1093

Canadian MineralogistVol. 30, pp. lW3-1112 (1992)

AMPHIBOLE-RICH POLYCRYSTALLINE CLOTSIN CALC-ALKALINE GRANITIC ROCKS

AND THEIR ENCLAVES

ANTONIOCASTRODepartamento de Geologia y Mineria, Universidad de Sevilla, E-21819 La Rdbida, Huelva, Spain

W. EDRYD STEP}MNSDepartment of Geography and Geology, Universiry of St. Andrews, St. Andrews, Fife KY16 9ST, Scotland

ABS'IRAC|

The presence of amphibole-rich polycrystalline aggregates (clots) of millimetric size is a widespread feature of calc-alkalinegranitic rocks and related microgranularenclaves. A petrological study ofthese clos, including field relationships, textures, mineralchemistry, and compositional zoning has been carried out to constrain possible origins and petrogenetic significance. The twoplutons chosen, the Quintana pluton in the Hercynian belt of Iberia and the Strontian pluton in the Caledonides of Scotland, differin age and tectonic setting. In both, amphibole-rich clots are present in microgranular enclaves and in the host granitic rocks. Theclots are generally zoned. The margin is rich in amphibole, but with significant proportions ofbiotite and plagioclase, normallyintergrown with other minerals of the groundmass. The interior is dominantly composed of multiple crystals of amphibole inmutual contact showing a typical granoblastic texture. Individual crystals there show compositional zoning. In general terms, thecore is actinolite, and the rim, magnesian homblende. This "reverse" zoning can be explained by a complex reaction with earlypyroxene. This reaction probably took place in two stages, one producing magnesian homblende in equilibrium with the melt, andthe otherproducing actinolite in the solid state, from reaction ofpyroxene armored by eady homblende. The origin ofthe pyroxeneprecurson to these amphibole clots is not well constrained by this study, but the prismatic form of these clots in the enclaves doessuggest that those preserved in enclaves were originally idiomorphic magmatic phenocrysts, whereas those in the host granitecould also have a restitic origin.

Keywords: granite, clot, amphibole, calc-alkaline, magma mixing, restite, Quintana pluton, Spain, Strontian pluton, Scotland.

SolaNaeRs

I-a pr6sence d'agrdgaa millim6triques polycristallins riches en amphibole est r6pandue dans les roches granitiquescalco-alcalines et les enclaves microgrenues qu'elles contiennent. Nous avons 6tudi6 ces agrdgats, leurs relations de terrain, leuntexfures, et la composition et la zonation de leurs min6raux, afin de placer des limites sur leur origine et leur significationp'6trog6n6tique. ks deux suites plutoniques choisies, Quintana (ceinture hercynienne ib6rique) et Strontian (Cal6donides d'Ecosse)difFdrent dans leur 6ge et leur contexte tectonique. Aux deux endroits, les agr6gats riches en amphibole sont prdsents dans lesenclaves microgrenues et dans les roches granitiques h6tes. Ces agrdgats sont zon6s, en gdndral. l,a bordure est enrichie enamphibole, mais contient une proportion importante de biotite et de plagioclase, normalement en intercroissance avec les autresmin€raux de la pdte. L'int6rieur est surtout compos6 d'une multitude de cristaux d'amphibole dont les contacts d6finissent unetexture granoblastique. t es cristaux individuels sont zon6s. En g6n6ral, le coeur est fait d'actinote, et la bordure, de homblendemagndsienne. Cette zonation "invers6e" peut s'expliquer au moyen d'une rdaction complexe avec un pyroxbne prdcoce. Lar6actionaurait eu lieu en deux stades, le premier donnant une hornblende magndsienne en 6quilibre avec le magma, I'autre produisantI'actinote i1'6tat solide, parr6action dupyroxbne recouvert de hornblende pr6coce. L'origine du pyroxbne pr6curseur d ces agrdgatsriches en amphibole n'est pas bien d6finie par nos r6sultats, mais la forme prismatique de ces agr6gats dans les enclaves fait penserqu'il s'agissait au d6part, dans ces cas du moins, d'un ph6nocristal magmatique idiomorphe. Ceux du granite h6te pourraient aussi6tre un mat6riau "restitique"'

oraduit par ra R6daction)

Mots-clis: granite, agr6gats, amphibole, calco-alcalin, m6lange de magmas, restite, pluton de Quintana, Espagne, pluton deStrondan. Ecosse.

r094 THE CANADIAN MINERALOGIST

IN'rRoDUcroN

MaJic clots (millimetric aggtegates of mafic miner-als) are remarkably common features of calc-alkalinetonalites and granodiorites of widely varying age andtectonic setting (e.9., hesnall & Bateman 1973, Vernon1983, Cantagrel et al. 1984, Znrpi et a/. 1989" Castro1990, Castro er al. l99la).InI-type granitic rocks, theseclots are most commonly amphibole-rich. Their originhas been variously attributed to restite, accidental xeno-liths, cognate xenoliths, remnants of mingled magmas,mineral cumulates, polycrystalline pseudomorphs ofphenocrysts, and crystal agglomerations, with effects ofrecrystallization where appropriate. A knowledge of theorigin of such clots is most desirable, as their presenceis generally taken as a sure indication of one or other ofthese petrogenetic processes, and their normally homo-geneous disribution over a pluton or plutonic unit istaken as evidence of the widespread extent of thatprocess. It is very difficult, howeveg to design field orpetrographic tests to choose among these models; in-deed, if recrystallization of clots has occuned, theultimate origin may be more obscure. In volcanic rocks,clots of similar form but rather different compositionalso may occur, and have even been inferred to controlthe composition of calc-alkaline series (Scarfe & Fujii1987). Interpretation oftextures in plutonic rocks is lesswell constrained owing to the possible modifications byrecrystallization during slow cooling. Thus, for thisstudy we have adopted an approach that places mostemphasis on mineral compositions and their patterns ofzonation.

In comparing within-clot and between-clot composi-tional and textural variations. we find the combinationofback-scattered electron imagery (BEI) and electron-microprobe analysis to be particularly useful. Whereassome important tgxtural features are visible using nor-mal polarized-light petrographic techniques, others thatare controlled solely by bulk composition are onlyrevealed by BEI (a Z-contrast technique: Lloyd 1987).Thus BEI can guide the points for analysis particularlyefficiently and disclose textures and compositionalcharacteristics that would otherwise only become appar-ent only with a high density of spot elecFon-microprobeanalyses.

MODEIS FORTHE ORIGINoF CLoTs

, The literature on amphibole-rich clots in graniticrocks is quite extensive. Here, we attempt to summarizethe models for their origin in terms of end-memberprocesses, combinations of which may be possible.I . Restite: a patch of mafic phases survives the meltingprocess and is transported along with the magma, eitherunmodified or afterrecrystallization in equilibrium withthe host magma. hesnall & Bateman (1973) interpretedclots in the Sierra Nevada batholith to be indicators of

the presence of solid material (restite fraction) canied upfrom depth by the granitic magma (Chappell et al. 1987).2. Crystal cumulate: an accumulation of melt-precipi-tated mafic phases becomes entrained and dispersedthrough the magma body.3. Mafic remnants of mingled magmas: rernnants fromthe quenching and dynamic disruption of mafic magmain a lower-temperafure felsic magma during magmaticflow.4. Disaggregated enclaves: a cluster ofattached maficminerals forms by the chemical or mechanical disaggre-gation of a mafic enclave (either cognate or accidentalxenolith).5. Pseudomorphs: phenocrysts recrystallize into acluster of secondary minerals (Vernon 1983, Mazzoneet al. 1987).

In several ofthe above cases, it may be appropriate toconsider that there has been a precursor phase, such aspyroxene, that recrystallizes to an amphibole-dominatedassemblage. This precursor is required for the formationof clots by the replacement of phenocrysts, but couldalso occur in the other models where the early-formedmafic phase was pyroxene that later reacted with melt,or fluid, to form a new assemblage dominated byamphibole.

Gsor-ocrcAr- SelI'INc

The two plutons chosen for this study are t}teStrontian pluton in the Scottish Caledonides and theQuintana pluton of the Spanish Hercynian (Fig. I ). Theircommon features include: (a) mineralogical and bulkchemical compositions (metaluminous) fypical of l-typeplutons, (b) the dominance ofgranodiorite, (c) abundantmicrogranular enclaves of diorite and tonalite composi-tion, and (d) the presence of abundant amphibole-richclots.

The Caledonian pluton of Strontian

The Strontian pluton (Fig. 1B) intruded Moine ho-terozoic metasediment ry and meta-igneous rock at 425Ma (Rogers & Dunning 1991) during the Caledonianorogeny in northern Britain. The metasediments wereregionally metamorphosed to the sillimanite zone of akyanite-sillimanite sequence flilinchester 1974), andthese were overprintedby an asymmetrical metamorphicaureole that reaches cordierite - K-feldspar assemblagesat the contact. The asymmetry is attributed to a gradientin the regional metamorphic isograds at the time ofemplacement (Ashworth & Tyler 1973). Emplacementof the pluton was passive, in part controlled by contem-poraneous movements on the Great Glen Fault (Hutton1988a b).

