o ( 8 0 3 ) 3

Ca M93 Al 5llg (803 ) 3

Na r "3* A16 (Bog)a

S i 6 0 1 s ( 0 H , F ) 4

S i 6 0 1 s ( 0 H , F ) a

S i 6 0 1 s ( 0 , 0 H , F ) q

S i 6 0 1 s ( 0 , 0 H , F ) 4

S i 6 0 1 g ( 0 , 0 H , F ) 4

s i 6 0 t 8 ( 0 H , r ) 4

S i 6 0 1 8 ( 0 H ) 3 F

Notes :*Hypothetical end renber

entiation of the pegmatite was caused by crystallizationat the walls. Tourmaline samples from the intermediatezones, however, have compositional differences that mayreflect vertical differentiation of the pegmatite.

The general compositional trends are in good agreementwith the results of previous studies (e.g., Staatz et al., I 9 5 5 ;Foord, 1976; Manning, 1982). In accordance with theseresults, the major trend ofdecreasing schorl and dravitecomponents, and of corresponding increase of elbaitecomponent, appears to be a useful indicator ofthe relativedegree of fractionation and the paragenetic sequence ofthe tourmaline and associated primary minerals.

A close inspection of Figure I I reveals considerableoverlap in values from the border zone, wall zone, andintermediate zones. The data points are averages of mul-tiple microprobe analyses, and Li values are obtained frombulk crystal powders. These trends may be studied in moredetail by considering individual microprobe analyses. Al-though Li cannot be determined by microprobe analyses,the good inverse correlation of Li and Fe [or (Fe + Mg)]provides a basis for the use of (Fe + Mg) as an index ofrelative evolution or paragenesis. Each individual micro-probe analysis can be plotted against this index to showthe covariation of the cations and thus demonstrate anysystematic variations with progressive evolution of thetourmaline. Also, by using (Fe + Mg) rather than Li, thecompositional variations in fine-grained, zoned crystalsmay be compared with trends among tourmalines of thedifferent types. In summary, the ideal evolutionary trendwould be reflected by values of (Fe + Mg) or [Li + Al(Y)]normalized to 2[Fe,Mg,Li,Al(Y)]; however, because Licannot be determined directly, the value of (Fe + Mg)may be used as the best approximation.

The (Fe * Mg) range of each of the types of tourmalinemay be used to illustrate their paragenetic sequence (Fig.l4). Although this diagram indicates extensive overlap ofthe various tourmaline types and broad ranges of (Fe +Mg) content, there are systematic compositional zonationtrends that account for much of the apparent range orscatter of the data. For example, in the wall zone, high-


Fig. 10. Portion of the structure of tourmaline projected onto (0001), looking up the c axis. Drawn from atomic coordinates ofthe refined de Kalb tourmaline, given by Buerger et al. (1962); a: 15.95 A, c : 7.24 A. Symmetry-related anion positions shownas O,-Or. Additional BO, groups shown for clarity.

"jf i , ., , ., , ., , , ,lo5r | , r

- '1

- l ' I I i I i t r i r l u n + z n

Fig. I 1. General compositional variations of the Y-site cat-ions in tourmaline of the Bob Ingersoll No. I dike. Data pointsrepresent averages of multiple microprobe analyses.

Fe tourmaline occurs toward the outer portions of thezone grading to low-Fe tourmaline toward the interiorportions (Fig. l6a). In addition, many individual crystalsare strongly zoned, as noted previously, contributing tothe range ofcompositions observed for each textural type.

Z,oning of individual crystals

The compositional variations in individual zoned crys-tals are of particular interest as they mirror variation trendswithin suites of texturally similar tourmalines. One com-mon cause of compositional zoning is rapid crystal growth(Z,oltai and Stout, 1984) during which the diftrsion ratesof select components do not keep pace with the rate ofcrystallization. This may have been the case for the borderzone crystals (e.g., Fig. 8). The compositional zoning isnot simply a smooth variation, but involves a repetitioussequence that may result from one or several changes inthe rate of crystallization. The initial Mg-, \ and Ti-depleted core of the crystal shown in Figure 8 mly haveresulted from an initial abrupt increase in the liquidustemperature of the pegmatite melt. If sufficiently abrupt,a substantial difference between the liquidus and melttemperatures might cause the crystallization of an unsta-ble tourmaline solid solution at a composition interme-diate to the stable coexisting liquid and solid composi-tions, as suggested by Dowty (1980) and Lofgen (1980).Following the nucleation of the unstable composition, ad-ditional crystallizing layers obtain the equilibrium com-position. This zoning phenomenon may be repeated ifanother temperature change occurs, such as from the rapidloss of volatiles from the melt (Lofgren, 1980). The rapidor sporadic loss ofvolatiles, such as by a pulse ofexsolvingfluid from an HrO-saturated melt of the border zone wouldalso provide an efrcient removal of latent heat of crys-

Coudry Bo.darRocl ZoM

I

Fzl

Jf-Eo

GU4

zo

a


Y-SITE OCCUPANCY

487

BORDE RZONE

N :

F F

WALL ZONE

:; ;F F

CORE

F%

Fig. 12. End-member compositions of the Bob Ingersoll tour-maline plotted in terms of schorl (Fe + Mn + Zn), dravite (Mg +Ti), and elbaite (Li + Al).

tallization and possibly a local pressure drop, causing rap-id disequilibrium crystallization.

Alternatively, border-zone crystallization may have be-gun when the pegmatite melt reached a pressure-temper-ature regime favorable for aqueous-fluid exsolution. Withthe first pulse offluid, the liquidus would have been raised,initiating border-zone crystallization. Compositionalvariations in the border-zone tourmaline may reflect dif-ferent element partitioning in a mineral-melt system ver-sus a mineral-melt-fluid system; therefore, if fluid exso-lution during border-zone crystallization occurred inpulses, then the tourmaline would contain repetitious zon-rng.

Another common cause of zoning is the change of in-tensive variables-pressure, temperature, or compositionof the crystallizing medium-during the normal course ofcrystallization (i.e., not necessarily as a result of rapidcrystallization). Changes in lithostatic pressure were prob-ably not important because the pegmatite was emplacedat 3 kbar or greater (Redden et al., 1982); however, vari-ations in P"ro may have been an important factor. Thesmooth compositional variation within single crystals oftype 3 tourmaline in the wall zone (Figs. 7, l6a) reflecteither changes in the composition of the closed melt-fluidsystem with progressive crystallization or a slowly de-creasing temperature, or both. The zoning observed intype 5 crystals appears to have resulted from the depletionof Fe in particular, as progressively more evolved fluidspermeated the wall zone.

The changing composition of the pegmatite melt-fluidsystem with progressive crystallization is closely relatedto (and in part caused by) the stability of a given tour-maline composition at a particular set of pressure-tem-perature conditions. For example, at hrgh temperatures,Fe-rich compositions are more stable than Li-rich com-positions; thus, early tourmaline crystallization contrib-utes to the Fe-depletion of the melt. Although tourmalinesolid solution is complex, tourmaline may be expected tobehave similarly to other minerals with octahedral sites.

Dislonce (fi) from psgmolite-counlry rock contocl

Fig. 13. Compositional variation with respect to distance fromcountry-rock contact. Representative samples from west-centralpart of Bob Ingersoll No. I dike as follows: type lb, country-rockinclusion, Gl l(a); type 2, border zone, Gl I (b); type 3, wall zone,coarse, Ela2;type 5, wall zone, intergown with muscovite, Elal;type 4, wall zone, fine, Elb; type 9, core, D'1. (First and secondintermediate zones absent in this area; third intermediate zoneis thin and contains no tourmaline.)

Of the major Y-site cations, Mg should be preferred athigher temperatures, Fe at intermediate, and Li at lowertemperatures based on ionic size and charge (Tauson,1965). It is therefore important to consider (and to iden-ti&) the different efects of simple depletion or enrichmentof elements in the melt-fluid versus crystal-chemical sta-bility under varying conditions.

