JMG_ShervDenn2003

8/16/2019 JMG_ShervDenn2003

1/16

Deep in the Heart of Dixie: Pre-Alleghanian Eclogiteand HP Granulite Metamorphism in the Carolina Terrane,South Carolina, USA

J . W . S H E R V A I S ,

1

A . J . D E N N I S ,

2

J . J . M C G E E

3

A N D D . S E C O R

3

1Department of Geology, Utah State University, Logan UT, 84322, USA ([email protected]) 2Department of Geology and Biology, University of South Carolina, Aiken SC 29801, USA3Department of Geological Sciences, University of South Carolina, Columbia SC, 29208, USA

AB ST R ACT The central part of the Carolina terrane in western South Carolina comprises a 30 to 40 km wide zone of high grade gneisses that are distinct from greenschist facies metavolcanic rocks of the Carolina slate belt(to the SE) and amphibolite facies metavolcanic and metaplutonic rocks of the Charlotte belt (to theNW). This region, termed the Silverstreet domain, is characterized by penetratively deformed felsicgneisses, granitic gneisses, and amphibolites. Mineral assemblages and textures suggest that these rocksformed under high-pressure metamorphic conditions, ranging from eclogite facies through high-Pgranulite to upper amphibolite facies.

Mafic rocks occur as amphibolite dykes, as metre-scale blocks of coarse-grained garnet-clinopyroxeneamphibolite in felsic gneiss, and as residual boulders in deeply weathered felsic gneiss. Inferredomphacite has been replaced by a vermicular symplectite of sodic plagioclase in diopside, consistent with

decompression at moderate to high temperatures and a change from eclogite to granulite facies con-ditions. All samples have been partially or wholly retrograded to amphibolite assemblages. We infer thefollowing P-T-t history: (1) eclogite facies P-T conditions at ‡ 1.4 GPa, 650–730 C (2) high-P granulite

facies P-T conditions at 1.2–1.5 GPa, 700–800 C (3) retrograde amphibolite facies P-T conditions at0.9–1.2 GPa and 720–660 C. This metamorphic evolution must predate intrusion of the 415 MaNewberry granite and must postdate formation of the Charlotte belt and Slate belt arcs (620 to 550 Ma).

Comparison with other medium temperature eclogites and high pressure granulites suggests that theseassemblages are most likely to form during collisional orogenesis. Eclogite and high-P granulite faciesmetamorphism in the Silverstreet domain may coincide with a 570–535 Ma event documented in the

western Charlotte belt or to a late Ordovician-early Silurian event. The occurrence of these high-Passemblages within the Carolina terrane implies that, prior to this event, the western Carolina terrane(Charlotte belt) and the eastern Carolina terrane (Carolina Slate belt) formed separate terranes. Thecollisional event represented by these high-pressure assemblages implies amalgamation of these formerlyseparate terranes into a single composite terrane prior to its accretion to Laurentia.

Key words: amphibolite; Carolina terrane; southern Appalachians; eclogite; HP granulite.

I NT R O DUCT I O N

High-pressure granulites, characterized by the ortho-pyroxene-free assemblage Grt + Cpx + Pl ± Qtz,comprise a newly recognized subfacies transitionalbetween plagioclase-free eclogites and orthopyroxene-bearing granulites (Pattison, 2003). O’Brien & Ro ¨ tzler,2003) distinguished two varieties of high-P granulite:ultra-high temperature assemblages with melt reactiontextures, and medium-T , high-P assemblages (700– 850 C, 1.0–1.4 GPa) that overprint former eclogitefacies assemblages. Like medium temperature (MT)eclogites (Carswell, 1990), the medium-T , high-Pgranulite subfacies is typically associated with colli-sional orogens, which form in tectonically thickenedarc or continental crust, typically in response to theattempted subduction of an arc or continental margin

during collision (e.g. Carswell, 1990; O’Brien & Ro ¨ t-zler, 2003). In many areas, these rocks are commonlyassociated with retrogressed felsic gneisses that wereoriginally cofacial with the enclosed eclogites (e.g.Cuthbert & Carswell, 1990; Cuthbert et al ., 2000;O’Brien et al ., 1990).

The eastern margin of North America in thesouthern and central Appalachians comprises a tec-tonic collage of terranes that formed in exotic locationsduring the late Neoproterozoic through early Palaeo-zoic, and were subsequently accreted to Laurentiaduring the mid- to late Palaeozoic (Williams &Hatcher, 1983; Secor et al ., 1983; Horton et al ., 1989,1991; van Staal et al ., 1998). These exotic terranesevolved independently of Laurentia for much of theirexistence, and preserve evidence of orogenic andmagmatic events that are not observed in Laurentia.

J. metamorphic Geol., 2003, 21, 65–80

Blackwell Science Inc., 0263-4929/03/$15.00 65Journal of Metamorphic Geology, Volume 21, Number 1, 2003


2/16

One of the most extensive of these exotic peri-Gondwana terranes is the Carolina terrane, whichcomprises a large portion of the southern Appalachianorogen east of the Blue Ridge province (Secor et al .,1983; Fig. 1). The Carolina terrane is an exoticAvalonian terrane that originally formed adjacent toGondwana in the late Neoproterozoic, and was notaccreted to Laurentia until the mid- to late Palaeozoic(Secor et al ., 1983; Williams & Hatcher, 1983).

We have recently re-examined a little knownoccurrence of high-P granulite and amphibolite, withan inferred MT eclogite precursor, within the centralpart of the Carolina terrane (Dennis et al ., 2000).These rocks, which were originally interpreted aspyroxene-bearing garnet amphibolites, contain relict

garnet-pyroxene-plagioclase assemblages that record apreviously unrecognized episode of eclogite trans-itional to medium temperature HP granulite faciesmetamorphism within the Carolina arc terrane. Thisevent has broad implications for the evolution of thesouthern Appalachians, and for models of metamor-phism and exhumation in accreted arc terranesin general. We present here a first look at thesenewly discovered high pressure rocks, their inferredP–T–t history and some tectonic implications of theiroccurrence.