The pluton is internally zoned (Fig. lB), with an outernonporphyritic granodiorite (ONPG), inside which is theouter porphyritic granodiorite (OPG), and an inner coreof granite (IG). These plutonic members are easily

AMPHIBOLE.RICH CLOTS IN CALC.ALKALINE GRANMC ROCKS 1095

Frc. lA, B. Geological sketch and regional sening of the two granitic plutons chosen for study. A. The Quintana pluton (Spain).

B. The Strontian pluton (Scotland).

mapped in the field. The names used here are those ofthe IUGS classification (Streckeisen 1976) applied to the"tonalite", "porphyritic granodiorite", and "biotite gran-ite", respectively, of the earlier investigators (MacGre-gor & Kennedy 1932, Sabine 1963, Munro 1973).

Detailed petrographic descriptions may be found inthese papers. The boundary between the ONPG and theOPG is a transition over about l0 metres, whereas theOPG-IG boundary is quite sharp. The outer members ofthe pluton are host to some large bodies of "appinite",

S n6r domdnsE ouar domains

Los Pedroches batholith

tana Pluton $

ffi Cordierirc monzogranitcs-@f Granodiorircs ,/ =

tl-

/CIkml-l

1096 THE CANADIAN MINERALOGIST

hornblendite - diorite series rocks that show lobate andcrenulate liquid-liquid contacts indicating contempora-neous intrusion. Enclaves of microdiorite aspect areabundant in the ONPG and OPG, locally occurring assw.ums of apparently disaggregated synplutonic dykes.The ONPG and OPG are characterized by abundantamphibole-rich clots aligned within the plane of folia-tion: the idiomorphic crystals of amphibole grew ran-domly. The Sr isotope initial ratios for the pluton areabout 0.7053-O.7060 for the ONPG and OPG. 0.7068-0.7072 for the IG, 0.7051-{.7066 for the enclaves. and0.7056-7060 for the appinites (Halliday et al. 1979,Hamilton et al.'|,980" Holden et al. 1987\.

The Hercynian pluton of Quintana

The Quintana pluton (Fig. lA) is part of the LosPedroches batholith, one ofthe largest granitic batholithsof the Iberian massif. It was emplaced as an elongatebody parallel to the NW-SE trend of regional structures,at the southem branch of the Central Iberian Zone(Julivert et al. 1974). The pluton was emplaced at highlevels into Upper Paleozoic metasediments, and em-placement took place after the first phase of Hercyniandeformation, and probably during the second phase,which was dominated in this sector of the chain byintracontinental shear movements (Castro 1985, 1986).Full descriptions and a petrogenetic interpretation wererecently presented by Castro (1990).

The pluton is a medium-grained biotite hornblendegranodiorite. Several facies are identified on the basis oftheir amphibole content (normally 0-57o) and colorindex. However, these facies do not show any regulardistribution within the pluton, and rhe pluton is consid-ered to be essentially unzoned on a broad scale. Maficmicrogranular enclaves are abundant (l-57a) and more-or-less evenly distributed over the pluton. Three typesof enclave were classified by Castro (1990) as: (a)porphyritic tonalite enclaves (PTE), (b) hornblende-pla-gioclase diorite enclaves (HPDE), and (c) transitionalenclaves (TRE) that have features of both PTE andHPDE. Petrographic descriptions of these types are alsogiven in Castro (1990).

Clot-bearing rocks in the Strontian andQuintanaplutons

In both plutons, amphibole-rich clots tend to bepreferentially concentrated in the more mafic graniticfacies and also in certain types ofmicrogranular enclave.

Granitic rocl<s. At Sfontian, amphibole-bearing clotsare abundant in the ONPG. They are also present,however, with somewhat lesser abundance. in both theOPG and ONPG facies of the pluton. It is important tonote that the ONPG, which is richest in clots, has fewcrystals of idiomorphic amphibole, whereas in theporphyritic granodiorites, the modal proportion ofidio-morphic amphibole to clots is approximately 9: l. On the

basis of whole-rock geochemical studies (including clotsbut not enclaves; Stephens e/a/., in prep.), the differencebetween the two facies is more textural than composi-tional. The boundary between the ONPG and the OPGis transitional along the Loch Sunart section of thenortherly part of the pluton (Fig. 1B). There is an inversecorrelation between the abundances of idiomolphicamphibole and mafic enclaves, except where there areswarms of enclaves.

At Quintana, there is no preferential disribution ofthe clot-rich facies, which appear in domains richerin amphibole. Most of the amphibole occurs in clots,with only a minor proportion of isolated idiomorphiccrystals. In general, the Quintana granodiorite is morebiotite-rich in comparison with the Strontian rocks,in which amphibole dominates over biotite. In bothplutons, there is an appreciable correlation betweenthe abundances of enclaves and clots. That the enclave-rich facies are also richer in clots perhaps indicates acommon genetic relationship between enclaves andclots. The same correlation has been observed in othercalc-alkaline granitic bodies such as the Central Systembatholith of Spain (Castro et al. 1991b), and the SierraNevada batholith of California (hesnall & Bateman1973't.

Enclaves. In general, and certainly in the two plutonsconsidered here, not all mafic, microgranular enclavescontain amphibole-rich clots. Clots are particularlyassociated with the so-called transitional enclaves (TRE)of the Quintana pluton (Castro 1990), where clotsrepresent more than lOVo by volume of the enclave. Asimilar type of enclave is also common in Strontian andother Caledonian plutons (e.g., the Criffell pluton, asdescribed by Holden et al. 1987). Characteristics ofthese enclaves are: (1) a fine-grained matrix, (2) thepresence of plagioclase megacrysts having a resorbedcore, and (3) the presence ofamphibole in the form ofclots varying in abundance between 5Vo and 307o byvolume. The fine-grained mafix of these enclavestypically has a serial texture and comprises biotite,plagioclase laths, quartz, amphibole, acicular apatiteand other accessory minerals including titanite, zirconand opaque minerals. Poikilitic quartz is a feature ofmany tonalitic enclaves (Castro et al. 1,991c) andsynplutonic dikes in zones of magma interaction (Castroet al. l99la), suggesting a complex history of crystal-lization. The two other types of enclaves (porphyritictonalites, PTE and hornblende-plagioclase diorites,HPDE) are poorer in amphibole clots. Amphibole-richenclaves (HPDEtype), with modal amphibole approach-ing 907o, are present in both plutons, but are moreabundant in the Strontian pluton; clots are less abundantin this type. The PTE have few if any clots; theseenclaves are particularly abundant in the Quintanagranodiorite but are very scarce at Strontian, generallysupporting the correlation noted between amphibolecontent of the host granite and the abundance ofamphibole-rich enclaves.

Tnxrunal- FsafllnEs oF AMprilBoLE Clors

Amphibole-bearing clots, as described from batho-liths from a wide variety of orogenic settings, appear tobe remarkably similar in their textural and compositionalcharacteristics. We have studied these features in boththe Strontian and Quintana plutons.

Size

Amphibole clots vary from less than I mm up to a fewcm, but with an even grain-size for the individualcrystals of amphibole. However, the dominant size

1097

ranges from 2 to 8 mm (Figs. 2, 3); we note acorrespondencebetween size ofthe clots and the averagegrain-size ofthe rock. In flne-grained enclaves, clots aretypically smaller than those of the coarser-grained hostgranite, but in both cases, unusually large clos (> l0mm) may be present.

Shape

In undeformed granitic rocks, amphibole-rich clotstypically have a round shape, whereas in deformedrocks, including the marginal facies and enclaves oftheStrontian pluton (Hutton 1988b), amphibole-rich clots

AMPHIBOLE.zuCH CLOTS IN CALC.ALKALINE GRANITIC ROCKS

FIc. 2. Aspect of a clot-bearing tonalite of the Strontian pluton. Note the homogeneous sizeof the clots (darker spots).

1098 TTIE CANADIAN MINERALOGIST

Ftc. 3. Comparisons between shape and size of amphibole clots in different rock types of the Strontian and Quintana plutons. Aand C are the Quintana granodiorite and related tonalitic enclaves, respectively. B and D are the Strontian tonalite and relatedenclaves, respectively. Note the compositional zoning in the clots of B; the lighter core is richer in actinolite. Scale bar = 6mm.

have ellipsoidal shapes, with the principal axis alignedparallel to other deformation-related structures in therock. In the case of deformed rocks, the absence ofintracrystalline deformation in the latest mineral phases

(quafiz and K-feldspar) suggests that clots were de-formed when the rock was in a magmatic state. Theabsence of any kind of intracrystalline deformation inthe amphibole of these oriented clots may indicare that

some recrystallization may have occurred either during,or after, magmatic flow. The fact that the textures ofthese clots do not correspond to those generally attrib-uted to magmatic processes (as illustrated below) makesit difficult to interpret the history of clots. Specifically,it is not possible to demonstrate that they formed fromthe crystallization of a silicate melt during laminarmagmatic flow.