Crystal-chemical confol of tourmaline composition

The initial purpose of Figures 15 through 17 was toillustrate how Mg and the four transition metals Ti, Fe,Mn, and Zn covary as the pegmatite crystallized, underthe premise that decreasing (Fe + Mg) is a good index of

a

o

E

o

Eo

E

0 2 0o t 6o t 2o 0 8o 0 4


^ . 1 2f l - n [ } n , . r , , , , r r c o u N r R Y R o c K ( T Y P E l o )

2 . 5 2 3 2 . 1 t . 9 t.7

n h = 2 3

. t -u - l | | , - , f - f - r , - . rT , , T- l - COUNTRY ROCK INCLUSIOi. IS (TYPE lb)2 . 3 2 . t t . 9

BORDER ZONE( TYPE 2 )

D -coo rse g ro i ned , n= 155 I - f i ne g ro i ned ,(TYPE 4 )

n = 3 2(TYPE 3)

W A L L Z O N E

l 7

WALL ZONE ( INTERGROWN

WITH MUSCOVITE )( T Y P E 5 )

FIRST INTERMEDIATE ZONE( T Y P E 7 )

VEIN TOURMALINE, WALL(TYPE 6)

THIRD INTERMEDIATE ZONE( T Y P E 8 )

CORE(TYPES 9A |o )

Fig. 14. Tourmatne paragenesis based on (Fe + Mg) values (atoms per formula unit), shown along abscissa; n: the numberof analyses for each group.

6oEo lc

o 3L O

P A= o

=.8z =

progressive crystallization. Owing to the similarity oftheirproperties, these cations should generally behave as fer-romagnesian elements, and thus each should in turn de-crease as Li increases. The sequence in which these fiveelements should first peak, then diminish as temperaturefalls during crystallization can be predicted, as suggestedby Foit and Rosenberg (1979). Under ideal melt condi-tions (no efects from cocrystallizing phases, no other frac-tionation processes), the cation with the highest fieldstrength should be favored at high temperature and thecation with the lowest field strength, at low temperature.Thus in an ideal crystallization sequence, the elementsmight covary as shown schematically in Figure 18. Cat-ions with successively lower field strengths should ideallybe at their peak concentration in tourmaline at succes-

sively lower temperatures, if their concentrations in themelt are of a similar order of magnitude (Tauson, 19651,Foit and Rosenberg, 1979).

The cation field strength may be approximated by Z/r2in which Z: charge and r: cation radius for octahedralcoordination. Using Zhdanov's (1965) values forthe ionicradii in octahedral coordination (in Bloss, l97l), the ap-proximate field strengths are Ti, 9.8; Mg, 3.7;Fe,3.l;7-n,2.9; and Mn, 2.4.Figure l8 is drawn such that the quantityofeach element is similar to the quantity actually observedin tourmaline from the pegmatite, that is, Fe and [Li +Al(Y)l are the principal substituents, and Mg, Ti,Zn, andMn are subordinate. The (Li * Al) couple may be ex-plained by the same reasoning if the field strength of Li(-2.1) is the governing factor. Al has a high field strength

Fe + M9 (per formulo uni i )

489JOLLIFF ET AL.: TOURMALINE AS A RECORDER OF PEGMATITE EVOLUTION

o ooo

o

Uo-

F

I9 g= o

N lE I

tt't Il fI U

EEE o

! . ro

o iO O

; 9' E

-l

. o8of, '"+ E,R.c

6 X . dr d b O t s

EECN ' O

x u - -; '> :'Q , ; 9v t t I

" i 3E:=ob .X 5i i E E

d h

Ic.oC'c.octe

|!

ctl.=U'C'o(tE

=c,J

o=Eo

oo-

@

o

o

N

lrlzo^N N(fuUJLo >E F

.D

anEo

RgE l=+

q oN L

N - oc i o

o oq oF g o r t po o c i o o

44

x

i<

rf<*

&eF,oa€a a

:'4i. iFtt'

G

lrJ

trF

=T Et:t ;e6r.rQo- -ia,,Jzo-ioeE= >E'r.f FQ ZFf ,

58o l

it

( lUn blnu.r.rol rad rI 'uZ'uyrl '6y1 'aJ 3 \\ t tunoff i /NolrcvujYrorv


u o r l c o J l u o l v

490

\.q\, s

uo! l3oJ l u ro lv

N ! 6E = :; : 3 .. P " F: ? 6

9 - ;

E ! $E 9 E@ E ES

F > l

9 9^ - i

EE ; F; ; a o

E; : .EE: € 'o i :E. T E . : N

' =E iF } . E I

TI

It ^

_ c P

I e iI F c

^ 1

e

;d

III

II

I,L

o Norrcvul v{orv


Fe +Mg (otoms p€r formulo unil )

Fe + Mg ( otoms per formulo unit )

Fig. 17. Variation in concentration of Ca in Bob Ingersoll tourmaline in terms of Ca atoms per formula unit vs. (Fe + Mg).

49r

8. :E 5

09 - s.9 ! o.,o 3o

o

E

oo

ooo5

O o

and thus readily fits the Y-site octahedra, its concentrationapparently limited only by charge-balance requirements.

Figure 19 is similar to Figure 18, but is constructedfrom compositional data in Table 3. The similarity oftrends shown in Figures 18 and 19 indicates the impor-tance of crystal-chemical control of composition as aninfluencing factor on the overall compositional trends ofthe tourmaline. Recogrrition of this factor allows furtherinterpretations regarding additional factors, such as dif-ferentiation due to the coexistence of melt and aqueousfluid, or the fractionation effects of other crystallizing min-erals.

Processes of crystallization

One of the foremost goals of this research is to deter-mine whether aspects of tourmaline composition and tex-ture or habit reflect the nature of the media from whichthe tourmaline crystallized. Three possibilities are sug-gested by three distinct trends of Fe and Mg variation, asfollows: boron metasomatism, crystallization from sili-cate melt, and crystallization from a melt-aqueous fluidsystem.

The first trend of Fe-Mg variation is that of tourmalinein the country rock and in inclusions ofcountry rock inthe pegmatite. Such tourmalines exhibit a range of Fel(Fe + Mg) from 0.544 to 0.799 (Fig. 9), with the lowestvalues generally corresponding to the highest (Fe + Mg)values. The lowest Fe/(Fe + Mg) values obtained for tour-maline appear to result from early replacement of biotite,prior to extensive re-equilibration with pegmatite fluidsenriched in Fe relative to Mg. As the biotite re-equili-brated, its Fe/(Fe + Mg) value increased, as did that oftourmaline replacing biotite. Thus the zoned tourmalinein the country rock formed with low Fel(Fe * Mg) coresthat were mantled and thus protected from later re-equil-ibration by rims of successively higher Fe/(Fe + Mg). Thebiotite re-equilibrated completely, whereas the history ofthe re-equilibration was retained in the zoned tourmaline.

The second trend is characteristic of tourmaline fromthe pegmatite border zone. The Fel(Fe + Mg) values rangefrom 0.718 to 0.938, but there is actually very little over-lap between these values and values for country-rock tour-maline (Fig. 9). Border-zone tourmaline crystals are zoned,with high-Fe, high-Mg cores, and the Fe/(Fe + Mg) ratioincreases as (Fe + Mg) decreases (e.g., Fig. 7a). One in-ference is that this Fel(Fe + Mg) ratio is governed by theFe and Mg content of the pegmatite melt. Alternatively,the border-zone tourmaline may have been uniformlycontaminated by Mg of the country rock, and the valueof Fel(Fe + Mg) a function of the equilibrium at theprevailing pressure, temperature, and composition. In anycase, the Fe/(Fe + Mg) ratios reflect different processesof formation: country-rock tourmaline formed by boronmetasomatism of Fe-Mg minerals in the country rock,andborder-zone tourmaline by crystallization from a meltor fluid-melt system.