ECLO GI T ES AND GR ANULI T ES O F T HE

CAR O LI NA T ER R ANE

Regional Setting

The Carolina terrane in the southern Appalachians is a calc-alkalineisland arc that is exotic to Laurentia and does not share a commonhistory with North America until the late Palaeozoic Alleghanianorogeny (Fig. 1). It is largely Neoproterozoic in age but includessections of early to middle Cambrian age (Secor et al ., 1983; Samsonet al ., 1990; Shervais et al ., 1996; Dennis & Shervais, 1996; Wortmanet al ., 2000). Recent field and geochronological studies showthat the Carolina terrane formed during two major episodes of arcmagmatism at 620 Ma and 550 Ma (Dennis & Wright, 1997;Heatherington et al ., 1996).

The Carolina terrane has been divided into three belts with dif-ferent metamorphic and petrological characteristics: (1) the Kings

Mountain belt, which consists of greenschist facies mafic metavol-canic rocks and forms the north-western margin of the Carolinaterrane; (2) the Charlotte belt, which consists largely of lower tomiddle amphibolite facies, dominantly mafic metavolcanic and meta-plutonic rocks; and (3) the Carolina Slate belt, which is dominated bylow-grade (greenschist to subgreenschist) felsic metavolcanic rockswith subordinate mafic lavas and mudstones (Fig. 1).

The Carolina terrane was metamorphosed and ductilely deformedduring the latest Neoproterozoic to early Cambrian (Dennis &Wright, 1995, 1997; Hibbard & Samson, 1995; Barker et al ., 1998).Metamorphism and ductile deformation resulting from the Alle-ghanian (320 Ma) collision of Laurentia and Gondwana isrestricted to narrow shear zones which separate broad zones con-taining older fabric and mineral assemblages (e.g. Secor et al ., 1986;Dallmeyer et al ., 1986; Horton et al ., 1989; Horton & Dicken, 2001).The Charlotte belt was intruded by a suite of undeformed Devoniangabbros and granitoids (400 Ma; McSween et al., 1991) that cross-cut regional foliation and mark the upper age limit of penetrativedeformation within most of the terrane.

The exotic nature of the Carolina terrane is shown clearly by theoccurrence of a diverse Middle Cambrian peri-Gondwanan trilobitefauna in the Carolina Slate belt (Samson et al ., 1990). In addition,combined field-geochronological studies have shown that metamor-phic fabric in most of the Carolina terrane formed prior to 535 Ma,approximately coeval with the rift-drift transition on the Laurentianmargin (Dennis & Wright, 1995, 1997; Hibbard & Samson, 1995;Barker et al ., 1998).

Field Occurrence of High-Pressure Rocks

The boundary between the Charlotte belt and the Slate belt in centralSouth Carolina comprises a 30-km wide zone of high grade gneissesthat are distinct from less highly deformed amphibolite facies(dominantly) mafic rocks of the Charlotte belt (to the northwest) andlow-grade felsic metavolcanic rocks of the slate belt (to the south-east; Fig. 2). This region, termed the Silverstreet domain, consists of high-grade felsic biotite gneisses, granitic gneisses, and amphibolitesthat form the SE margin of the Charlotte belt (Secor et al ., 1982,1988; Halik, 1983; Hauck, 1984). The Silverstreet domain is intrudedby the undeformed early Devonian Newberry granite (415 ± 9 Ma;Fullagar, 1981; Samson & Secor, 2000), which cross-cuts regionalfoliation and includes xenoliths of sheared and foliated country rock.The age of this pluton represents an uppermost age limit forformation and deformation of rocks in the Silverstreet domain.

The Silverstreet domain is bounded to the north and south byshear zones. A variety of field studies have demonstrated that the

Fig. 1. Regional geology of the southernAppalachians, showing principal sub-divi-sions, including the Carolina terrane (palegrey), the Blue Ridge terrane (dark grey), theInner Piedmont terrane, and the AtlanticCoastal Plain. Rocks of the Inner Piedmontterrane (including the Chauga belt) and theCarolina terrane (including the Charlottebelt, the Carolina Slate belt, and the KingsMountain belt [KMB]) are all exotic toNorth America.

6 6 J . W . S H E R V A I S E T A L .


3/16

boundary between the high-grade gneisses of the Silverstreet domainand the Carolina Slate belt is a fault over much of its length, but itsgeometry and kinematics are not known (e.g. Dennis et al ., 2000;Offield, 1995; Offield & Sutphin, 2000; Secor et al ., 1988, 1989). It isinferred to be a normal fault in part because it juxtaposes high-gradeCharlotte belt rocks against low-grade Slate belt rocks. Locally,however, it can be demonstrated that most recent ductile motionalong the Stony Hill orthogneiss was right lateral, based on com-posite planar fabric and asymmetric porphyroclasts (Dennis et al .,2000). The northern margin of the Silverstreet domain is a 10-kmwide, E-W trending ductile shear zone (Beaver Creek shear zone)

with dextral shear sense indicators that separates it from lessdeformed rocks of the Charlotte belt (West, 1998).Mafic rocks in less deformed parts of the Silverstreet domain form

amphibolite dykes up to 20 cm thick that are oriented parallel toregional foliaton. In more deformed areas, isolated metre-scaleblocks within felsic gneiss are interpreted to represent boudinagedmafic dykes (Fig. 3). In many areas, these blocks form residualboulders that have weathered out of the felsic gneisses; where theseoccur in flat upland terrain they are interpreted to be approximatelyin place.