Occasionally, amphibole clots show approximatelyprismatic shapes; this feature is most commonly seen inclots within mafic enclaves, whereas the clots of hostgranites tend to be more inegular, and generallyrounded. The prismatic shape is conspicuous in the TREof the Quintana granodiorite (Fig. 3C) and in similarenclaves of the Strontian pluton. A comparison betweenclot shape in enclaves and in related granitic rocks isshown in Figure 3. This feature may be indicative ofpseudomorphism of early ferromagnesian silicates (py-roxene, olivine), as described by Vernon (1983) andTnrpi et al. (1989).

Texture

A major difficulty with the interpretation of rexturesof amphibole-rich clots is that they do not exhibit thetexture typical of serial crystallization in magmaticrocks. Most clots are characterized by a granoblastic-liketexture in which individual grains display near-120otriplejunctions (Fig.4). This texture is widespread in thecore of t}re clots, in which amphibole alone is present.Toward t}le rim, plagioclase or biotite or both may bepresent together with amphibole. In these marginaldomains, the texture is serial, and plagioclase and biotitefill interstices between idiomorphic crystals of amphibole (see below, Fig. 9A). This sequential, apparentlymagmatic, texture is best developed in the outermostparts of the clots, in which amphibole is intergrown withother minerals of the groundmass, namely biotite,plagioclase and quartz. Generally, this concentric zoningof texture is reflected in a weak optical zoning ofamphibole, with the marginal amphibole having deeperpleochroic colors than amphibole in the core.

Rehrtonships with other minerals

Biotite, plagioclase, quafrz, titanite and opaque min-erals are commonly present in modal quantities each lessthan 5Vo in amphibole-rich clots. Plagioclase typicallyshows an intergranular texture, indicating that it formedas a late-crystallizing phase. The composition of thisplagioclase is similar to that ofplagioclase in the hostgranite. These crystals are normally zoned, with themore albitic, extemal bands contouring amphibole crys-tals. However, lobate amphibole-plagioclase contactsalso are present, suggesting simultaneous growth. Simi-lar intergranular textures are shown by biotite crystalsclose to therim of the clots. However, biotite may exhibittriple junctions with amphibole, suggesting simultane-

1099

ous growth in the solid state. Most biotite occurs nearthe rim, in part encircling the clot and intergrown withthe outermost amphibole crystals, and overlaps in thecrystallization sequence with the groundmass plagro-clase and quafiz. Quartz and opaque minerals are foundas inclusions in amphibole, withquartzoccurring as verysmall vermicular-like grains typical of symplectiticintergrowths. Opaque minerals are principally sulfides(pyrite and chalcopyrite), except in a few cases whereprobable chrome-rich spinel was qualitatively identifiedwith the electron microprobe. Apatite and intergranulartitanite are common phases in amphibole-rich clots.Apatite tlpically has an acicular habit and appears asinclusions in amphibole within the outer parts of clots.

TABLE 1. REPFESENTATIVE FLECTRON-MICROPROBE DATAON AMPHIBOLE FROM THE QUINTANA GRANODIORTTE

AND RELATED ENCLAVES

Rml( typ€ PT snct. PT eml. PT eml. TR €nd. TR mcl. GsEd. G6odsample G2 c2 O2 c3 B3 D-l D-lAnalysb 10 12 37 2, 4t g.ga 19Desipdon Oor clot oot Llatix clot Oot Matix

AMPHIBOLE-RICH CLOTS IN CALC.ALKALINE GRANMC ROCKS

Sio2wto/o 50.66 51.16 50.94Ttrc2 0.5r 0.17 0.49At203 3.55 S.49 3.57CY2O3 0.04 0.04 0.00FeO 10.34 7.57 8.51Fo2@' 2.79 5.87 5.73F€ot 12.85 12.€5 19.66lt.lr|o 0.47 0.63 0.55MO 0.03 0.04 0.05t go 14.41 14.61 14.66CaO 11.68 10.86 11.48lla2o 0.7S 0.64 0.65Kn 0.3 0.9 0.28

TOTAL 95.52 95.66 96.98

Str|rtel iffmulae (O-3)q

Ar 0v)x(r)Ar (vr)riCr0+FB S+'Fo2+MnMMSx(c)

Fl2+

Bal,la (M4)>(B)

t'la (A)Kr(A)

7.41 7.495 7.4120.509 0.505 0.5s8

8 8 8

0.113 0.@ 0.0250.057 0.051 0.05s0.@4 0.@4 0.010.311 0.651 0.6s1.41 0.9tts 1.0410.06s 0.071 0.0670.003 0.@4 0.@53.17t 9.191 3.17S5.003 5.01 5.(X}9

0.003 0.01 o.@9't.851 1.705 1.79

0.147 0.285 0.2t 1.so i.99

0.0620.056 0.055 0.0590.118 0.055 0.053

50.61 48.68 47.13 49.910.62 O.7e 0.95 0./l54.17 5.34 5.92 4.120.2 0.14 0.611 11.75 12 10.7

2.35 29 2.24 e6919.12 14.36 14.01 13.12o.17 0.48 0.48 0.42

0.0214.01 1S.A 12.U 13.@11.8 11 .89 11 .1 11 .340.62 0.95 o.Ss 0.&0.4 0.52 0.6 0.31

95.24 96.@ 91r.35 94.42

7.443 7.N 7.216 7.850.55/ 0.795 0.7U 0.537

s 8 8 8

0.167 0.137 0.285 0.t910.069 0.088 0.11 0.050.023 0.015 0.670.26 0.s24 0.268 0.303

1.355 1.456 1.59S 1.U20.059 0.081 0.06t] 0.053

0.@23.O7 2,919 2.718 3.056

6.0q1 5.0G 5.@1 5.@2

0.@1 0.@ 0.001 0.@21.959 1.885 1.921 1.&.

0.14 0.112 0.178 0.1782 2 2 2

0.037 0.161 0.084 0.0590.078 0.097 0.118 0.0590.112 0.259 0.42 0.117

15.112 f5.259 15.n2 1F.11738.4t 21.* n.41

Toblcdm 15.118 14.952 15.0s5%An in Plsg. 26.65 30.06

') Caloiabd (s t€td)

I 100 THE CANADIAN MINERALOGIST

MINERAL Cgstvrrsrnv

Silicate phases in clots were analyzed by electronmicroprobe (JEOL JCXA733 Superprobe) at the Uni-versity of St. Andrews, with more than 500 pointanalyses performed. Analyses were carried out using abeam current of 20 nA and an acceleration potential ofl5 kV. Reference standards were a combination of puremetals and silicate minerals; the ZAF procedure wasused to conect apparent concentrations. Back-scatteredelectron imaging (BEI) was used to reveal features ofcompositional zoning in amphibole, and textural andreaction relationships between amphibole and contigu-ous phases.

Amp hibo le comp o s itions

Representative results of analyses are listed in Tables1. and2. Structural formulae were calculated on the basisof 23 oxygen atoms, with recalculation of ferric ironbased on the maximum per formula unit (pfu) to satisrystoichiometry (Robinson er a/. 1982), with the ideal totalof cations set at 13 (exclusive of Ca Na and K).Amphibole compositions, including those from clotsand idiomorphic crystals in both hostrocks and enclaves,display a wide range of Si pfu, from 6.8 to 7.8 as shownin the IMA classificauon (Fig. 5; trake 1978). [n mostcases, the amphibole is magnesian hornblende andactinolitic homblende; pargasitic hornblende only ap-

pears in the Strontian pluton in the core of idiomorphiccrystals.

Complex coupled substitutions between various cat-ions in different amphibole sites are common. Theiridentification is possible using simple schemes(Czamanske & Wones 1973, Blundy & Holland 1990)in which any amphibole composition is related to thetremolite (Tr) end-member. Charge compensation dueto the introduction of Al in tetrahedral sites may bebalanced in various ways, the most cornmon in naturaland synthetic amphiboles being incorporation of alkalisin theA site (Blundy & Holland 1990). This is the edenite(Ed)-type substitution, formulated as:

Si +no = lvAl + Naa

Another common way of achieving charge balance isby incorporation of octahedrally coordinated Al insubstitution of divalent cations in M1-M3 sites. This isthe tschermakite (Ts)-type substitution, formulated as:

Si + Rff,r,_r., - IvAl + vrAl

Figures 6 and 7 show cation plots used to identifythese coupled substitutions. Ed-type substitution isdominant in the amphiboles, as evidenced by the goodconelation between tetrahedrally coordinated Al andalkalis in the A site, for amphibole from both theStrontian and Quintana plutons. This plot also showscompositional differences between amphibole within

l 1 0 l

the host granodiorite and within enclaves of the Quin-tana pluton (Fig. 7B). However, analogous differencesare not found in the amphibole crystals from theStrontian pluton (Fig. 68). It must be noted that there isa smaller difference in bulk composition between en-claves and host in the Strontian pluton (both aretonalites) compared with the Quintana pluton.