The third trend oftourmaline compositions on the basisof Fe and Mg values is that of samples of tourmalineinterior to the border zone (with the exception ofthe thirdintermediate zone, discussed separately). All such tour-malines consistently contain Mg in quantities below 0.10

Fig. 18. Schematic illustration of ideal covariation of Y-sitecations of tourmaline in response to decreasing temperature andincreasing fractionation of melt.

ozt t r- Uo oo f fU O

DECREASTNG TEMPERATURE +INCREASING FRACTIONATION


+ F c

o M g

x L i

o z n

o M nco

oL

Fo O+

Trcnd of M9-depleted tourmol ine_-.r__ 1_*_ /_T r e n d o l M 9 - b s o r i n g l o u r m o l i n e

- - t : ' - - *

- - - - \ - - - - - - - -+ - - - - , U<t - ' - - - + + \ - x + ' ) * '

^

+ t*.a

x x/-rn"tai-Jrt--.-/ Y** '

-/- r^ t1

2.O o 8

Fe + Mg (oloms per formulo unit)

Fig. 19. Covariation of Y-site cations of Bob Ingersoll tourmaline with decreasing (Fe + Mg). Abscissa plotted as Fe + Mg asan approximation of increasing fractionation of the melt-fluid system. Data are from Table 3 (cf. Fig. l8).

- x - x / - x

wto/o MgO (and most below 0.01). To further emphasizethis, crystals that contain less than 0.10 wto/o MgO occurin the wall zone within inches of the border zone (Table3). Values of FeO in these same tourmaline samples, how-ever, are greater than FeO values of border-zone tour-maline. It is proposed that the differentiation of Fe andMg is due to the presence of an exsolving aqueous fluid.This is supported by the very coarse grain size of tour-maline immediately interior to the border zone (outer wallzone). The coarse grain size may be attributable to en-hanced diftrsion and crystal growth in an aqueous me-dium (Jahns, 1953, 1982).

Other compositional aspects of tourmaline in the thirdtrend (Mg depleted) provide evidence of crystallizationfrom a fluid-melt system that was differentiated relativeto the border zone. Aside from Mg depletion, Zn and Mnbecome enriched as Fe decreases with inward crystalli-zation of the wall zone (Fig. l6a). Additionally, Fe, Mn,and Zn concentrations constitute well-defined trends asFe decreases, implying that these elements were efficientlydifferentiated and approached the ideal crystal-chemicalconcentration in tourmaline at the given pressure, tem-perature, and composition of the crystallizing medium.These trends and the habit of tourmaline in the outer wallzone may indicate the presence of an aqueous-fluid in-terface between melt and crystallizing tourmaline. Alter-natively, exsolved fluid rising through the melt may havecaused differentiation by depleting lower portions of meltof the elements that favor the fluid, thus enriching meltin upper parts of the pegmatite in these same elements.Evidence supporting this process is the occurrence oftype3 tourmaline predominantly on the hanging wall of thepegmatite.

The differentiation ofan aqueous fluid exsolving from

the silicate melt should have several effects on tourmalinecompositions. Certain elements such as Zn and Mn shouldfavorably partition into the aqueous fluid from the silicatemelt and thus be enriched in the tourmaline. Holland(1972) found Zn and Mn to be enriched in the aqueousfluid (Zn > Mn) in the presence of Cl. Fe is also favorablypartitioned into aqueous fluid coexisting with silicate melt(Burnham, 1979), but Mg is strongly retained in the melt(Holland, 1972). Another effect on composition of crys-tallizing tourmaline results from the increased diffirsivityof elements in an aqueous phase, the viscosity of whichmay be seven orders of magnitude less than the silicatemelt (Jahns, 1982). With increased diff.rsivity, diffusingions should respond more efficiently to chemical potentialgradients, and therefore the composition may approachthe ideal crystal-chemical-controlled (equilibrium) con-centrations of each of the elements involved in tourmalinesubstitution.

The halogens F and Cl should also show characteristicpartitioning behavior in the presence ofaqueous fluid andsilicate melt. Cl concentrates in the fluid, and F prefer-entially remains in the melt (Hards, 1976; Fuge, 1977;Burnham, 1979; Manning, l98l). The ratio of these ele-ments in tourmaline should therefore provide insight asto the presence of aqueous fluids; however, Cl has notbeen determined in this study. F contents (Table 3), whichare lowest in tourmaline of the outer wall zone, indicatethat the crystallization of this tourmaline (type 3) mayhave occurred in the presence ofan aqueous-fluid interfacebetween the growing tourmaline and silicate melt, or thatthe crystallizing medium was diluted with respect to F byan abundance ofaqueous fluid.

On the basis of the F content of all the tourmaline ofthe Bob Ingersoll No. I dike, the pegmatite is F-rich (cf.,

o.44t 62 .42 6

Nemec, 1969). Amblygonite-montebrasite in the Bob In-gersoll No. l. dike contains -3 wto/o F (&*o = 0.25), andlepidolites contain as much as 7 wto/o F (F/OH . 0.8).From the preponderance offluorapatite over chlorapatitein Black Hills pegmatites in general (Roberts and Rapp,1965; Papike et al., 1984), the pegmatite was probablynot enriched in CI (e.g., F = 3-4 wt0/0, Cl < 0.01 wto/o inapatite, Jensen, 1984; cf., Roegge et al., 1974).It followsthat although elements such as Zn and Mn were enrichedin the aqueous fluid, the enrichment was only by a smallfactor and they were not completely extracted from themelt.

There is also textural evidence from the coarse tour-maline of the outer wall zone (type 3) for the presence ofan aqueous fluid during crystal growth. External, heter-ogeneous nucleation of these crystals, with the inward-crystallizing wall of the magma chamber as the substrate,is indicated by both the paucity and irregular distributionof crystals, and also by their orientation relative to thewall (Lofgren, 1980). Such nucleation characteristics com-bined with the very coarse grain size suggests crystalli-zation in the presence of a medium of reduced viscosity,e.g., melt and exsolving aqueous fluid. Additionally, theorientation of type 3 tourmaline crystals is analogous toan extremely coarse grained variant of comb layering, theorigin of which has been similarly interpreted (Moore andLockwood, 1973).

The trend of MnandZn enrichment in tourmaline con-comitantly with progressive crystallization of the wall zone(Fe depletion), holds well for coarse-grained (type 3) andfine-grained (type 4) tourmaline ofthe wall zone. Althoughtexturally different, the similarity of trends indicates asimilar crystallization mechanism for both. The fine-grained tourmalines simply grew at a stage when much ofthe surrounding portion ofthe wall zone had previouslycrystallized, consistent with their occurrence in the inte-rior parts of the wall zone and their subhedral texture.The location of the fine-grained crystals and the orien-tation of coarse-grained crystals perpendicular to the bor-der zone-wall zone contact are evidence of an orderlyinward crystallization of the wall zone. However, tour-maline that occurs with coarse-grained segregations ofmuscovite in the wall zone (type 5) deviates somewhatfrom the trend (Fig. 16b). Some ofthese crystals are stronglyzoned (Fig. 7d), commonly with rims very low in Fe. Athigh Fe values, Zn and Mn fall on the main wall-zonetrend (see Fig. l6b), whereas at values between 1.0 and0.7 Fe atoms per formula vr;lit, Zn and Mn are generallylow relative to other wall-zone tourmaline. This may bedue in part to the phengitic nature of the coexisting mus-covite (Table 6). Although muscovite may incorporatesubstantial amounts of the Y-site cations of tourmaline,the amounts are quite variable. Cation exchange withmuscovite as the tourmaline crystallized may have buff-ered the Mn and Zn concentrations, as fluids became pro-gressively more Fe depleted. It is evident from the rangeof Fe contents of different samples of type 5 tourmaline

493

Table 6. Values ofselected oxides for coexisting tourmalineand muscovite


sampl enumber *Fe0

Tr* M** r

Hn0

T i l

ZnO Zone

T N

foumal i ne

T M

0 . 5 2 0 . 1 50 . 2 9 0 . 1 20 . 6 9 0 . 1 70 . 2 1 0 . 0 30 . 4 4 0 . 0 80 . 5 9 0 . 0 40 . 5 2 0 . 0 70 . 3 9 0 . 0 50 , 7 4 0 . 0 50 . 3 7 0 . 0 40 . 3 8 < 0 . 0 11 . 6 4 0 . 2 3