Retrogressed eclogite and high-P granulite assemblages commonlyare preserved in the cores of these isolated blocks. Blocks with relicthigh-pressure assemblages are found within the Beaver Creek shearzone along the north side of the Silverstreet domain and as residualboulders near the centre of the terrane, south of the Newberrygranite, and clearly outside of the Beaver Creek shear zone (Fig. 2).Foliation in the shear zone wraps around the eclogite blocks and

clearly postdates eclogite formation.

Mineralogy and Mineral Chemistry

Mafic rocks that preserve high-pressure assemblages are modallydiverse with 20–40% pink garnet, 20–60% green, diopsidic clino-pyroxene, 15–45% hornblende, up to 10% plagioclase, and 3–5%ilmenite, with accessory rutile, epidote, apatite, zircon, titanite andcalcite (e.g. Libby & Carpenter, 1969). Relict garnet and pyroxenegrains are up to 1 cm diameter, but typical grain size for relict phasesis 1–3 mm (Fig. 4). In thin section, hornblende, plagioclase andilmenite replace clinopyroxene, while garnet is replaced along itsmargins by kelyphitic intergrowths of plagioclase with hornblende

and minor epidote. Calcite forms irregular veins and patches. Epi-dote, quartz, plagioclase, hornblende and oxides are also found as

inclusions in garnet, with epidote being the dominant inclusionphase. Representative electron microprobe analyses from one sampleare presented in Table 1; these data are presented graphically inFig. 5. Analytical methods are presented in Appendix A: Methods.

Where it has been well preserved, diopside is characterized by avermicular symplectite of sodic plagioclase (An15)22) that we inferrepresents the breakdown of omphacite; this is clearly shown by bothBSE images and high-resolution X-ray composition maps of thesymplectites (Fig. 6). The diopside contains about 15% jadeitecomponent, but modal reconstruction (see Appendix A: Methods)suggests that primary omphacite contained 30% jadeite. Thereconstructed omphacite is presented in Table 1.

The breakdown of omphacite to diopside + plagioclase sym-plectite is commonly observed in high-P granulites after a medium

Fig. 3. Field photo of eclogite block in felsic gneiss of BeaverCreek shear zone. Block is about 1.5 m across, with a trapezoidalshape; foliation in the shear zone wraps around the block. Thisand other blocks were sampled using portable core drill.

Columbia

Lake Murray

L . G r e e n w

o o d

Clinton Joanna Newberry NW BlairWhitmire S

Cross Hill Bush River

Good Hope

Newberry W

Saluda N

Pomaria

Chappells

Jenkinsville

DysonLittle Mtn ChapinSilverstreet Prosperity

DelmarDenny

Newberry E

W h i t m i r

e R ee n t r

a n t : I n n e

r P i e d mon t

Wateree L.

Carolina sla te bel t and K io

kee b e l t

8 2 ¡ W

34¡N

34¡30'N

8 1 ¡ W

Salem X-roads

B e a v e r

C r e e k s h

ear zone

E

?

S t o n e y H i l l

o r t h o g n e i s s

Clinton

Winnsboro granite295 ± 2 Rb-Sr w.r.

Newberry granite414 ± 8 Ma U-Pb z.

Little Mtnmetatonalite 550 ± 4 Ma U-Pb z.

E

E

E

E

E

E

High grade Silverstreet domain:eclogite, high-P granulite andenclosing felsic gneisses

c . 414 ± 8 Ma (U-Pb z) Newberry granite

c . 295 ± 4 Ma (Rb-Sr w.r.)Winnsboro granite

Little Mountain metatonalite orthogneiss

E known eclogite - high P granulite localities

area of Secor and others (1982)investigation outlined in dash

Mesozoic brittle faults

Orientation of some major structuresin Carolina slate belt and Kiokee belt

7.5' quadrangles indicated and namedin SE corner. This report focussed onquadrangles shown in bold.

++

+

++

Fig. 2. Geological map showing location of eclogite ⁄ high pressure granulite-bearing Silverstreet domain of the Charlotte belt relativeto the Carolina Slate belt and the Whitmire reentrant of the Inner Piedmont. Capital E shows location of known eclogite ⁄ granuliteblocks.

D E E P I N T H E H E A R T O F D I X I E 6 7


4/16

temperature eclogite assemblage, and is consistent with decompres-sion at moderate to high temperatures (e.g. Elvevold & Gilotti, 2000;O’Brien & Ro ¨ tzler, 2003). The inferred primary assemblageomphacite + garnet ± rutile is consistent with formation undereclogite facies conditions, whereas the observed assemblage diopside+ plagioclase + garnet represents high pressure granulite faciesconditions (Galan & Marcos, 2000; Cooke et al ., 2000; Pattison,2003; O’Brien & Ro ¨ tzler, 2003). The breakdown of diopside andgarnet to form amphibole + plagioclase + epidote + ilmeniterepresents final equilibration under amphibolite facies conditions.Formation of amphibolite occurred in two stages. The first is rep-resented by aluminous pargasite and relatively calcic plagioclase(An26)53), which replace diopside + sodic plagioclase symplectites;the second is represented by magnesian pargasite and more sodicplagioclase (An17)22) which replace both diopside and garnet. Cal-culated hornblende-plagioclase temperatures (next section) suggestthat the aluminous pargasite-calcic plagioclase pairs formed at highertemperatures than the magnesian pargasite-sodic plagioclase pairs,and that they are closely associated with the high-P granuliteassemblages.

High resolution X-ray composition maps of garnet show twodistinct growth zones (Fig. 7). The inner zone is enriched in Mn andFe, the outer zone is enriched in Ca and Mg. The Ca and Fe X-raymaps show a sharp interface between the inner core and the outer

mantle, while Mg and Mn show smooth, continuous zoning profiles(Fig. 7). Note that these garnet are generally not symmetricallyzoned: the centre of growth typically lies close to one edge of thegrain. In the example shown here, several small spessartine-richgarnet cores (seen as high Mn spots in the X-ray maps) have beensubsumed by the garnet mantle as it grew.