The linear correlation between IvAl and other tetra-valent and trivalent cations (Figs. 6,4,, 7A) is better inamphibole from Strontian than in that from Quintana,but in both cases correlations are less significant than forthe Ed-type substitution. This indicates that other cou-pled substitutions involving hornblende (Hbl), pargasite(Prg) and tschermakite (Ts) are rather less importantthan the Ed-type substitution in determining amphibolecomposition in these two plutons. Combinations of thesesubstitutions are shown [Fig. 8, based on Blundy &Holland (1990)1. The dominant compositional vectorforboth plutons starts from the Tr end-member and movestoward a position intermediate between Ed and hg butrather closer to Ed. The dominant substitutions aresimilar for the amphibole cornpositions of clots fromenclaves and host granite in both plutons. Granitic rocksand related enclaves are similar in overall bulk compo-sition, but the PT conditions (mainly P) during crystal-lization were different. The Quintana pluton was em-placed at shallower levels than the Strontian pluton.

Total Al content is strongly dependent on T, asdemonstrated by Blundy & Holland (1990), but balanc-

tryryrffi{

Rc. 4. Textural relationships in several amphibole-rich clots from microgranular enclavesof the Quintana granodiorite. A. The minerals at the edge are intergrown with mineralsofthe fine-grained matrix. B. Polygonal packing, with abundanttriplejunctions betweendifferentamphibole crystals andbetween amphibole and biotite (arrow). C.Intergranulartexture is displayed by plagioclase (arrow) near the edge of the clot. Scale barcorresponds to 3 mm in A, and I mm in B and C.

tta2 THE CANADIAN MINERALOGIST

TABLE 2, REPRESENTATTVE ELECTRON.MICROPROBE DATA OIiI AMPHIBOLE FROMTHE STRONTIAN.GRANMC ROCKS AND RELATED ENCI.AVES

Rek type Tmallb Ends Etrlm Tmalhe Tonallb EndaF Erdaw &dm EndaE Emlaw llybtd Hybridsample sRz sR2 sR2 sR4 SR4 SRo gRB sR6 sR6 SRB sFo gRE

Anat6ls 8 15 & U 41 a(@n) 5(rtn) l9(@D)28(dm) 36 2S(coro)24(ttft)Doscdpdon ldbm. ldom. ml.Plag ldom. Fumclot Clot Clor Oor Cbt Rimclot Oot Clot

Sfo2wto, 8,m O.78 47.48 48.00 48.46 62.s1io2 0.30 |u 0.96 1.06 0.96 0.2,4f203 4.53 6.dr 8.05 6.68 6.16 L72G203 - 0.01 0.06 0.06 0.03F6O S.&f 9.At 10.11 10.70 9.97 6.88Fe2O3' 4.63 5.08 4.U 4.42 4.3s 4.61Mno O.a7 0.29 0.35 0./O' 0.31 0.35Nlo O.O2 0.04 0.03MSO 15.33 13.60 13.74 13.31 13.V 16.4CeO 11.96 l l.te 11.7s 11.71 11.69 11.73Na2O 0.91 '1.44 1.11 1.19 1.03 0.56l<n 0.36 0.78 0.62 0.68 0.57 0.4

TOTAL 57,62 57,31 98.3r! 98.22 97.34 96.92

Sn|shmlbmulae (O-23)

At0v) 0.681 o.s6 0.s59 1.@1 0.914 0.404tcD 8.0@ 8.o0 8.000 8.@0 8.0e 8.0o

Al (VD 0.090 0.147 0.OSg 0.145 0.1,4 0.05611 0.0311 0.114 0.107 0.116 0.105 0.024cr& - 0.ol 0.@8 0.007 o.@4Fegr' 0.505 0.662 0./162 0.486 0.484 0.5@Fe2+ 1.0115 1.174 1.257 13@ 1.8 O.8AMn 0.045 0.@6 0.0/t4 0.049 0.09 0.043i0 - - o.@3 0.@4 0.0@Mg 9.N 2.972 9.038 2.892 3.@2 3.541>(o) 6.006 6.@7 5.@7 5.010 6.008 5.@5

R2+ 0.006 0.O7 0.007 0.010 0.008 0.005Cs 1.849 1.753 1.872 1.82s 1.831 1.802l,,la(M4) 0.146 0.2/O 0.'121 0.161 0.161 0.166>(B) e@ a@ 2.000 2.000 2.@ 1.S

/a.49 55.d2 50.94 /15.34 55.30 51.84o.75 o.to 0.46 1.21 0.10 0.505.80 1.38 4.42 7.8 1.52 3.730.05 - 0.10 0.01 0.03 0.079.19 5./tt 8.01 1r.18 5.18 7.794.52 3.80 4.73 4.5't 9.03 3.e0.36 0.38 0.28 0.n 0.20 0.&0.m 0.14 0.07 0.05 0.@ 0.08

14.27 17.@ t5.50 1a&4 18.68 16.6411.67 11.85 11.76 11./A 12.4 12.24

't.21 0.35 0.82 1.51 0.32 0,740.58. 0.@ 0.31 0.80 0.@ 0.34

7.@90.9178.@

0.@80.0820.06o.afii1.1270.0440.o23.1085.@

7.M 7.U4 6.7n 7.7U 7.440.200 0.656 1.m. 0.216 0.s748.0@ 8.0@ 8.00 8.@ 8.@

0.030 0.@5 0.168 o.@ 0.660.010 0.050 0.139 0.010 0.054

- 0.011 0.@2 0.@4 0.o80./O7 0.515 0.5@ 0.321 0.s590.7/2 0.9@ 'l.rl02 0.728 0.S50.0i16 0.035 0.0i8 0.024 0.0390.015 o.@ 0.006 0.010 0.m9g.ru 3.3it1 2.750 3.913 3.5525.019 5.014 5.012 5.012 5.012

0.@9 0.019 0.014 0.012 0.012 0.012't.s28 1.800 1.817 1.8ri1 1.873 l.S7r0.155 0.@5 0.170 0.15'/ 0.@7 0.t@2(rl0 1.914 2.000 2.@ 1.972 /@

r\b(A) 0.1€ 0.1@ 0.108 0.174 0.132 - 0.176 - 0.061 0.282 - O.@7K 0.067 0.142 0.118 0.127 0.16 0.040 0.108 0.016 0.058 0.t6tr 0.015 0.002>(A) 0.171 0.311 0.316 0.301 0.23s 0.040 0.284 0.016 0.117 0.415 0.015 0.159

ToiElcadm: 15.182 15.318 15.323 15.311 15.2,18 15.@ 15.2S9 14.949 15.131 15.447 14.9S 15.171%&t in Pla( 16.17 21.50 20.80

') Calculalgd (s ren)

ing ofcharges by coupled substitutions depends more oncomponent activities than PT conditions. Ed-type sub-stitution is dominant in amphibole crystallized in equilibrium with plagioclase, the norm in calc-alkalinemagmas.

Compositional zoning in the clot amphiboles

A significant feature of the clots is the presence ofdiscontinuous, regular and sector zoning in the amphi-boles. There exists a compositional gap of abott 57oSiO2 and 2Vo MgO (see Table 2). Geometrically, thecompositional zoning appears in three different situ-ations: (1) core zoning, (2) oscillatory zoning, and (3)patchy zoning. In all three cases, the compositional

differences are very similar, and the boundary betweenadjacent zones is sharp.

Zoning in the core is displayed by the amphibolegrains displaying a granoblastic texture, in the centralpart of the clots (Figs. 9A, B). Cores have an actinoliticcomposition richer in Mg and Si and poorer in Al, Fe,and alkalis. than the rim" which is hornblende incomposition. The shape of actinolitic core is more or lessinegular, but generally follows the external shape ofthezoned crystals. Figure l0A is a compositional profile ofa zoned crystal in a clot within the Strontian NPG.Zoning, although revealed as being sharp by BEI @igs.9A, B), does not appear to be sharp in terms of theMg(Mg+Fe) ratio, indicating either that this parameterwas not important in the primary zonation or that

subsequent re-equilibration affected the distribution ofFe and Mg. The compositional vector on the substitutiontriangle @g. l0B) is coincident with the general vectordisplayed by the amphibole grains (Fig. 8), suggestingthat zoning is controlled by magmatic processes. Thezoned crystal of Figure 10 can be taken as representativeof all the zoned crystals of amphibole, as the samecompositional relationships were found in many ana-lyzed cases in both the Strontian and Quintana plutons.If zoning is related to magmatic crystallization, a

I 103

problem arises in explaining the apparent sequence ofactinolite to hornblende, the reverse of that normallyobserved, which suggests that the activities of silica andthe alkalis decreased as crystallization proceeded. Thegranoblastic-like texture also is not aprimary magmaticfeature. Many amphibole grains in clots were found tohave oscillatory zoning near the edge, and this feature ismore cornmon in idiomorphic crystals of amphiboletexturally intergrown with biotite (Fig. 9A). The zoningconsists of several alternatins continuous bands of

AMPHIBOLE-RICH CLOTS IN CALC-ALKALINE GRANMC ROCKS

1r0

+*olr+rDEcttE

O STRONTIANO QUINTANA

:f+o

Magnesio-hornblende

-a

Tremolite

ffiActinol.

a

, 1o1o?of o

P &oEg8 o

Actinolite

7rO0,5 L

6r5 7,5 Si (pfu) 8,0

Frc. 5. lrake's (1978) classification diagram showing the amphiboles ofthe Strontian and

Quintana plutons.