G 18 b bG l tc 7 bE l a l8 l bF 5c 1E l bG4cBt20 ' 1

4 . 6 7 0 . 9 76 . 7 9 < 0 . 0 16 . O 2 0 . 7 2s . 9 6 0 . 2 45 , 8 8 0 . 9 15 . 1 0 0 . 9 52 . 9 5 0 . 1 45 . 6 2 0 . 8 04 . 8 4 0 . 6 48 . 2 5 0 . 6 33 . 5 0 0 . 1 30 . 9 2 0 . 0 7

0 . 4 8 0 . 3 30 . 0 1 < 0 . 0 1

< 0 . 0 1 < 0 . 0 1< 0 . 0 1 < 0 . 0 10 . 0 2 0 . 0 2

< 0 . 0 1 0 . 0 50 . 0 1 0 . 0 1

< 0 . 0 1 < 0 . 0 10 . 0 1 0 . 0 10 . 0 6 0 . 0 30 . 4 7 < 0 , 0 1

< 0 . 0 1 < 0 . 0 1

0 . 4 3 0 . 1 7 B o r d e r1 , 2 7 O . 4 2 U a 1 11 . 3 7 0 . 2 6 l a l l0 . 7 8 0 . 0 5 H a l l0 . 9 4 0 . 0 6 | ' l a l l0 . 3 9 0 . 0 5 i l a l l0 . 3 7 0 . 1 4 i l a l l1 . 2 2 0 . l l i l a l l1 . 5 7 0 . 0 9 { a l l1 . 0 8 0 . 0 6 l s t l n t .0 . 1 3 0 . 0 3 3 r d I n t .1.09 0.06 Corei t* i

N o t e s :r V a l u e s a a e i n t r . g* r T - t o u r [ a ] i n e*** 14 - muscovite* * a * 0 r 1 m i c a i s l e p i d o l i t e : L i 2 0 = 4 w t . % , F : 5 w t . X

that the crystallization of this assemblage occurredthroughout the crystallization of the wall zone (Fig. la).

Tourmaline of the flrst intermediate zone (type 7) iscompositionally similar to type 5 tourmaline (Fig. 16d).The range of Fe concentrations suggests an overlappingparagenesis with wall-zone tourmaline; however, the com-positional trends of tourmaline from this zone may beincomplete owing to sampling restrictions imposed byprevious mining of the interior parts of the zone. Tour-maline was not an early-forming mineral in this zone,based on its occurrence as an anhedral constituent of thematrix between feldspar megacrysts. This suggests crys-tallization from the interstitial, residual melt followingearly K-feldspar crystallization. The residual nature of theliquid is supported by a slight Mg content, and by elevatedSc and Co concentrations of one of the samples (G4c,Table 3).

Tourmaline from the third intermediate zone (type 8)is compositionally similar to border-zone tourmaline (Fig.l5) and almost identical to the outer rim of the crystal inFigure 8 except that Mn and Zn ate depleted relative tothe border-zone tourmaline. Type 8 tourmaline is en-riched in Mg, Co, Sc, and F and depleted in Fe, Mn, andZn relative to type 3 tourmaline. On the basis of thefractionation of these elements between coexisting silicatemelt and aqueous fluid, as discussed earlier, it appearsthat this composition reflects crystallization from a resid-ual magma that had previously undergone aqueous-fluidseparation. Additionally, the distribution of the tourma-line crystals is regular, and they have no preferred ori-entation (Fig. 5b), both of which support homogeneousnucleation and growth directly from the melt (Lofgren,1980). This raises the question of how exsolution ofaqueous fluid can cease once it has begun in a saturatedmelt.

A saturated granitic melt should contain at least -10

wto/o HrO at 3 kbar (Tuttle and Bowen, 1958); thereforeonce saturated, the crystallization of any minerals, all ofwhich contain less than l0 wt0/0 HrO, should cause the


remaining melt to be continuously oversaturated. How-ever, from recent work by London (1986), the solubilityof HrO in a silicate melt may be increased significantlyby the buildup of an alkali borate (specifically LirBoO,)component in the melt. This may have been the case inthe third intermediate zone in which tourmaline occursonly at the inner contact with the core. Otherwise, thethird intermediate zone consists of a barren assemblageof quartz and Na-feldspar. The reason for tourmaline in-stability throughout most of the zone is unclear, althoughthe coordination state of 83* in the melt may be a factor(Chorlton and Martin, 1978; Pichavant, l98l; London,1986). The implication is that in the absence oftourmalinecrystallization, the activities of Li, F, and B in the meltincrease and thus allow increased solubility of water. Thisresults in melt depolymerization (Burnham, 1979) andthus an increase in the availability of structural sites inthe melt for network-modifying cations. When tourmalinebecomes stable and crystallizes, the LirBoO, componentis removed from the melt, water exsolves, and a diversemineral assemblage results from the loss of structural sitesin the melt (London, 1986). The presence ofthe accessory-mineral assemblage of tourmaline, beryl, apatite, and cas-siterite at the contact of the third intermediate zone andthe core supports this model.

The Mg content of type 8 tourmaline poses a problemrelating to the enrichment of Mg during the crystallizationof the pegrnatite. The normal compositional variation withprogressive crystallization in a magmatic system shouldinclude Mg depletion. If the Mg content of the melt isprogressively depleted following border-zone crystalliza-tion, how then does type 8 tourmaline form with the sameFel(Fe + Mg) as border-zone tourmaline? This may resultfrom an increase in Mg concentration in residual meltduring the hiatus in tourmaline crystallization at the timeof formation of the third intermediate zone and from thepreferential loss of components other than Mg to previ-ously exsolved aqueous fluid.

The similarity of Fe/(Fe + Mg) values between types 2and 8 tourmaline raises another question: is Fe/(Fe + Mg)in tourmaline temperature dependent in a given system,e.g., melt-crystals? If so, the temperature of crystallizationof the border zone may have been nearly the same as thatof the third intermediate zone. This may be possible ifthe border zone formed by a rise in the solidus temper-atures on the order of 25 to 50'C. The crystallization ofthe pegmatite through the third intermediate zone maywell have taken place over a narrow temperature interval.The solidus may have been less than 600t owing ro thehigh B content (Pichavant, I 98 l) and perhaps even lowerbecause of the combined effects of HrO, B, F, and Li(Manning et al., 1984). Once the pegmatite began to crys-tallize and volatiles were gradually removed from the sys-tem, crystallization may have proceeded under the influ-ence of only a slight temperature gradient.

Tourmaline of the pegmatite core (types 9 and l0) isMg depleted and ranges in Fe content from 0.5 atoms per

formula unit to below the detection limit of the micro-probe (Fig. l6e). The two distinct varieties of tourmalineof the core may reflect both crystallization from the re-sidual melt and the presence ofthe last portion ofexsolvedaqueous fluid. The light-bluish-gay tourmaline (type 9),which occurs in a fine-grained intergrowth of quartz,cleavelandite, and lepidolite, may result from crystalli-zation ofthe last portion of residual melt. This tourmalineis embayed and crosscut by the other minerals; therefore,it formed early in the crystallization of the core. The lackof a continuum in Mg and Fe contents from third-inter-mediate-zone tourmaline to core tourmaline may resultfrom a sampling gap, or it may reflect a dilution of theseelements due to their uptake by lepidolite as well as bytourmaline (Table 6).