Compositional profiles selected to traverse from the true core torim confirm these trends. A 1500-lm traverse of the grain mapped inFig. 7 shows smooth profiles for pyrope (Prp) and spessartine (Sps),and sharp steps in profiles for grossular (Grs) and almandine (Alm;Fig. 8). Profiles for three additional garnet are shown in Fig. 9,scaled to percentage of total grain radius. Although these threegrains vary somewhat in their innermost core compositions (Grs,Sps), they display consistent profiles for all elements, with (a) an

inner zone (0 to 55% of radius) that is low in Grs and Prp, and highin Alm and Sps, and (b) an outer zone (60 to 100%) that is higher inGrs and Prp, and lower in Alm and Sps (Fig. 9). All grains exhibit asharp increase in Grs and decrease in Alm at the transition (c. 60% of grain radius) that implies an abrupt change in growth history.

R ESULT S

Geothermobarometry

We infer from the data presented above that the garnetcores formed during prograde metamorphism atgreenschist or amphibolite facies conditions, followedby growth of the garnet mantles at eclogite (Grt-Omp-Rt) and then high-P granulite facies conditions(Grt-Di-Hbl-Pl-Ilm); retrograde metamorphism in theamphibolite facies resulted in the breakdown of garnetand formation of the late Hbl-Pl-Ilm assemblage.Because these rocks experienced a range of metamor-phic conditions, a number of assumptions are made indetermining which compositions to use for thermo-barometry.

We assume that the reconstructed omphacite was inequilibrium with the more Mg-Ca-rich mantles of thegarnet, and that the diopside-plagioclase symplectiteswere in equilibrium with the more Mg-Ca-rich,Fe-poor outermost rims of the garnet (e.g. Fig. 9). Forpurposes of calculation, three garnet mantle composi-tions were used: (a) an average of all garnet mantlesfrom the profile shown in Table 1a (b) an average of allgarnet mantles in Fig. 9, from 65 to 99% of grainradius, and (c) the garnet mantle farthest from the

Fig. 4. Probe mount (2.5 cm diameter) of

eclogite ⁄ granulite from central Carolinaterrane, sample NEW-1–3. Pink ¼ garnet,pale green ¼ pyroxene and pyroxene-plagio-clase symplectite, brown ⁄ darkgreen ¼ hornblende, clear ¼ plagioclase orcalcite, black ¼ ilmentite or Fe-oxides. Noteplagioclase-rich kelyphite rims on garnet.

6 8 J . W . S H E R V A I S E T A L .


5/16


6/16


7/16

Fig. 6. X-ray composition maps of diopside-sodic plagioclase symplectites (¼ former omphacite) surrounded by hornblende, plagioclasNew-1; (B) Ca map, New-1; (C) Al map, 3080E (D) Al map, 3080E. (A, B) Field of view ¼ 5 mm, hotter colours equal higher concenplagioclase-hornblende aggregates, which seem to pseudomorph the primary omphacite. (C, D) Field of view 2.5 mm. Lighter shades ¼

A B C D

Fig. 7. X-ray maps of zoned garnet surrounded by hornblende with minor plagioclase and ilmenite. A ¼ Fe, B ¼ Mn, C ¼ Mg, Din core, with higher Mg and Ca in mantle. Note the sharp contact between the inner garnet core and the outer garnet mantle seen cleaNote also the small garnet cores (high Mn and Fe, low Ca and Mg) that have been subsumed by the garnet mantle. Hotter colours equand Mg; darker blue equals higher Ca. Black line in A is approximate location of line profile (Table 1a). Field of view is 5 mm in all m


8/16

outermost rim analysis of the garnet in Table 1a andselected hornblende-plagioclase pairs from Table 1c.Hornblende-plagioclase temperatures (Holland &Blundy, 1994) were also calculated using the data fromTable 1c, which represent adjacent hornblende-plagio-clase pairs from a range of textural associations. Wealso used GRIPS (garnet-rutile-ilmenite-plagioclase-silica, Bohlen & Liotta, 1986) to estimate pressure, sinceilmenite is associated with the hornblende-formingreactions.

Calculation of equilibrium T-P conditions was car-ried out using the program GTB of Spear & Kohn(2001), and the hornblende-plagioclase program of Holland & Blundy (1994). The pyroxene quadrilateralthermometer of Lindsley & Anderson (1983) is usedfor diopside (but not reconstructed omphacite) becauseits nonquadrilateral components are less than 20%

(Fig. 5). The results from this graphical solvus thermo-meter 700–800 C, or up to 50 C higher thanresults from the garnet rim-diopside thermometers(700–750 C). Garnet-clinopyroxene temperatureswere calculated using stoichiometry to partition totalFe between Fe2+ and Fe3+; calculation of Fe3+ stoi-chiometrically is strongly dependent on analytical

precision and generally overestimates Fe3+ in pyroxenedue to vacancies in the pyroxene lattice (e.g. Robinson,1980). Calculation of garnet-diopside temperaturesassuming total Fe as Fe2+ results in temperatures50–80 C higher than those using calculated Fe3+

(Table 2). The c. 50 C difference between the pyrox-ene solvus temperatures and garnet-diopside tempera-tures calculated here probably results from theoverestimation of Fe3+ using stoichiometry.

Variations in hornblende-plagioclase temperaturescorrelate with texture and composition, as noted ear-lier. Aluminous pargasite and calcic plagioclase thatreplace diopside-plagioclase symplectites formed at

higher temperatures than the magnesian pargasite andsodic plagioclase that replace garnet and form thegroundmass in highly amphibolitized samples. Fig-ure 10 shows hornblende-plagioclase temperatures as afunction of plagioclase composition at 1.2 GPa, whichis the mean pressure calculated for these assemblages.The high temperature hornblende-plagioclase assem-blage clearly replaces pre-existing diopside-plagioclasesymplectites, but formed at similar temperatures andpressures, probably in response to increased PH

2O

during the thermal peak; we speculate that this watermust have come from dehydration reactions in theenclosing felsic gneisses.