B)B)Nat

A ^ - o pF q i { , G ' T

.-.tr.,aibRE atlr^ f f i

O L0.0 o,2 0.4 6a+K) In A 0.6

Fro.7. Cationplots forthe amphibole compositions at Quintana.Compositions from clots and other textural situations areincluded in order to make comparisons. The grouping in thealkalis versas,IvAl plot @) is better than in the octahedralsites versus ''Al plot (A).

Lsgond:

A)lvAl

AA

^";HEFfiff'" "1r1fd. .

- E

1" rf'^":

A)2 -

O Lo.2

1.0 12vlAr * Fe&*2Tl

MA

+ A A r

' af,t.IEdssf a

..6 ;tr r

A ^^d

o0 02 0.4 (!,la+K) In A 0.6

Ftc. 6. Cation plots for the Strontian amphiboles. Compositionsfrom clots and other textural situations are included in orderto make comparisons. The sroupine in the alkalis yersrrr'Ivat ptot 1n; *Oetterttranln-ttre octa-[eAra sites yerszs ryAl

plot (A), indicating the dominance of the edenite+ypesubstitution.

o Mdrlx Granod.. Clot Granod.A MddxHPsncA ClolHPonc.tr MdilPTsnc.r Clol PT enc.

0.4 0.0 0.8 1.0 1.2vlAr * F"&*fi

lvlt

l l M TI{E CANADIAN MINERALOGIST

E A vlnr

Frc. 8. Triangular plot for the amphiboles of Quintana (A) andStrontian (B) showing the substitution vector, dominatedbv a combination of Ed and Ps.

actinolite and hornblende. Occasionally, actinolite alsoforms irregular lamellae in hornblende; this is thepatchy zoning referred to above and shown in Figure9C. The association of actinolite lamellae with quartzinclusions may be accidental in this case, and they areunrelated in most cases.

Amphib ole-biot it e re hrtons hip s

Biotite is the most abundant subordinate phase in theclots that we have studied. Indeed, some clots containonly biotite. These are very abundant in the biotitegranodiorites and related enclaves of the Central Sys-tembatholith of Spain (Castro et al. tgglb).Many clotsrich in amphibole have a marginal zone rich in biotite;biotite may show complex textural relationships withamphibole crystals inside the clots (Figs. 48, 9A).Biotite may show a granoblastic-like texture, with triplejunctions involving biotite and amphibole (Fig. 4B),which suggests that they recrystallized in equilibrium,with no apparent signs of reaction. Elsewhere, biotiteshows reaction textures with amphibole in both direc-tions, that is Bt reacts to Amp and Amp reacts to Bt, thelatter being very common near the margin of theamphibole clots. Finally, biotite appears to fill intersti-ces between the amphibole grains (Fig. 9A). Biotite inthese different textural forms was analyzed with theelectron microprobe, and representative compositionsare presented in Tables 3 and 4. The most significantfeature is the rather homogeneous composition of the

biotite (Fig. I l). Biotite from the Strontian pluton has alarger Mg/(Mg+Fe) ratio (Mg#) than biotite from theQuintana pluton. This difference may be related to theslightly more mafic composiuon of the Strontian bulkrocks. It is not possible to distinguish biotite from

lver

different textural occurrences in either the Strontian orQuintana plutons, which suggests that re-equilibrationhas extensively affected this mineral, or that this parame-ter is not sensitive to paragenetic environment. Biotite-amphibole pairs were analyz,ed from the margins ofclots, in order to compare the respective Mg# and toestimate the average partition coefncient, expressed as(xMcrF")Bd/(xMc/XF")a-p (after Speer I 987).

I 105

Figure 12 is an AFM plot for these biotite-amphibolepairs. In the Strontian pluton, there are no significantdifferences between the slopes oftie lines correspondingto pairs from the margins of clots compared with pairsfrom idiomorphic amphibole in contact relationship withbiotite. This strongly suggests that at least the marginalpart of the clot crystallized in equilibrium with amphi-bole ofthe host rock. The calculated D is approximately

Frc. 9. Back-scattered electron images (BEI) of representative amphibole clots from therocks studied. Three types ofzoning are displayed: core zoning (A,B), oscillatory zoning(A) and patching zoning (C). The darker zones are richer in actinolite. Note the triplejunctions and the intergranular biotite in A. The zoned crystal at the center of A alsoshows oscillatory zoning near the rim. A compositional traverse across the labeled crystalin B is shown in Fig. 10. A comes from a clot in an enclave of the Quintana pluton. Band C come from clots in the Strontian tonalite. Scale bar = 100 ttm.

8.0f

A)f7.6f

F7.2L

to'r F0.4 F

l0.2 f.

t0.06

|.

I'*fo.o0 Lnrf

In'fr

o.7L

B)

- 1 0 0 l r m +

1 106 TI{E CANADIAN MINERALOGIST

Amphibole-plagioclase relations hip s

Plagioclase and amphibole are nonnally the mostabundant felsic and mafic phases, respectively, in inter-mediate calc-alkaline plutonic rocks. Both may crystal-lize over a wide range of T (see Blundy & Holland 1990)and may record physicochemical changes duringmagma evolution. Many of the amphibole clots studiedcontain plagioclase as a subordinate phase. Mutualrelationships between amphibole and plagioclase areimportant for recognition of whether clots have amagmatic origin and whether the amphibole-plagio-clase thermometer (Blundy & Holland 1990) is applica-ble. Plagioclase normally occurs in these clots as aninterstitial phase showing lobate contacts with amphi-bole. These features are indicative of simultaneouscrystallization, and thus geothermometric applicationsare appropriate. Figure 13 shows compositions of se-lected pairs (Tables 1 and 2) plotted in a graphicalsolution of the Blundy & Holland (1990) thermometer.The selected pressures, I kbar for the Quintana plutonand 3 kbar for the Strontian pluton, are based on

TABLE 3. REPRESENTAT]VE ELECTRON-MICROPROBE DATAON BIOTITE FROM THE QUINTANA GRANODORrrE

AND REI.ATED ENCLAVES

Rocktypo PTsncl. PTond. Pfsd. TRmd. Omod Gnnod.Senple C2 C-2 G2 B-2 D.l D.1A n a t 8 i 6 3 0 3 2 3 4 3 2 3Dsshton Amph. dot Amph. dot Anph. olol lilatix Marn Amph. olot

FIc. 10. A. Cation plots showing variations along a travene inan amphibole crystal from the clot of Fig. 9B. Note thediscontinuous zoning involving Si,IvAl andTi andthe moregradual variation in Mg#. B. The triangular substitution plotshowing that zoning in the core is dominated by anedenite-type coupled substitution.

0.86 for the Quintana pluton and 092 for the Strontianpluton. These values are within the range of normalvalues for calc-alkaline granitic rocks as reported bySpeer (1987). Ifthe uniform compositions displayed bybiotite indicate extensive re-equilibration, then the Dvalues may be unrelated to conditions of magmaticequilibrium. The observed differences in the D valuesbetween the two plutons may be explained by differ-ences in the intensive variables (e.g., the greater depthof emplacement of the Strontian pluton) and may not beprimarily related to differences in the bulk compositionofthe respective magmas. Speer (1987) suggested thatthe calculated D depends on the mineral assemblage andbulk composition, with D increasing toward more felsiccompositions. In our case, although the differences inbulk composition are quite small, this correlation isreversed. with lower D values in the more felsiccomposition (Quintana).

902wr./. 97.11"Ito2 3.50At&{r 13.57F€o! 17.18Crzog 0.13MnO 0.32Nlo 0.12MgO t256CaoBaO 0.06l,la2O O.'17t<20 9.25

97.97 36.s{l3.73 3.8t1

t4.1a 13.8817,U 17.070.05 0.140.26 0.20.04 0.09

12.47 't2.38

0-18 0.o70.11 0.128.57 0.09

97.13 96.373.37 4.24

15,77 14.U15.22 18.650.07o.gtt o.0.08

12.49 11.41- 0.02- 0.70

0.10 0.11g.fi 5.42

9{.63 95.21

5.@2518.O

37.43.79

13.66r8.20

0.34

',"r)0.040.30o.129.59

95.45

5.682.92s.@

0.13 0.140.49 0.4{t

z& 2.920.0t) 0.040.@261 2.755.67 5.68

- 0.010.04 0.o20.03 0.03't.85 1.861.92 1.92

16.69 15.61

Stuctml bmuls (O-22)si 5.70 5.69Al0v) 2.s 2.e1Zdla 8.@ 8.00

Al(vr) 0.16 0.4'n 0./O 0.411cr 0.02 0.01Fe2+ 2.21 2.21Mn 0.04 0.03M 0.01Lq 2.8 2.8t1Y db 5.71 5.73

Ca 0.Ol 0.0t1Ba O.OS 0.0t1NaK 1.81 t.66x db '1.a7 1.72

TOTALCAT. 15.58 15.46

€ R O

241s.@

5.672.338.@

0.18 0.17o.44 0.@0.022.m 4330.03 0.(x

- 0.012.AS 2835.7 8.71

0.01o.04

1.78 1.751.gr 1.81

AMPHIBOLE.RICH CLOTS IN CALC.ALKALINE GRANTNC ROCKS

TABLE 4, REPRESENTATIVE ELECTRON_MICROPROBE DATAON BIOTITE FROM THE STRONTIAN GRANITIC ROCKS

AND RELATED ENCLAVES

lt07

Rock typ€ Enclave Enclave TonaliteSample SRz SR6 SR4Analysis 14 I 26Desoiplion ldom. amph. Amph. clot Plag. incl

Tonalite Hybrid HybfidsR4 SRs SB5

st{t 3 32ldiom. amph. ldom. amph. Amph. clot

Structural formulae (O=22)si 5.66Ar (rv) 2.UZ sito 8.00

sio2Ti02AL2O3FoOtCr2oSMnONioMsoCaONa2OK20

TOTAL

Ar (vr)TiFe2+CrMnNiMgY sib

CaNaKX sib

TOTALCAT.