Pink elbaite (type l0) is the most highly evolved tour-maline in the pegmatite in terms of Li enrichment and Fedepletion. The commonly euhedral form and splayed hab-it, and the apparent preferential occlurence ofthis tour-maline with small quartz segregations (Norton, 1983), mayreflect crystallization in the presence ofthe latest-exsolvedaqueous fluid. F depletion in type l0 relative to type 9tourmaline may also indicate crystallization of type l0 inthe presence ofan aqueous fluid; however, dilution ofFby incorporation in lepidolite is also possible. The distri-butions of the minor elements (Mn, Zn, Ti, Ca) betweentypes 9 and l0 tourmaline are not easily interpreted interms of differentiation between the aqueous and mag-matic phases. Manganese is most concentrated in type l0tourmaline, and the Zn concentration decreases relativeto type 9 tourmaline. This may result from the crystal-chemical preference of the tourmaline structure. If, how-ever, Mn andZn are depleted in the residual melt becauseof preferential partitioning into the coexisting aqueousphase, why are they not depleted in tourmaline that isinferred to have crystallized from the residual melt (type9)? The partition coefficients based on a simple Cl-richsystem may be an order of magnitude too high for anF-rich system, owing to the complexing effect of F (Man-ning, l98l). Although Mn and Zncan be effectively frac-tionated between the aqueous phase and silicate melt byCl complexing, they may not be completelyremoved fromthe melt (in an F-rich system), as might occur in a Cl-richsystem. It may therefore be possible for Mn and Zn topersist in the residual melt. The reason that the two dis-tinct tourmaline types (9 and l0) follow the same trends(Fe and Zn depletion; Li, Mn, and Ca enrichment) maybe that Cl was depleted and F dominant during core for-mation, and therefore the crystal-chemical behavior (sitepreference) of tourmaline was not overprinted by a strongfractionation of elements between the melt and aqueousfluid.

The reason for the increase in Ca concentration oftour-maline throughout the pegmatite, for samples with rela-tively low total (Fe + Mg), is not certain; however, thistrend was also noted by Staatz et al. (1955) and Foord(197 6). Staatz et al. suggested that this might be a crystal-


Hoplogron i le min imum o l

PHOO = 4 hb, futitc ond Bos.n' 1958

Avarogr nornolivr comporil ionof pagmolltr tona!

NoAlS i30s KAlSi30s

Fig. 20. Plot of approximate compositions of pegmatite zones normalized to 100 wto/o quartz-albite-orthoclase. Minimum at 4

kbar in Qz-Ab-Or-HrO haplogranite system (Tuttle and Bowen, 1958) shown for reference. Dashed circles show general trends of

estimated bulk compositional variations of pegmatite zones.

chemical response, driven by charge-balance require-ments ofthe tourmaline. The occurrence of the Ca buildupin high-Li tourmaline regardless of zonal location or tex-tural type supports this suggestion. Another possible causeis the continuous subsolidus release of Ca from early-formed plagioclase as it equilibrates with progressivelymore-evolved, lower-temperature fluids (Jahns, I 982). Theobserved increase in Ca content of apatite from the wallzone to the core at the Bob Ingersoll No. I dike (Papikeet al., 1984) and the occurrence of microlite, a Ca-Natantalum oxide, in the pegmatite core support this pos-sibility. Additionally, Iondon and Burt (1982a) have not-ed postmagmatic Ca-for-Li and Ca-for-Mn exchange inprimary phosphates. An actual increase in the activity ofCa is favored over the charge-balance argument as thecause of increasing Ca in late tourmaline because the mostCa-enriched tourmaline contains Al(Y) in excess of Li, acondition that does not appear to require excess positivecharge for balance.

Sequence of crystallization of pegmatite zones andinternal evolution

Textures observed at the Bob Ingersoll No. I dike cou-pled with compositional data on tourmaline provide in-sight to the sequence of crystallization of the pegmatite

zones and possibly to mechanisms that were in part re-sponsible for the compositional differentiation of the peg-matite. One possible model into which the data can beintegrated with current thinking (esp. Jahns, 1982) re-garding pegmatite internal evolution is described below.

The early stages ofcrystallization during orimmediatelyfollowing emplacement of the pegmatite consisted of aninitial B metasomatism of the country rock and the for-mation ofthe border zone. The initial solidus temperaturemay have been depressed to well below 600'C on the basisof the HrO, Li, B, and F enrichment of the pegmatite(Chorlton and Martin, 1978; Stewart, 1978; Manning,l98l;Pichavant, l98l;Manning et al., 1984). The fine-grained border zone may represent a "quench" due tovolatile loss (causing an abrupt increase in the solidus).The border zone is silica rich (Fig. 20) relative to thehaplogranite minimum and is also peraluminous. Theperaluminous nature may only be apparent, resulting froma loss of alkalis, especially K, to the country rock (Norton,1983). Alkali loss has been documented for many BlackHills pegmatites (e.9., Sheridan et al., I 9 5 7; Redden, I 963 ;Shearer et al., I 986) and specifically for the Bob Ingersollpegmatite (Tuzinski, 1983). Zoned tourmaline crystals inthe border zone may reflect enrichments and depletionscaused by initial pulses ofexsolving aqueous fluids. The

tordar | \7o"f

-'t x

'\ honging

' . , - l - wol l

--1 jjaa;yffi"%*i * t t r

!crrr"4 .l

-t. ;. . ,


Mg content of the border-zone tourmaline probably re-sults from trace Mg in the melt and from contaminationby country rock.

In the wall zone, the presence of large tourmaline crys-tals that are perpendicular to and tapered toward the bor-der zone indicates unobstructed crystallization from theborder inward. The coarse grain size provides evidencefor the presence of exsolved aqueous fluid (Jahns andBurnham, 1969). The aqueous fluid may have existedinitially along growing-crystal boundaries or as fine, dis-persed droplets in the melt (Jahns, 1982). The composi-tion of tourmaline in the outer wall zone reflects the pres-ence ofthe aqueous fluid in terms of the efficient, systematicdifferentiation of Fe, Mg, Mn, Zn, and,F.

Tourmaline paragenesis, based upon the extent of thesubstitution of (Li + Al) for (Fe + Mg), suggests overlapin the timing of crystallization of the interior part of thewall zone, the first intermediate zone, and the third in-termediate zone. The crystallization of these three zonesmay be complementary. The first intermediate zone isgreatly enriched in K, based on the modal abundance ofK-feldspar, whereas the third intermediate zone and muchof the interior of the wall zone are depleted of K. Whetherthe reason for this differentiation lies with the responseto a temperature gradient or with melt-site preference forNa relative to K, the first intermediate zone, which formeda broad thick hood in the upper portion of the pegmatite,was greatly enriched in K, a typical feature in Black Hillspegmatites (Norton and others, 1964). Transport in a fluidphase that rises through the melt because of its lowerdensity is an appealing mechanism to achieve such dil:ferentiation. Although K rnay also have been depletedfrom the top of the pegmatite by the exsolving aqueousfluid that moved upward into the country rock, it wasreplenished from lower zones. The K-feldspar-bearing zonetherefore crystallized with a bulk composition far re-moved from the quartz-K-feldspar-albite minimum to-ward the K-feldspar-quartz sideline (Fig. 20) or towardthe K-feldspar apex. The effect of wholesale removal ofK from lower parts of the pegmatite is to displace thecomposition of the rest magma from the ternary mini-mum toward the quartz-albite join, or locally even intothe quartz or albite single-phase fields, but not far fromthe minimum, on the basis of the presence of minoramounts of muscovite or K-feldspar in the inner wall zoneand third intermediate zone. Additionally, the displace-ment from the minimum composition should result inforced crystallization of quartz and albite (Jahns, 1982).

As the composition of the residual melt moved awayfrom the minimum composition, its ability to retain waterin solution, in terms ofweight percent, may have increased(Fig.27, Tuttle and Bowen, 1958). Alternatively, the lackof tourmaline crystallization in outer parts of the thirdintermediate zone allowed the component LirBoO, (or oth-er alkali borate) to concentrate in the residual melt, caus-ing increased H,O-silicate liquid miscibility as shown byLondon (1986). The significance of these compositional

shifts is that the magma forming the third intermediatezone was forced to at least partially crystallize and thatthere may have been a hiatus in the aqueous-fluid exso-lution. Tourmaline of this zone does not show the com-positional or textural characteristics that are inferred toresult from coexistence with an aqueous-fluid componentwith the melt. Tourmaline eventually became stable andresumed crystallization. Removal of LirBoO, from the meltcaused HrO exsolution to resume, which in turn causedincreased melt polymerization and deposition of beryl,apatile, and cassiterite. Another important compositionalaspect ofthe third intermediate zone is the abundance ofquartz (relative to cleavelandite), which appears to in-crease gradually, beginning in cleavelandite-rich portionsof the wall zone. The increase in modal quartz is alsoconsistent with the possibility that the third intermediatezone is a product of a residual magma. As such, it shouldbe enriched in F, and this is indeed indicated by the highF content of third-intermediate-zone tourmaline. An in-crease in F concentration in the silicate melt has the effectof shifting the quartz-albite join away from the quartzapex (Manning, l98l), thus causing an increasing pro-portion of quartz crystallization.