Using the data and methods described above, the

following equilibration conditions are suggested foreach stage of metamorphism, with each diagnosticassemblage shown in parentheses:(1) Eclogite facies metamorphism of a mafic protolith(garnet mantle + omphacite + rutile) at 650–730 Cand ‡ 1.4 GPa.(2) HP granulite facies conditions during decompres-sion (garnet rim + diopside + pargasite I + plagio-clase) at 700–800 C and 1.2–1.5 GPa.(3) Amphibolite facies conditions (pargasite II +plagioclase + ilmenite ± epidote) at 660 C to 720 Cand 0.9–1.2 GPa.

These data are summarized in Table 2 and Fig. 11,which depicts the calculated equilibria for variousmineral associations and facies with ellipses thatoverlie the intersection of geothermometer and geo-barometer equilibria for each stage of metamorphism.In the absence of primary plagioclase, the pressureestimated for the eclogite assemblage (P ¼ 1.4 GPa)is a minimum equilibration pressure; this may belowered somewhat based on the high Fe content of thesample, but may also be considerably higher (Carswell& Harley, 1990). The equilibration pressures of theHP granulite assemblage (1.2–1.5 GPa) are well

Fig. 8. Profile of single large garnet grain, from centre to rim.Distance from rim in lm. Note sharp increase in Grs (CaO) anddecrease in Alm (FeO) at around 1150 lm. Pyr (MgO) showsless precipitous increase, Sps (MnO) decays exponentiallytoward rim.

7 2 J . W . S H E R V A I S E T A L .


9/16

constrained since plagioclase and quartz are present inboth assemblages.

The conditions calculated for these assemblagesimply a clockwise pressure-temperature-time (P-T-t)path (Fig. 11). The clockwise P-T-t configuration isconsistent with models that involve collision of largecontinental or arc blocks, where one block is thrustbeneath another (high pressure at relative low tem-peratures) and then rebounds to an equilibriumgeothermal gradient when the block is exhumed duringuplift (England & Thompson, 1984). Based on thepreservation of primary zoning profiles in the garnet attemperatures up to 800 C, we suggest that uplift andcooling must have been relatively rapid after peakmetamorphic conditions were reached (O’Brien, 1997;Cooke et al ., 2000).

Protolith of mafic boudins and layers

Field relations suggest that protoliths of the maficboudins and layers were originally mafic dykes intru-ded into the more abundant felsic gneisses that com-pose the country rock of the Silverstreet domain.Whole rock analyses presented here (Table 3) andelsewhere (Dennis & Shervais, 1991, 1996) show thatthe felsic gneisses were derived from arc-related felsicto intermediate composition metavolcanic and meta-plutonic rocks of Charlotte belt affinity.

Most of the mafic boudins and dykes are basaltic incomposition, with SiO2 50%, MgO 6–8%,FeO* 10% and TiO2 1–3% (Fig. 12). Thesecompositions are typical of oceanic basalts, but are toohigh in TiO2 to represent arc-related high-alumina

0.5

1.5

2.5

3.5

Table 2. Summary of T-P data for high-P mafic lithology from Silverstreet domain. Numbers in brackets refer to references listedbelow.

Assemblage Facies Themometer Temperature Barometer Pressure

1 Eclogite Grt-Cpx 650–730 C @1.4 GPa [1,2] Grt-Cpx-Pl-Qtz

Jd-Ab-Qtz

> 1.3 GPa @700 C [6]

> 1.5 GPa @700 C [7,8]

> 1.4 Gpa @700 C [9]

2 High P Granulite Grt-Cpx 700–750 C @1.2–1.5 GPa [1,2] Grt-Cpx-Pl-Qtz 1.2 GPa @800 C [6,7]

Grt-Cpx 760–780

C @1.2–1.5 GPa [1,2]* 1.5 GPa @800

C [7,8]Cpx Solvus 700–800 C [3]

Grt-Hbl 770–820 C @1.2 GPa [4] Grt-Hbl-Pl-Qtz 1.2–1.5 GPa @800 C [10]

Hbl-Pl 760–830 C @1.2 GPa [5] GRIPS 1.2–1.5 GPa @800 C [11]

3 Amphibolite Grt-Hbl 660–775 C @1.2 GPa [4] Grt-Cpx-Pl-Qtz 1.0–1.2 GPa @700 C [10]

Hbl-Pl 690–740 C @1.2 GPa [5] GRIPS 1.0–1.2 GPa @675 C [11]

References [1] Ellis & Green (1979); [2] Powell (1985); [3] Lindsley & Anderson (1983); [4] Graham & Powell (1984); [5] Holland & Blundy (1994); [6] Newton & Perkins (1982); [7] Powell &

Holland (1988); [8] Moecher et al . (1988); [9] Holland (1980); [10] Kohn & Spear (1989); [11] Bohlen & Liotta (1986). All temperatures with Fe 3+ correction except *.

Fig. 9. Combined profiles of three garnet grains in sample NEW-1, scaled to percentradius of grain. Note sharp increase in Grs (CaO) and decrease in Alm (FeO) at around60% of total radius. Pyr (MgO) shows less precipitous increase, Sps (MnO) decaysexponentially toward rim.



10/16

basalts. Two samples are somewhat unusual ferro-basalts, with SiO2 41%, MgO 6%, FeO* 15–18%, and TiO2 2.5–3.5% (Fig. 12). They arechemically equivalent to evolved tholeiitic basalts, withvery low Cr and Ni (


11/16

rocks were preserved in thrust sheets during the Alle-ghanian collision of North America with Gondwana inthe late Palaeozoic (Stewart et al ., 1997).