36.47 57.293.34 3.52

14.4 14.62't6.81 17.980.010.26 0.230.02

13. t6 12.210.10 0.020.08 0.109.27 9.68

93.80

5.602.408.00

0.190.392.'t8

0.03

3.0;5.78

0.020.021.421.86

15.64

95.65

5.642.368.00

0.250..rc2.27

0.03

2.7;5.71

0.031.871.90

15.60

38.48 37.614.O4 3.7S

14.8 14.6314.96 15.07o.25 0.040. ' t7 0.14

14.& 14.550.09 0.o7o.12 0.039.54 9.7S

96.80

o.bo

2.358.@

0.160.451.840.030.02

9.215.71

0.010.031.79't.83

95.60

5.6 t2.398.@

0.18o.421.980.010.02

3.233.74

0.010.011.851.87

15.54 15.6't

37.@s.41

14.1216.880.0s0.'t80.02

13.700.010.109.79

95.93

0.160.392.12

0.02

3.0;5.75

0.03'r.88

1.91

15.66

s7.81s.60

14.3217.S80. t50.140.02

15.420.01o.149.71

96.69

5.642.368.@

0.150.€2 .170.02o.o2

2.985.75

0.041.851.89

15.63

estimates of their respective depths of emplacementdeduced from aureole assemblages (Ashworth & Tyler1973). Many amphibole-plagioclase pairs indicate sub-solidus temperatures for both plutons. In the Quintanapluton, mineral pairs from the margin of clots are notsignificantly different from other textural occunencesin the same rock. In the Strontian pluton, mineralpairs within the amphibole clots show the sametrends as idiomorphic crystals of amphibole in contactwith plagioclase either in enclaves, host tonalites and

hybrid rocks. These similarities suggest that plagioclaseand amphibole in the clots crystallized under similarconditions and in equilibrium with other amphibole-plagioclase pairs of the groundmass, where theirmagmatic origin is unambiguous on textural grounds.However, plagioclase only appears near the margin ofthe amphibole clots. Consequently" the interpretation ofa magmatic origin may not be applicable to the innerparts ofthe clots, at least not as simple precipitates froma melt.

I 108

3.0

THE CANADIAN MINERALOGIST

eastonile siderophyllite

phlogopite annite

Frc. 1 l. IvAl versus Fe ratio diagram for biotite related to theamphibole-bearing rocks of Strontian and Quintana. Notethe nearly homogeneous composition for both plutons.

Comparisons with amphibole in the host rock

Amphibole in textural types other than clots has beenanalyzed in order to compare clots with apparentlynormal magmatic parageneses in the same rock. Thesecomparisons are made using the substitution diagrams(Figs. 6,7). The differences between amphibole in clotsand that in idiomorphic crystals are not particularly

QUINTANA

Amphlbole

0 . 7Mol Mg(Mg+Fe)

z0+0.0

STRONTIAN

Amphlbole

large. All the analytical data from different samples aregrouped together in this diagram. For reasons ofclarity,comparisons are made using data from one particularsample from the Strontian pluton (SR4), including anenclave and the host tonalite. Figure 14 shows compo-sitions of amphibole from clots and of idiomorphiccrystals within this sample. In diagrams showing Mg#versus tv Al ardTi versus tvA1, compositions of the clotamphibole vary differently within enclaves and withinthe host rock, and both are distinct from the idiomorphicamphibole in the host granodiorite. Cores, rims andidiomorphic amphiboles of the enclaves show widevariations, whereas compositions of amphibole fromsimilar occurrences in the host granodiorite are tightlygrouped. If re-equilibration during slow cooling affectedamphibole compositions, then it is clear that it was moreeffective in the host tonalite than in the enclave (whichprobably crystallized earlier and more rapidly, as sug-gested by the fine-grained matrix and the presence ofacicular apatite). However, the core-to-rim composi-tional trend for amphibole within the clots differs fromthat for amphibole in clots within enclaves, and bothdiffer from the trend for idiomorphic amphiboles in thehost tonalite (Fig. l4). These trends might reflectdifferences in original composition of the magmas. Ifthis is so, then the amphibole in clots within the hostgranodiorite and the enclaves probably crystallized fromdifferent magmas and in equilibriumwith differentmeltsprior to the mingling event that formed the enclaves(Holden et al. 1987). This conclusion assumes that theamphibole in clots crystallized from a melt phase;however, there is an alternative possibility, that theamphibole is not a primary precipitate from melt, but areaction product of an earlier ferromagnesian phase.Idiomorphic crystals of amphibole probably crystallizedlater than the inner parts of the clots.

DISCUSStoN

Any model for the origin of amphibole-rich clots inthe Strontian and Quintana tonalite and granodioriteplutons must account for cerlain general features ofclotshosted by granite or by enclave. In both cases, a criticalfactor in interpreting the clot-forming process is whetheractinolite in such tonalite-granodiorite rocks and dioriticenclaves is magmatic in origin, as proposed by Pe-Piper(1988), or whether it formed in the subsolidus aftercalcic pyroxene, as conventionally interprete.d (e.9.,Chivas 1981). We have observed oscillatory zoning inactinolite, which might favor a magmatic origin, thoughthe evidence is equivocal, as such zoning might beinherited from similar zoning in a precursor calcicpyroxene. HeIz (1982) observed that tremolite-actino-lite is absent from all experimental runs in metaluminoussystems. She attributed this finding to the inability ofAl-poor amphibole to coexist with liquids of normal(10-2580) Al2O3 content.

0.1

c'o 0.0+

Y+

5 o.r

- 42oE

-0.3

0.1

6o.t 0.0-+dz ^ .

-

0 . 8

6 -0.2E

-0.30 . 6 0 . 7 0 . 8

Mol Mg/(Mg+Fe)

FIc. 12. AFM plots (Speer 1987) showing several biotite-am-phibole pairs from the marginal zone of amphibole-richclots (full circles). Pairs of idiomorphic amphibole incontact witl biotite inclusions also are shown for theStrontian tonalites (open circles) and related enclaves (opensquares).

o

s&f.Fe(Fe+Mg)

T fc)

0 L0 , 5 Sl pfu In Amphlbole

8 0 0 7 0 0("c)ToAR,ESlKDAr ]_ *srRbi TIAN-1

Ir - u l 5

o0 Sl Plu ln AmPhlbole 7 '5

Fro. 13. Graphical solutions to the plagioclase-amphibolethermometer (Blundy & Holland 1990) for 1 kbar (Quinta-na) and 3 kbar (Strontian). The analyzed pairs around clotsfrom enclaves axe plotted as open squares, and the hostgranitic rocks as full squares.

The constraints we consider most important to anymodel of clot formation are:

a) The dimensions of clots are more or less compara-ble with the long axes of the largerprismatic amphibolecrystals in the host rocks.

b) The distribution of clots through a magmatic unitin the host granite is relatively uniform. There are nozones where clots are more numerous.

c) Texturally, the interior and outer margin of clotsare distinct. The interior is dominantly composed ofmultiple crystals of amphibole in mutual contact. Themargin also is amphibole-rich, but with significantproportions of biotite and plagioclase, and virtually noquartz or alkali feldspar. An idiomorphic (prismatic)outline ofclots is observed only in the enclaves, whereasthe outline of clots in the granitic matrix is alwaysinegular.

d) Individual crystals of amphibole in the interior ofa clot tend to be zoned. The core of these crystals is lowin Al (actinolitic), whereas the rim is a calcic amphibolesimilar in composition to the idiomorphic crystals in thehost granitic rock. The amphibole in the margin of clotsis a calcic amphibole indistinguishable from that of thehost rock.

The textural and compositional differences betweenthe interior of a clot and its outer margin stronglysuggest that they have distinct histories of crystal-lization.

I 109

0.1 Ti (ptu) o.2

Frc. 14. Cation plots showing core-rim relationships in zonedamphiboles from enclaves and host tonalites of the Stron-tian pluton (from a single sample).