As a consequence of pegmatite crystallization, the ac-tivity of Li in the residual silicate melt increases (Stewart,1978). Eventually, with the continued crystallization ofquartz and alkali feldspars, the composition ofthe residualsilicate melt reaches a minimum in the quartz-alkali feld-spar<ucryptite (LiAlSiO4) ternary system (Fig. 2l), andone of the lithium aluminosilicates begins to crystallize.At the Bob Ingersoll No. I dike, the primary Li mineralthat formed is lepidolite, probably owing to the abundanceofF (London and Burt, 1982b; eern! and Burt, 1984).This zone (quartz-cleavelandite-lepidolite) forms thestructural center of the dike and is inferred to be the corein the sense that it was last to crystallize, on the basis ofthe occurrence of highly evolved, primary tourmaline.

CoNcr,usroNs

Tourmaline from the internally zoned Bob Ingersoll No.I pegmatite dike occurs as a number of texturally andcompositionally distinct types that are confined to partic-ular zones. The tourmaline compositions are mostly alongthe schorl-elbaite join, ranging from slightly over 500/o ofthe schorl end member to nearly end-member elbaite.Several of the groups contain appreciable Mg; Zn, Mn,Ti, and Ca are minor constituents. Tourmaline exhibitssystematic compositional variations that indicate B meta-somatism of the country rock, followed by the progressivecrystallization of the pegmatite in the sequence borderzone to wall zone to intermediate zones to core. The majorsubstitution corresponding to proglessive evolution of thepegmatite is (Li + Al) replacing (Fe2* * Mg) in the Y site.This is generally consistent with the findings ofother stud-ies. However, variations in minor-element concentrationssuggest that a complex crystallization relationship existsamong the intermediate zones of the pegmatite.


EucryptiteLi Ats i04

Fig. 2 I . Plot of composition of pegmatite core, normalized to 100 wto6 combined alkali feldspars , quart4 and eucryptite (after

Stewart, 1978). Minimum of the system Ab-Qz-eucryptite-HrO at Pr,o:2 kbar and approximate 6eld boundaries for Li minerals

in general (Stewart, 1978) shown for reference.

FeldsporA b + O r

Interpretations regarding the nature of crystallizationrelationships of the zones and the nature of crystallizationmedium depend upon the recognition ofthree factors thatafect the compositional variation of tourmaline: (l) bulkcompositional evolution of the pegmatite melt-fluid sys-tem during progressive crystallization; (2) the crystalli-zation process, particularly the nature of the crystallizingmedium; and (3) crystal-chemical preferences of the tour-maline structure in response to changing physical condi-tions. Regarding the second factor, three processes iue ofconcern: metasomatism, crystallization from a silicatemelt, and crystallization from a silicate melt in the pres-ence of an exsolving aqueous fluid. Of particular impor-tance is the differentiation capability of the separateaqueous fluid, whether in a "static" sense or in the senseof a mobile fluid that differentially sqlvenges elementsfrom one portion of the melt and enriches the melt else-where. By distinguishing the effects of these factors, com-positional variations of the tourmaline may be useful inthe interpretation of crystallization processes that affectedthe differentiation of a given pegmatite.

Textures and crystal-zoning trends support the com-positional trends and are consistent with the followingmodel for the Bob Ingersoll No. I pegmatite. The border

zone crystallized as a quenched margin owing to volatileloss during an initial pulse of exsolving fluid synchro-nously with B metasomatism and formation oftourmalinein the country rock. The pegmatite initially crystallizedfrom the walls inward and exsolved aqueous fluidthroughout wall-zone crystallization. The aqueous fluidprovided an effective medium for differentiating the peg-

matite melt with respect to several constituents, includingthe Y-site cations of tourmaline and the alkalis Na andK. As the aqueous fluid concentrated in upper parts ofthe pegmatite, K became enriched and formed the K-feld-spar-rich first intermediate zone. The depletion of K re-sulted in the increasingly sodic nature of the inner wallzone and the adjacent third intermediate (quartz-cleave-landite) zone in lower parts of the dike. The suppressionof tourmaline crystallization coupled with the depletionof K in the third intermediate zone changed the bulkcomposition of the rest magma, and crystallization of thequartz-cleavelandite zone occurred without concomitantexsolution of aqueous fluid. When conditions again be-came favorable, tourmaline crystallization resumed, de-pleted the melt in Li and B fluxing components, and con-tributed toward renewed exsolution of aqueous fluid. Therest magma achieved Li saturation and crystallized a min-

Minimum in th€ 3yltcmAb-Qz-Eu- Hlo or Pli i 2 kb


imum-melt assemblage of quartz-cleavelandite-lepidoliteat the core of the pegmatite, which also contains the most-evolved tourmaline, Fe-free elbaite.

Many of the interpretations presented in this study arespeculative. Research in several specific areas may pro-vide additional insights. These include the stability oftourmaline and its equilibrium composition, in terms ofmajor end members, under diverse conditions. Data per-taining to the partitioning behavior of elements in thesystem melt-fluid-crystals, with different combined con-centrations of the elements F, Cl, and B, are also criticalto improved understanding of chemical differentiationduring the internal evolution of zoned pegmatites.

AcrNowlnncprrNrs

We are greatly indebted to the following individuals: V. Jensen,word processing; R. Talbot, drafting; D. Gosselin and A. Settlefor assistance with mapping and sampling; C. I-arive, fluorinedeterminations; S. Simon, computer applications; W. L. Robertsfor supplying sample Sl; J. C. Laul and Battelle, Pacific North-west Laboratories, rNAA; B. L. Davis, X-ray diftaction patterns;and E. E. Foord for providing tourmaline sarnples for analyticalcomparison. The manuscript was greatly improved by the adviceand review of J. A. Redden, E. E. Foord, D. London, R. V.Dietrich, P. J. Dunn, J. J. Norton, and R. J. Walker. This workwas supported by the National Science Foundation under NSFGraduate Fellowship RCD-8450127 (B. L. Jolliff) and by theU.S. Department of Energy under DOE grant DE-FGOI-84ER13259 (J. J. Papike).

RnrnnrNcns

Barton, R., Jr. (1969) Refinement of the crystal structure of buer-gerite and the absolute orientation of tourmalines. Acta Crys-tallographica, 825, | 525-1 533.

Bence, A.E., and Albee, A.L. (1968) Empirical correction factorsfor the electron microanalysis of silicates and oxides. Journalof Geology, 7 6, 382-403.

Bloss, F.D. (1971) Crystallography and crystal chemistry. Holt,Rhinehart and Winston, New York.

Bodkin, J.B. (1977) Determination of fluorine by use of an ion-selective electrode following fusion with lithium metaborate.The Analyst, 102, 409-413.

Bradley, J.E.S., and Bradley, O. (1953) Observations on the col-oring of pink and green zoned tourmaline. Mineralogical Mag-azine,30,26-38.

Buerger, M.J., Burnham, C.W., and Peacor, D.R. (1962) Assess-ment of several structures proposed for tourmaline. Acta Crys-tallographica, I 5, 583-590.

Burnham, C.W. (1979) Magmas and hydrothermal fluids. In H.L.Barnes, Ed. Geochemistry of hydrothermal ore deposits, 7l-136. Wiley, New York.

Carobbi, G., and Pieruccini, R. (1947) Spectrographic analysisoftourmalines from the Island ofElba with correlation ofcolor

- and composition. American Mineralogist, 32, l2l-I30.Cemf, Petr. (1975) Granitic pegmatites and their minerals: Se-

lected examples of recent progress. Fortschritte der Mineral-o$e,52,225-250.

- (1982) Petrogenesis of granitic pegmatites. In petr Grn!,!,{, Shgrt goyLrse in granitic pegmatites in science and industry,405-461. Mineralogical Association of Canada Short CourseHandbook 8.