Regional Implications

One conclusion that seems inescapable at this time isthat the Carolina terrane, as it is now understood, is acomposite terrane composed of two previously unre-lated arcs: the Charlotte belt arc and the Carolinaslate belt arc. Amalgamation of these two arcs toform the Carolina terrane must have occurred far

from Laurentia, and prior to accretion of the Caro-lina terrane to the Laurentian margin (Late Palaeo-zoic). Since the Silverstreet domain appears to be partof the Charlotte belt and may represent in part theinfrastructure of the Charlotte belt arc, subductionpolarity during amalgamation of the Charlotte beltarc with the Carolina slate belt arc must have been tothe SE (present day co-ordinates; Fig. 14). That is,the Charlotte belt arc formed part of the lower plateassemblage that was over-ridden by the Carolina slatebelt arc (which formed part of the upper plate of thesubduction zone).

Table 3. Whole rock major and trace element analyses by X-ray fluorescence.

Sample#

Rock

type

3080

Ferro-

basalt

NEW-2

Ferro-

basalt

3070–1

Basalt

3070–2

Basalt

3070–3

Basalt

3070–4

Basalt

3071UP

Amphib-

olite dyke

3071 A

Mafic

boudin

3071B

Mafic

boudin

3071C

Mafic

boudin

3071–4

Amphib-

olite dyke

3071–5

Amphib-

olite dyke

3071 FG

Felsic

Gneiss

3025 A

Felsic

Gneiss

3025B

Felsic

Gneiss

SiO2 41.7 41.5 50.1 49.8 49.1 50.4 51.0 49.3 50.4 49.3 50.3 48.0 54.9 60.6 60.0

TiO2 2.77 3.60 0.54 1.20 1.31 1.32 1.14 2.56 3.15 3.12 2.05 1.24 1.16 0.89 0.88

Al2O3 16.24 13.01 16.86 13.50 13.70 13.32 15.10 16.20 15.36 15.67 12.69 15.01 19.88 16.53 16.79

Fe2O3 16.70 20.36 7.38 12.11 11.71 13.12 10.69 11.66 12.70 12.37 14.23 11.66 9.93 8.37 8.54MnO 0.13 0.18 0.13 0.20 0.22 0.20 0.17 0.17 0.16 0.20 0.25 0.20 0.08 0.33 0.33

MgO 5.61 6.48 8.65 8.00 7.83 5.83 6.67 5.81 5.84 5.82 5.83 6.75 2.49 1.17 1.24

CaO 13.51 11.30 11.48 12.31 12.87 13.18 10.92 10.43 9.46 10.10 10.36 13.06 2.96 3.69 3.87

Na2O 2.09 2.10 2.72 2.57 2.20 2.03 2.91 1.70 1.78 1.57 2.21 1.99 4.98 4.16 4.29

K2O 0.176 0.209 0.606 0.39 0.398 0.132 1.091 0.956 0.495 0.697 1.228 0.707 2.948 3.006 2.903

P2O5 0.100 0.076 0.045 0.106 0.094 0.163 0.090 0.498 0.450 0.481 0.224 0.112 0.031 0.123 0.145

Total 99.03 98.84 98.47 100.18 99.43 99.71 99.79 99.25 99.82 99.36 99.40 98.76 99.31 98.84 98.94

p.p.m.

Nb 2.4 2.5 1.4 2.8 3.3 3.2 4.7 21.1 25.9 25.2 9.5 3.3 23 12.1 12.3

Zr 66 32 33 60 64 61 65 230 273 244 122 69 177 259 273

Y 17 14 14 23 23 20 22 37 42 37 37 28 13 59 63

Sr 937 100 180 59 116 85 210 548 367 436 133 274 255 422 418

Rb 5 4 19 7 6 5 15 47 13 33 20 15 160 55 53

Sc 50.1 52.6 40.3 38.3 44.2 41.7 36.1 35.1 39.2 37.6 39.2 51.2 10.2 8 9.4

V 666 734 171 309 324 341 293 363 438 397 406 325 172 70 82

Cr 30 43 259 288 283 76 125 160 185 186 65 268 123 10 15

Ni 8 18 158 78 63 69 65 67 49 55 48 112 79 8 10

Cu 46 107 107 96 90 44 62 31 9 30 138 148 121 41 29

Zn 117 124 47 88 87 93 86 90 98 89 112 99 116 81 85Ba 34 19 13 85 117 12 77 bdl bdll bdl 26 168 365 659 586

Note: Major elements in weight percentage oxide, trace elements in p.p.m.

(a) (b) (c)

(d) (e) (f)

Fig. 12. Harker diagrams for eclogite ⁄ high-P granulites of the Silverstreet domain, along with three felsic gneiss host rocks. Plots showSiO2 vs. (a) TiO2 (b) FeO* (c) MgO and CaO (d) Na2O and K2O (e) Cr and Ni p.p.m., and (e) Zr p.p.m.



12/16

Alternatively, Hibbard & Samson (1995) have sug-gested that collision between Carolina and the Gren-ville-aged Goochland terrane of eastern Virginia(Farrar, 1984) might be responsible for the meta-morphic fabric of the western Carolina terrane.Mueller et al . (1996) and Heatherington et al . (1996)have presented results based on their work in NorthCarolina that suggest a swing in Nd isotopic compo-sitions of Carolina lavas from strongly positive eNd tonear 0 or even negative at approximately the Cambrianboundary. A complex collision between a c. 1.1 Gacontinental fragment and some portion of the Carolinacomposite terrane could explain this observed swing inisotopic compositions and some of the xenocrysticzircon in Albemarle Group lavas from the NorthCarolina slate belt. Subsequent terrane dispersalduring postcollisional strike-slip faulting could beresponsible for the present disposition of high pressurerocks far from the present outcrop belt of theGoochland terrane. However, since the Charlotte belt-Silverstreet arc must have been in the lower plate of thecollision, the upper plate would have to include an arc

assemblage that was active at the time of collision; thisis not observed in the Goochland terrane.