Origin of clot interiors

The interior of the clots is characterized by multipleindividual grains of amphibole, in mutual contact, eachhaving an actinolitic core surrounded by a rim ofmorenormal calcic amphibole similar in composition to thatof idiomorphic amphibole in the host rock. If theactinolitic core precipitated directly from a magma, itmust have been in equilibrium with a liquid of differentcomposition than the magma that presently hosts theclots. Alternatively, the actinolitic core may be aproductof reaction of pyroxene with magma, fluid, or ofadaptation to changed P-T conditions. The problem isto determine the origin of the pyroxene and to explainwhy individual grains preserved these distinct cores anddeveloped a rim of calcic amphibole in magmaticequilibrium with the host while also developing agranoblastic texture.

The observed compositional zoning is the reverse ofthat anticipated for magmatic crystallization. Accordingto Wones (1981), Cawthorn (1976), and Allen &Boettcher (1978), the silica content of an amphiboleincreases sympathetically with that of the host magma.Cawthom (1976) also observed that an amphibole tendsto be richer in Ti than its host magma, but the core of ouramphibole is relatively depleted in Ti compared to therim. Both features suggest that the actinolitic coreprobably did not crystallize from a magma of similarcomposition to the present host-rock.

Pyroxene precursors to the clot cores are a likelyrationalization of these features:

AMPHIBOLE-RICH CLOTS IN CALC.ALKALINE GRANTNC ROCKS

6 0 07 0 0

0.2 L0.0

, 5

lvlro/.An

t =n kb@r f QUINITANA_

- n

; n

Mg(Mg+FeP+)

t 1 l 0 TI{E CANADIAN MINERALOGIST

(a) Early pyroxene phenocrysts in tonalitic-grano-dioritic magma may be expected before the H2O activityincreased sufficiently to stabilize amphibole. Calcicpyroxene cores to phenocrystamphiboles are commonlyobserved, though it must be stated that this is not the casefor the two plutons in this study.

(b) Mafic and intermediate magmas containing py-roxene phenocrysts have been reporled as mingling andmixing with more silicic magmas (e.g., Cwtro et al.1991a). Survival of pyroxene crystals as grains andnetworks would leave tlese minerals in a new composi-tional and thermal environment, which may lead toextensive re-equilibration. A variant of this proposalmight include pyroxenes formed by the disruption ofmafic igneous enclaves, themselves incorporated in thegranitic magma by processes of magma mingling(Stephens et al.199l).

(c) Pyroxene will dominate the mineral assemblageof a mafic restite after partial melting of a typical I-typesource @utter & Wyllie 1988). Entrainment and trans-port in the magna of such restite are regarded by someworkers as important features of l-type granites (e.9.,Chappell et al. 1987).

Whatever the origin of inferred pyroxene in thernagma, such pyroxene, on entering the stability field ofhornblende, will react with melt to produce an amphi-bole in equilibrium with that crystallizing from the melt.This process of crystal-melt re-equilibration will bemost effective on therim of crystals in contact with me1t,and perhaps along some grain boundaries. Contempora-neously, the magma will precipitate new amphibolemost rapidly on appropriate sites of nucleation, and amass of amphibole crystals in the form of a clot mightrepresent a favorable substrate in this regard. If the rateof this amphibole-accretion process is considerablymore rapid than the process of crystal-melt re-equilibra-tion, then the effectiveness of the laner will rapidlydiminish as the clot becomes armored by a new cycle ofamphibole growth. The net effect might be to a:rest tfiere-equilibration of pyroxene grains with melt, leaving apyroxene core surrounded by a hornblende rim.

The most difficult problem in explaining these char-acteristics ofclots is to account for the actinolitic core togranoblastic crystals within the clot interiors. In our firstmodel for re-equilibration, we assume that:

pyroxene + meltl -+ hornblende + me1t2.

The low aluminum content of the pyroxene favors theformation of actinolite in the solid state in a reactionisolated from melt, such as:

pyroxene + fluids (OH, ...) + actinolite.

Such a reaction may occur once the pyroxene isisolated from melt by the armoring of the clot byhornblende crystals.

The second model regards the interior ofapyroxene-rich clot as isolated from contact with the melt and isbased on the observation that hydrogen diffrrses through

silicate melt and along grain boundaries much morerapidly than all other elements (Goldsmith 198O. It isproposed that a hypothetical precursor clinopyroxenegrain isolated within a clot reacts with hydrogen to forman actinolitic amphibole and some iron, according to themodel reaction:

4Caa.5Mge.75Fq.75Si2O6 + 2H -->Ca2Mg3FqSisO22(OH)2 + Fe

cpx amph

Goldsmith (1987) proposed that AVSi interdiffusionin albite is facilitated by hydrogen hopping and sug-gested that the extent of equilibration may be as muchdependent on the availability ofhydrogen as on time. Inan analogous way, we propose that the above reactiondominates while the supply of protons is large comparedwith other cations. The rate of H+ diffr.rsion in pyroxenecrystals is very high, and this will not limit ttte rate ofthe reaction (Skogby & Rossman I 989). However, as themore sluggishly diffusing Al and other cations becomeavailable at the grain boundaries, then equilibration ofthe actinolite with the more aluminous amphibole of thehost magma may take place, according to:

Ca2Mg3FqS(O zz(O$z + 2 Al -)Ca2Mg3FqSi6,Al2o22(OH)2 + 2Si

low-Alamphibole aluminousamphibole

Thus the actinolite represents a fiansient state andintermediate stage between the precursor pyroxene andthe hornblende ofthe host magma, and is preserved onlyto the extent that the kinetic control on the availabilityof other cations allows. This model could explain thetexture of individual grains of the clot interior. Completereaction with protons transforms the whole grain toactinolite. whereas with mass transfer within the clotstaking place dominantly by grain-boundary diffirsion,equilibration with aluminous amphibole of the magmatakes place inward fromthe grain margin. This processmight be prevented from going to completion either bynew melt-precipitated growth on the clot exterior,choking off access to the clot interior, or by loss ofenergy from the system in a cooling pluton, reducing therates of diffusion to ineffective levels.

Origin of the clot m.argins

Characteristics of the margin of the clots are entirelyconsistentwith crystallization from the hostmagma. Thecompositions of amphibole, biotite and plagioclase donot differ from those in the host granitic rock, andtextures are magmatic. It seems that the clot as a wholemay have acted as a suitable substrate for nucleationduring the early and marn stages of crystallization.Late-stage minerals such as quartz and alkali feldspar arenot abundant amongst the minerals making up the clotmargins. Clots in the enclaves are distinct from those inthe host granitic rocks in tending to maintain a prismatic

AMPHIBOLE.RICH CLOTS IN CALC-ALKALINE GRANTTIC ROCKS n l l

at the University of St Andrews during 1990 wassupported by a sabbatical grant from the lPlnlthtvtinistry of Education and Science (DGICYT). WES isgrateful for financial assistance from the NERC (GrantCnfl+O::'l for fieldwork at Strontian. Constructivecomments by S.E. Haggerty and an anonymous refereeare acknowledged with thanks.

REFERENCES

Ar-r-eN. J.C. & BosmcrEn, A.L. (1978): Amphiboles inandesite and basalt. II. Stability as a functron ofP-T-fs2e-f62. Am. Mineral. 63, 1074-rc87.

Asuwonrs, J.R. & Tvr-sn, I.M. (1973): The distribution ofmetamorphic temperatures around the Strontian granodio-irte. Geol. Mag. 120,281-290.

BLrrNDy, J.D. & Hor-r-lND, T.J.B. (1990): Calcic amphiboleequilibria and a new amphibole-plagioclase geothermome-ter. Contrib. Mineral. Petrol lM,208-224.

CANTAGREL. J.-M., DprEn, J. & Gouncaun' A. (1984): Magmamixing: origin of intermediate rock and "enclaves" fromvolcanism to plutonism. Piys. Eanh Planet. Int.35'63-76.

CAsrRo, A. (1985): The Central Extremadura Batholith: geo-tectonic implications @uropean Hercynian Belt) - anoutline. T e c t ono p hY s. l2A, 51 -68.

- (1986): Structural pattern and ascent model in theCentral Extremadura batholith, Hercynian belt' Spain. "/.Struct. G e ol. 8, 633-&5.

- (1990): Microgunular enclaves of the Quintana grano-diorite (-os Pedroches batholith): pefiogenetic signifi-cance. ftev. Soc. Geol. Espafla 3' 7'21.

-, DE LA RosA, J.D. & SrepmNs, W.E. (1991a): Magmamixing in the subvolcanic environment: petrology of theGereni interaction zone near Seville, Spatn. Contrib.M ineral. P et rol. 106, 9-26.

-, MoRENo-VEMAS, I. & DE LA RosA, J.D. (1991b):H-(hybrid) type granitoids: a proposed revision of thegranite type classification and nomenclature. Earth-Sci'Rev.31.237-253.

& - (1991c): Multistage crystallization oftonalitic enclaves in granitoid rocks (Ilercynian belt,Spain): implications in magma m\xing. Geol. Randschnu80. 109-120.

CAw'moRN, R.G. (1976): Some chemical controls on igneousamphibole compositions. Geochim- Cosmochim. Acta 40,t3t9-1328.

CHAppEr.L. B.W., WHrrr, A.J.R. & WYBoRN, D. (1987): TheimDortance of residual source material (restite) in granitepetrogenesis. "/. Petol. ?3, 1 I 1 1-1 138.