Cernf, Petr, and Burt, D.M. (1984) Paragenesis, crystallochem-ical characteristics, and geochemical evolution ofmicas in peg-

matites. In S.W. Bailey, Ed. Micas,259-297. MineralogicalSociety of America Reviews in Mineralogy, 13.

Chaudhry, M.N., and Howie, R.A. (1976) Lithium tourmalinefrom the Meldon aplite, Devonshire, England. MineralogicalMagazine, 40,747-751.

Chorlton, L.B., and Martin, R.F. (1978) The effect of boron onthe granite solidus. Canadian Mineralogist, 16,239-244.

Deer, W.A., Howie, R.A., and Zussman, J. (1962) Rock formingminerals. Volume l, Ortho- and ring silicates, fourth edition,300-319. Wiley, New York.

Donnay, G. (1968) Crystalline heterogeneity: Evidence from theelectron-probe study of Brazilian tourmaline. Carnegie Insti-tution of Washington Year Book 67, 219-220.

Donnay, G., and Barton, R., Jr. (1972) Refinement of the crystalstructure of elbaite and the mechanism of tourmaline solidsolution. Tschermaks Mineralogische und PetrographischeMitteilungen, 18, 273-286.

Donnay, G., Ingamells, C.O., and Mason, B. (1966) Buergerite,a new species of tourmaline. American Mineralogist, 51, 198-199.

Dowty, E. (1980) Crystal growth and nucleation theory and thenumerical simulation of igneous crystallization. In R.B. Har-graves, Ed. Physics ofmagmatic processes, 4 I 9-485. PrincetonUniversity Press, Princeton, New Jersey.

Dunn, P.J. (1977) Chromium in dravite. Mineralogical Maga-zine. 41.408-410.

Dunn, P.J., Appleman, D.8., and Nelen, J.E. (1977a) Liddicoa-tite, a ne\il calcium end-member of the tourmaline group.American Mineralogist, 62, I l2l-l 124.

Dunn, P.J., Appleman, D.E., Nelen, J.A., and Norberg, J. (1977b)Uvite, a new (old) common member of the tourmaline groupand its implications to collectors. Mineralogical Record, 8,100-108.

Ekambaram, V., Rosenberg, P.E., and Foit, F.F., Jr. (1981) Syn-thesis and characterization of tourmaline in the system NarO-AIrO3-SiOr-B2O3-HrO. EOS American Geophysical UnionTransactions, 62, 1065.

El-Hinnawi, 8.E., and Hofmann, R. (1966) Optische und chem-ische Untersuchungen an neun Turmalinen @lbaiten). NeuesJahrbuch fiir Mineralogie, Monatshefte, 8G89.

Epprecht, W. (1953) Die Gitterkonstanten der Turmaline.Schweizerische Mineralogische und Petrographische Mittei-lungen, 33,481-505.

Faye, G.H., Manning, P.G., Gosselin, J.R., and Tremblay, R.J.(1974) The optical absorption sp@ctra of tourmaline: Impor-tance of charge-transfer processes. Canadian Mineralogist, I 2,370-380.

Foit, F.F., Jr., and Rosenberg, P.E. (1977) Coupled substitutionsin the tourmaline group. Contributions to Mineralogy and Pe-trology, 62,109-127.

- (1979) The structure ofvanadium-bearingtourmaline andits implications regarding tourmaline solid solutions. Ameri-can Mineralogist, 64, 788-798.

Foord, E.E. (1976) Mineralogy and petrogenesis oflayered peg-matite-aplite dikes in the Mesa Grande District, San DiegoCounty, California. Ph.D. thesis, Stanford University, Stan-ford, California.

- (1977) The Himalaya dike system, Mesa Grande district,San Diego County, California. Mineralogical Record, 8, 461-474.

Fortier, S., and Donnay, G. (1975) Schorl refinement showingconpositional dependence of the tourmaline structure. Ca-nadian Mineralogist, 13, 173-177 .

Frondel, C., Biedl, A., and Ito, J. (1966) New type of ferric irontourmaline. American Mineralogist, 5 l, I 50 l-l 505.

Fuge, R. (1977) On the behavior offluorine and chlorine duringmagmatic differentiation. Contributions to Mineralogy and Pe-trology, 6l, 245-249.

Hanley, J. B. (1953) Bob Ingersoll mine (Keystone district). InL.R. Page and others, Pegrnatite investigations, 1942-1945,


Black Hills, South Dakota. U.S. Geological Survey Profes-sional Paper 247, 7 5-7 8.

Hards, N. (1976) Distribution ofelementsbetweenthe fluid phaseand silicate melt phase of granites and nepheline syenites. InProgress in experimental petrology. Natural Environment Re-search Council Report D6, 88-90.

Henry, D.J., and Guidotti, C.V. (1985) Tourmaline as a petro-genetic indicator mineral: An example from the staurolite-grade metapelites of NW Maine. American Mineralogist, 70,l -15 .

Hermon, E., Simkin, D.J., Donnay, G., and Muir, W.B. (1972)The distribution of Fe2* and Fe3+ in iron-bearing tourmalines:A Mtissbauer study. Tschermaks Mineralogische und Petro-graphische Mitteilungen, 19, 124-132.

Holland, H.D. (1972) Granites, solutions, and base metal de-posits. Economic Geology, 67, 281-301.

Ito, T., and Sadanaga, R. (1951) A Fourier analysis ofthe struc-ture of tourmaline. Acta Crystallographica, 4, 385-390.

Jahns, R.H. (1953) The genesis of pegmatites I. Occurrence andorigin of giant crystals. American Mineralogist, 38, 563-598.

- (1955) The study of pegmatites. Economic Geology, Fif-tieth Anniversary Volume, Part II, 1023-1 I 30._ (1982) Internal evolution of pegmatite bodies. In PetrCernf, Ed. Granitic pegmatites in science and industry, 293-327. Mineralogical Association of Canada Short Course Hand-book 8.

Jahns, R.H., and Burnham, C.W. (1969) Experimental studiesof pegmatite genesis: I. A model for the derivation and crys-tallization of granitic pegmatites. Economic Geology, 64, 843-864.

Jenks, W.F. (1935) Pegmatites at Collins Hill, Portland, Con-necticut. American Journal ofScience, 5th series, 30,177-197 .

Jensen, M. (1984) Major, minor, and trace elements (REE's) inapatite as recorders ofpegmatite petrogensis. M.S. thesis, SouthDakota School of Mines and Technology, Rapid City.

Laul, J.C. (1979) Neutron activation analysis of geological ma-terials. Review article for Atomic Energy Review, IAEA, 17,No. 3, 603-697.

Leckebusch, R. (1978) Chemical composition and color of tour-malines from Darre Pech (Nuristan, Afghanistan). Neues Jahr-buch liir Mineralogie Abhandlungen, 133, 53-70.

l,ofgren, G. (1980) Experimental studies on the dynamic crys-tallization of silicate melts. In R.B. Hargraves, Ed. Physics ofmagmatic processes, 487-551. Princeton University Press,Princeton, New Jersey.

London, D. (1986) Magmatic-hydrothermal transition in theTanco rare-element pegmatite: Evidence from fluid inclusionsand phase-equilibrium experiments. American Mineralogist,7 1, 37 6-395.

London, D., and Burt, D.M. (1982a) Alteration of spodumene,montebrasite and lithiophilite in pegmatites of the White Pi-cacho District, Arizona. American Mineralogist, 67, 97 -l 13.

- (1982b) Lithium minerals in pegmatites. In Petr Cern!,Ed. Granitic pegmatites in science and industry, 99-133. Min-eralogical Association ofCanada Short Course Handbook 8.

Manning, D.A.C. (1981) The effect of fluorine on liquidus phaserelationships in the system Qz-Ab-Or with excess water at IKb. Contributions to Mineralogy and Petrology, 76,206-215.

- (1982) Chemical and morphological variation in tour-malines from the Hub Kapong batholith of peninsular Thai-land. Mineralogical Magazine, 45, 139-147.