The high pressure metamorphic event documentedhere must predate the c. 415 Ma Newberry granite(which is unaffected by the amphibolite facies over-print, and contains xenoliths of the sheared felsicgneisses which host the eclogites), and it must postdate

formation of the Carolina arc (c. 620 to 550 Ma;Dennis et al. 1997; Dennis & Wright, 1997; Samsonet al ., 1995; Wortman et al ., 1996, 2000; Barker et al .,1998). High-P metamorphism may coincide with the c.535 Ma collisional event documented in the westernCharlotte belt by Dennis & Wright (1997), or it mayrepresent an older, as yet unrecognized collisionalevent that occurred within the Carolina terrane priorto its amalgamation with Laurentia.

The occurrence of MORB-chemistry dykes in theroots of the arc implies that prior to its collision andpartial subduction beneath the Slate belt arc, theCharlotte belt-Silverstreet arc may have collided with

and over-ridden an active spreading centre. This isconsistent with the collision of two arcs that wereoriginally separated by a spreading centre, whichwould have to be consumed before collision of the twoarcs could occur (Fig. 14). It may also explain the hightemperatures and isobaric heating observed in theSilverstreet domain.

Uplift and Exhumation

High-pressure, MT granulites ⁄ eclogites of the Silver-street domain were exhumed and cooled extremelyrapidly, as shown by the preservation of progradezoning profiles in garnet that formed at 660–820 C

(e.g. O’Brien, 1997). These profiles could not persist if the boudins were held for long times at such hightemperatures. Exhumation of high-P metamorphicrocks involves two related problems: (1) a driving forcefor uplift of the crust, and (2) the dominant mechanismfor exhumation – erosion vs. low-angle normal faulting(Jamieson & Beaumont, 1989). In this context, upliftrefers to upward movement with respect to a fixeddatum, whereas exhumation refers to unroofing andmovement to lower lithostatic pressures (Jamieson &Beaumont, 1989; Ring et al ., 1999).

There are two possible models that may drive theuplift of deeply buried rocks in subduction zones orcollision zones, both related to the buoyancy of crustalrocks at depth: (1) the buoyancy of tectonically thickencrust when material is removed from the upper crustby erosion or faulting, or (2) break-off of the sub-ducting slab, thus removing slab pull and allowing thesubducting crust to return buoyantly to the surface(von Blanckenburg & Davies, 1995; Ernst et al ., 1997).We favour the slab breakoff model as the releasemechanism for uplift, because it is a logical conse-quence of the partial subduction of a buoyant blockattached to oceanic lithosphere, and because the twoarcs involved in this collision (Charlotte belt, Slate

0

5000

10000

15000

20000

25000

0 50 100 150 200 250 300

Ti ppm

Zr ppm

Calc-alkaline

ferro-basalts

MORB

IAT

felsic gneiss

0

100

200

300

400

500

600

700

800

0 5000 10000 15000 20000 25000

Ti ppm

V ppm 10

20

50

MORBARC

Felsic Gneiss

Ferrobasalts

Basalts

Fig. 13. Ti-V and Ti-Zr plots for eclogite ⁄ granulite blocks

from Silverstreet domain, showing MORB ⁄ OIB affinities of the mafic rocks.

7 6 J . W . S H E R V A I S E T A L .


13/16

belt) were likely too thin to generate great thicknessesof crust. In addition, it would allow isobaric heatingfollowed by sudden and rapid pressure release andcooling.

Exhumation, defined as the return of deeply buriedrocks to the surface, is generally driven either by ero-

sion, low-angle normal faulting, or both; ductileextension of the crust or lithosphere is slow and cannotaccount for more than a small fraction of exhumation(Ring et al ., 1999). Erosion and low-angle normalfaulting both operate at similar rates (5–13 km Myr) 1

for erosion, 5–10 km Myr)1 for normal faulting) andeither can account for the rapid cooling required topreserve the growth zoning in garnet. In the case of theSilverstreet domain, we suggest that low angle faultingwas the dominant process, because the fault contactbetween the Charlotte belt and Slate belt removes asignificant thickness of metamorphic section (c. 14 km)and juxtaposes terranes with significantly differentmetamorphic facies. Clearly, this process must havebeen aided by rapid erosion to remove material fromthe footwall, but erosion alone cannot account for the juxtaposition of terranes with distinctly differentmetamorphic grades and palaeodepths (Platt, 1986,1993).

CO NCLUSI O NS

Mafic meta-igneous rocks found along the boundarybetween the Charlotte belt and the Carolina Slate beltpreserve metamorphic phase assemblages that imply

progressive ⁄ retrogressive metamorphism under eclog-ite, high pressure granulite, and amphibolite faciesconditions. These rocks define a clockwise P-T-t path,consistent with collision and partial subduction. Wesuggest that this collision may have occurred duringamalgamation of the Charlotte belt to the Carolina

Slate belt to form the composite Carolina terrane thatwas later accreted to Laurentia.

Our conceptual model for this collision, based on thedata discussed above, is outlined in Fig. 14. In stage 1,the Charlotte belt arc and the Slate belt arc face oneanother across an active spreading centre. The Char-lotte belt arc over-rides this spreading centre in stage 2,leading to the emplacement of MORB compositiondykes in the dominantly felsic arc basement. By stage3, the Charlotte belt arc became extinct and attachedto oceanic crust still subducting beneath the Slate beltarc. Collision occurred during stage 4, with theCharlotte belt arc in the lower plate being partiallyover-ridden by the Slate belt arc, leading to the high-pressure metamorphism at eclogite and high-P granu-lite facies conditions. During stage 5 the subductingoceanic slab broke off, allowing rapid exhumationof the Charlotte belt arc basement, possibly along alow-angle normal fault. Finally (stage 6) subductionwas re-established beneath the combined Charlottebelt-Slate belt arc with renewed volcanism and plut-onism, and possible reactivation of the suture as ahigh-angle transcurrent structure. Further work isneeded to refine and test this model in other parts of the Carolina terrane.