CHNAs, A.R. (1981): Geochemical evidence for magmatic

shape indicative of an origin by pseudomorphic replace-ment ofpyroxene. Enclaves in the Strontian pluton havebeen shown to have formed from exotic magmas ofmantle derivation (Holden et al. 1987), probably asnetworks of mafic and other high-T phases resultingfrom quenching of a mafic magma in a felsic graniticmagma (Stephens et al. 1991). These networks formedcoherent structures in which the pyroxenes reacted in alocal environment to give amphibole.

CoNcLUSIONS

Our study of mafic clots in two plutons has revealedsome apparently general characteristics of amphibole-dominated clots. The clots are zoned, with an interiorcomprising multigain polygranular aggregates of am-phibole surrounded by an amphibole-biotite-plagio-clase assemblage forming a continuous margin. Theinterior grains are themselves zoned from an actinoliticcore to a calcic amphibole rim.

We conclude that the clots are primarily the productsof reactions with pyroxene occurring in a metaluminousmagma precipitating calcic amphibole. Two possibleorigins are proposed: (a) $roxene re-equilibrated withthe melt to form calcic amphibole in equilibrium withcalcic amphibole crystals growing in the host, and newamphibole also grew with plagioclase and biotite atmargins. When this marginal growth choked off accessof the melt to clot interiors, the remaining cores ofpyroxene reacted in the solid state to form actinoliticamphibole, presumably in the presence of a fluid phase.(b) Pyroxene in the clot interiors was isolated from melt,and grarn-boundary diffusion of protons from the meltpromoted recrystallization of pyroxene to actinolite.Later, the more sluggish diffitsion of Al and othercomponents brought about partial re-equilibration withthe more aluminous amphibole crystallizing from themelt. This re-equilibration was prematurely arrestedwhen growth of amphibole, plagioclase andbiotite attheclot margins choked off access, or when diffirsion ceasedto be an effective method of mass transfer for thermo-dynamic reasons.

The origin of the precursor pyroxene crystals is notobvious from this study. Possibilities include originalphenocrysts, or derivation through the mingling of thegranodioritic magma with more basic magm4 or theymay simply have been restitic from the source. Theiruniform distribution is consistent with either an originas phenocryst or a restite, whereas their granoblastictexture favors a restitic origin.

AcKNowLEDGETdENTS

The authors are grateful to Donald Herd for assistancewith the electron microprobe, and to Peter Holden andBruce Paterson for discussions about the Strontianpluton. Roger Bateman provided useful suggestionsfrom a critical reading of the manuscript. The stay of AC

fluids in porphyry copper mineralization. Contrib. Mineral.Petrol. 78.389403.

CZAMANSKE, c.K. & WoNEs, D.R. (1973): Oxidation duringmagnatic differentiation, Finnmarka complex, Oslo area,Norway. 2. The mafic silicates. "/. Petrol. L4,349-380.

Got-osunH, J.R. ( I981: AVSi interdifflrsion in albite: effect ofpressure and the role ofhydrogen. Cantrib. Mineral. Petrol.9 5 . 3 I t - 3 2 1 .

HALLIDAY, A.N., Arrelrol, M., vAN BREEMEN, O. & JocELyN,J. (1979): Petrogenetic significance of Rb-Sr and U-Pbisotopic systems in the 400 Ma old British Isles granitoidsand their hosts. 1lr The Caledonides of the British Isles -reviewed (A.L. Harris, C.H. Holland & B.E. Leake, eds.).Geol. Soc. lnndon, Spec. Publ.8,653-661.

HAMILToN, P.J., O'NtoNs, R.K. & PaNrnunsr, R.J. (1980):Isotopic evidence for the provenance of some Caledoniangranites. N at ure 2f7, 27 9-284.

Hnz, R.T. ( 1982): Phase relations and compositions of amphi-boles produced in studies of the melting behavior of rocks.1n Amphiboles: Petrology and Experimental Phase Rela-tions (D.R. Veblen & P.H. Ribbe, eds.). Rev. Mineral.9B,z t9-5) 5 -

HoLDEN, P., HA[lDAy, A.N. & SrerHrNs, W.E. (1987):Neodymium and strontium isotope content of microdioriteenclaves points to mantle input to granitoid production.Nature 330, 53-56.

Hu'rroN, D.H.W. (1988a): Granite emplacement mechanismsand tectonic controls: inferences from deformation studics.Trans. R. Soc. Edinburgh, Earth S,ci.79,245-255.

- ( 1988b): Igneous emplacement in a shear-zone termina-tion: the biotite granite at Srrontian, Scotland. Geol. Soc.Am. Bull. 100, 1392-1399.

Julrvnnr, M., FoNrBo'r€, J.M., Rrnrrno, A. & CoNDE. L.N.(1974): Memoria explicativa del Mapa Tect6nico de laPeninsula Ib6rica y Baleares. IGME, Madrid.

Lrers, B.E. (1978): Nomenclature of amphiboles. Can. Mine-ral.16,501-520.

Lt-ovo, G.E. (1987): Atomic number and crystallogaphiccontrast images with the SEM: a review of backscatteredelectron techniques. Mineral. Mag. 51,3-19.

MACGREGoR, A.G. & KrNlely, W.Q. (1932): The Morvem-Strontian "Granite". Summ. Progress Geol. Surv. for l93 l- I I ' 105-119.

MAzzoNE, P., S'rEwARr, D.C. & Huours, J.M. (1987): Sub-so-lidus dehydration of amphiboles in an andesitic magma.Contrib. M ine ral. P etrol. W, 292-296.

t l t2 THE CANADIAN MINERALOGIST

PRESNALL, D.C. & BamvnN, P.C. (1973): Fusion relations inthe system NaAlSi3Os-{aAl2Si2O8-KAlSi3Os-SiOz-H2O and generation of granitic magmas in the SierraNevada batholith . Geol. Soc. Am. Bull. M,3181-3202.

RoBrNsoN, P., SPEAR, F.S., SCHUMAGDR, J.C., LAIRD, J., Kr-srN,C., EvANs, B.W. & DoouN, B.L. (1982): Phase relationsofmetamorphic amphiboles: natural occurences and theo-ry. /z Amphiboles: Petrology and Experimental PhaseRelations (D.R. Veblen & P.H. Ribbe, eds.). Rev. Mineral.9B.. t-227.

RocERs, G. & DUNNTNG, G.R. (1991): Geochronology ofappinitic and related granitic magmatism in the WestHighlands of Scotland: constraints on the timing oflranscunent fault movement. J. Geol. Soc. Inndon 14,i.l '7-)'7

RurrER, M.J. & WvLLE, P.J. (1988): Melting of vapour-absenttonalite at l0 kbar to simulate dehydration-melting in thedeep crust. Natare 331, 159-160.

SasNr, P.A. (1963): The Strontian granite complex, Argylls-hire. Bull. Geol. Surv. Great Britain 20, 642.

Scanr, , C.M. & Furu, T. (1987): Petrology of crystal clots inthe pumice of Mount St. Helens' March 19, 1982 eruprion;significant role of Fe-Ti oxide crystallization. "/. Volcatnl.Ge othe rm. Re s. 34. | - | 4.

Srocsv, H. & RossuaN, G.R. (1989): OH- in pyroxene: anexperimental study ofincorporation mechanisms and stab!lity. Am. Mineral.74, 1059-1069.

Spesn, J.A. (1987): Evolution of magmatic AFM mineralassemblages in granitoid rocks: the homblende + melt =biotite reaction in the Liberty Hill pluton, South Carolina.Am. M ine ral. 72, 863-87 8.

SrEpHENs, W.E., HoloeN, P. & HENNEY, P.J. (1991): Microdio-rific enclaves within the Scottish Caledonian granitoids andtheir significance for crustal magmatism. 1z Enclaves andGranite Petrology (J. Didier & B. Barbarin, eds.). Elsevier,Amsterdam (125-134\.

S'rREcKErsEN, A. ( 1976): To each plutonic rock its proper name.Earth-Sci. Rev. 12. l-33.

VsnroN, R.H. (1983): Restite, xenoliths and microgranitoidenclaves in granites. J. Proc. R. Soc. New South Wales lL6,77-103.

WNCHESTER, l.A. (1974): The zonal pattem of regional meta-morphism in the Scottish Caledonides.J. Geol. Soc. lnndon130,509-524.

WoNss, D.R. (1981): Mafic silicates as indicators of intensivevariables in granitic magmas. Mining Geol.3l,l9l-212.

MuNRo, M. (1973): Struclures in the south-eastem portion ofthe Strontian granitic complex, Argyllshire. Scott. J. Geol. ZoRPl, M.J., CouLoN, C., ORsNl, J.B. & Coctnra, C. (1989):9, 99- 108. Magma mingling, zoning and emplacement in calc-alkaline

granitoid plutons. Tectonophys. 157, 315-329.Pe-PtpEn, G. (1988): Calcic amphiboles ofmafic rocks ofthe

JeffenBrookplutoniccomplex,NovaScoti4Canada.Arz. Received August 13, 1991, revised manuscript acceptedMineral.73,993-1006. Januam 11.1992.