Manning, D.A.C., Martin, J.S., Pichavant, M., and Henderson,C.M.B. (1984) The effect of F, B and Li on melt structures inthe granite system: different mechanisms? In Progress in ex-perimental petrology, 36-41. Natural Environment ResearchCouncil Report D25.

Moore, J.G., and Lockwood, J.P. (1973) Origin of comb layeringand orbicular structure, Sierra Nevada batholith, California.Geological Society of America Bulletin, 84, l-20.

Moore, P.B. (1982) Pegmatite minerals of P(V) and B(IID. In

Petr eern!, Ed. Granitic pegmatites in science and industry,267-291. Mineralogical Association of Canada Short CourseHandbook 8.

Neiva, A.M.R. (1974) Geochemistry of tourmaline (schorlite)from granites, aplites and pegmatites from Northern Portugal.Geochimica et Cosmochimica Acta, 38, 1307-1317.

Nemec, D. (1969) Fluorine in tourmalines. Contributions to Min-eralogy and Petrology, 20, 235-243.

Norton, J.J. (1983) Sequence of mineral assemblages in differ-entiated granitic pegmatites. Economic Geology, 7 8, 854-87 4.

Norton, J.J., Page, L.R., and Brobst, D.A. (1962) Geology of theHugo pegmatite, Keystone, South Dakota. U.S. GeologicalSurvey Professional Paper 297 -8, 49-127.

Norton, J.J., and others (1964) Geology and mineral deposits ofsome pegmatites in the southern Black Hills, South Dakota.U.S. Geological Survey Professional Paper 297 -8, 293-3 41.

Nuber, B., and Schmetzer, K. (1979) The lattice position of Cr3*in tourmaline: Structural refinement of a chromium-rich Mg-Al-tourmaline. Neues Jahrbuch fiiLr Mineralogie Abhandlun-gen, 184-197.

- (1981) Strukturverfeinerung von Liddicoatit. Neues Jahr-buch fiir Mineralogie, Monatshefte, 215-219.

Page, L.R., and others (1953) Pegmatite investigations 1942-1945, Black Hills, South Dakota. U.S. Geological SurveyProfessional Paper 247 .

Papike, J.J., Jensen, M., Shearer, C.K., Simon, S'B., Jolliff, B.L.,Walker, R.J., and I,aul, J.C. (1984) Apatite as a recorder ofpegmatite petrogenesis. Geological Society of America Ab-stracts with Programs, 16,617.

Pichavant, M. (1981) An experimental study ofthe effect ofboronon a water saturated haplogranite at I Kbar vapour pressure.Contributions to Mineralogy and Petrology, 76,430-439.

Power, G.M. (1968) Chemical variation in tourmalines fromsouthwest England. Mineralogical Magazine, 36, I 07 8- I 09 8.

Redden, J.A. (1963) Geology and pegmatites of the Fourmilequadrangle, Black Hills, South Dakota. U.S. Geological SurveyProfessional Paper 297-D, 199:29L

Redden, J.A., Norton, J.J., and Mclaughlin, R.J. (1982) Geologyof the Harney Peak Granite. U.S. Geological Survey Open-File Report 82-481.

Roberts, W.L., and Rapp, G., Jr. (1965) Mineralogy of the BlackHills, Bulletin Number 18, South Dakota School of Mines andTechnology, Rapid City, South Dakota.

Roegge, J.S., Logsdon, M.J., Young, H.S., Barr, H.B', Borcsik,M., and Holland, H.D. (1974) Halogens in apatites from theProvidencia Area, Mexico. Economic Geology, 69,229-240.

Rose, D., Langer, K., and Tennyson, CH. (1981) Turmalin Farbeund Paragenese. Fortschritte der Mineralogie, 59, 164-1 66.

Rosenberg, P.E., and Foit, F.F., Jr. (1979) Synthesis and char-acterization of alkali free tourmaline. American Mineralogist,6 4 , 1 8 0 - 1 8 6 .

Rumantseva, E.V. (1983) Chromdravite, a new mineral. ZapiskiVsesoyuznogo Mineralogicheskogo Obshchestva, | 12, 222-226.(English summary in American Mineralogist, 69,210.)

Sahama, Th.G., v. Knorring, O., and Tdrnroos, R' (1979) Ontourmaline. Lithos, 12, 109-1 14.

Schmetzer, K., and Bank, H. (1984) Crystalchemistryoftsilaisite(manganese tourmaline) from Zambia. Neues Jahrbuch fiir Mi-neralogie, Monatshefte, 6 1-69.

Schmetzer, K., Nuber, B., andAbraham, K. (1979) Crystal chem-istry of magnesium-rich tourmalines. Neues Jahrbuch fiir Mi-neralogie Abhandlungen, 136, 93-l 12.

Shearer, C.K., Papike, J.J., and Laul, J.C. (1985) Chemistry ofpotassium feldspars frorn three zoned pegmatites, Black Hills,South Dakota: Implications concerning pegmatite evolution.Geochimica et Cosmochimica Acta, 49,663-473.

Shearer, C.K., Papike, J.J., Simon, S.B., and I-aul, J.C. (1986)Pegmatite/wallrock interactions, Black Hills, South Dakota:Interaction between pegmatite-derived fluids and quartz-micaschist wallrock. American Mineralogist, 71, 518-539.

500

Sheridan, D.M., Stephens, H.G., Staatz, M.H., and Norton, J.J.(1957) Geology and beryl deposits of the Peerless pegmatite,Pennington County, South Dakota. U.S. Geological SurveyProfessional Paper 297-A, 147.

Slivko, M.M. (1965) On inclusions of solutions in crystals oftourmaline. In E. Roedder, Ed. Research on the nature of min-eral-forming solutions. International Series of Monographs inEarth Science, 22, Partlll, 626-634.

Snetsinger, K.G. (1966) Barium-vanadium muscovite and va-nadium tourmaline from Mariposa County, California. Amer-ican Mineralogist, 51, 1623-1639.

Staatz, M.H., Murata, K.J., and Glass, J. J. (1955) Variation ofcomposition and physical properties of tourmaline with itsposition in the pegmatite. American Mineralogist, 40, 789-804.

Stewart, D.B. (1978) Petrogenesis of lithium-rich pegmatites.American Mineralogist, 63, 970-980.

Tauson, L.V. (1965) Factors in the distribution of the trace ele-ments during the crystallization of magmas. Physics andChemistry of the Earth, 6,215-249.

Taylor, A.M., and Terrell, B.C. (1967) Synthetic tourmalinescontaining elements of the first transition series. Journal ofCrystal Growth, l, 238-244.


Tuttle, O.F., and Bowen, N.L. (1958) Origin of granite in thelight of experimental studies in the system NaAlSi3O8-KAl-SirOs-SiOr-HrO. Geological Society of America Memoir 74.

Tuzinski, P.A. (1983) Rare-alkali ion halos surrounding the BobIngersoll lithium-bearing pegmatite mine, Keystone, Black Hills,South Dakota. M.S. thesis, Kent State University, Kent, Ohio.

Walenta, K., and Dunn, P.J. (1979) Ferridravite, a new mineralof the tourmaline group from Bolivia. American Mineralogist,64,945-949.

Werding, G., and Schreyer, W. (1984) Alkali-free tourmaline inthe system MgO-AlrOr-BrO3-SiO2-HrO. Geochimica et Cos-mochimica Acta, 48, 133l-1344.

Wilkins, R.W.T., Farrell, E.F., and Naiman, C.S. (1969) Thecrystal field spectra and dichroism of tourmaline. Journal ofPhysics and Chemistry ofSolids, 30,43-56.

Zoltai,'|., and Stout, J.H. (1984) Mineralogy, concepts and prin-ciples, I 35-1 83. Burgess Publishing Company, Minneapolis.

MlNuscnrpr REcETvED Fsspueny 28, 1985MnNuscnrpr AccEprED NovrMsen I I, 1985

Tourmaline as a recorder of pegmatite evolution: Bob ... · during the formation of the pegmatite...

Documents

Transcript of Tourmaline as a recorder of pegmatite evolution: Bob ... · during the formation of the pegmatite...