Fig. 14. Model for collision and amalgamation of the Charlotte belt ⁄ Slate belt arcs. Stage 1: convergence of Charlotte belt and Slatebelt arcs, separated by mid-ocean ridge spreading centre; Stage 2: collision of Charlotte belt with MOR, emplacement of MORBcomposition dykes into infrastructure of the arc; Stage 3: continued convergence of Charlotte belt and Slate belt arcs, with detachmentand sinking of subducted lithosphere; Stage 4: collision of Charlotte belt and Slate belt arcs, with eclogite facies metamorphism of Charlotte belt arc infrastructure; Stage 5: delamination and sinking of oceanic lithosphere at leading edge of Charlotte belt arc,followed by rapid uplift and exhumation of Charlotte belt infrastructure; granulite and then amphibolite facies overprint duringexhumation; Stage 6: postcollisional Carolina terrane compositer arc; later arc volcanism and sedimentation may represent overlapassemblages that postdate suture; later reactivation of suture during Alleghanian sinistral shear?



14/16

ACK NO WLEDGEMENT S

The authors thank R. Jamieson and J. Gilloti for theirthoughtful and detailed reviews, A. Indares for edi-torial handling, and especially M.J. Kohn for discus-sions of thermobarometry issues. We also thankA. Bell and M. A. Jones (deceased) for assistance in the

field. This work was supported by NSF grantEAR9805159 to Dennis and Shervais.

R EFER ENCES

Adams, M. G., Stewart, K. G., Trupe, C. H. & Willard, R. A.,1995. Tectonic Significance of High-Pressure MetamorphicRocks and Dextral Strike Slip Faulting in the SouthernAppalachians. In: New Perspectives in the Appalachian-Caledonian Orogen, Geological Association of Canada Special Paper (eds. Hibbard, J., van Staal. C. R. Cawood, P. &Colman-Sid, S.), 41, 21–42.

Adams, M. G. & Trupe, C. H., 1997. Conditions and timing of metamorphism in the Blue Ridge thrust complex, northwest-ern North Carolina and eastern Tennesee. In: PaleozoicStructure, Metamorphism, and Tectonics of the Blue Ridge of

Western North Carolina, Field Trip Guide, (eds Stewart, K. G.Adams, M. G. & Trupe, C. H.), pp. 33–48. CarolinaGeological Society, Banner Elk.

Barker, C. A., Secor, D. T., Pray, J. R. & Wright, J. E., 1998.Age and deformation of the Longtown metagranite, SouthCarolina Piedmont: a possible constraint on the origin of theCarolina terrane. Journal of Geology, 106, 711–723.

von Blanckenburg, F. & Davies, J. H., 1995. Slab breakoff: amodel for syncollisional magmatism and tectonics in the Alps.Tectonics, 14, 120–131.

Bohlen, S. R. & Liotta, J. J., 1986. A barometer for garnetamphibolites and garnet granulites. Journal of Petrology, 27,1025–1034.

Carswell, D. A., 1990. Eclogites and the eclogite facies: defini-tions and classification. In: Eclogite Facies Rocks. (ed. D.A.Carswell), p. 1–13. Blackie, London.

Carswell, D. A. & Harley, S. L., 1990. Mineral barometry andthermometry. In: Eclogite Facies Rocks. (ed. D.A. Carswell),p. 83–110. Blackie, London.

Carswell, D. A. & O’Brien, P. J., 1993. Thermobarometry andgeotectonic significance of High-Pressure Granulites: Exam-ples from the Moldanubian zone of the Bohemian Massif inlower Austria. Journal of Petrology, 22, 85–126.

Cooke, R. A., O’Brien, P. J. & Carswell, D. A., 2000. Garnetzoning and the identification of equilibrium mineral composi-tion in high-pressure-temperature granulites from the Molda-nubian zone, Austria. Journal of Metamorphic Geology, 18,551–569.

Coward, M. P., Butler, R. W. H., Khan, M. Asif & Knipe, R. J.,1987. The tectonic history of Kohistan and its implications forHimalayan structure. Journal of the Geological Society of London, 144, 377–391.

Cuthbert, S. J. & Carswell, D. A., 1990. Formation and

exhumation of medium temperature eclogites in the Scandi-navian Caledonides. In: Eclogite Facies Rocks. (ed. D.A.Carswell), p. 180–203. Blackie, London.

Cuthbert, S. J., Carswell, D. A., Krough-Ravna, E. J. & Wain,A., 2000. Eclogites and eclogites in the Western GneissRegion, Norwegian Caledonides. Lithos, 52, 165–195.

Dal Piaz, G. V. & Lombardo, B., 1986. Early Alpine eclogitemetamorphism in the Penninic Monte Rosa-Gran Paradisobasement nappes of the northwestern Alps. In: Blueschists and Eclogites, Memoir 164, (eds B. W. Evans & G. H. Brown),pp. 249–265. Geological Society of America, Boulder.

Dallmeyer, R. D., Wright, J. E., Secor, D. T. & Snoke, A. W.,1986. Character of the Alleghanian orogeney in the southern

Appalachians. Part II. geochronological constraints on thetectonothermal evolution of the eastern Piedmont in SouthCarolina. Geological Society of America Bulletin, 97, 1329– 1344.

Dennis, A. J. & Shervais, J. W., 1991. Evidence for arc riftingalong the Carolina terrane boundary in northwestern SouthCarolina. Geology, 19, 226–229.

Dennis, A. J. & Shervais, J. W., 1996. The Carolina Terrane in

northwestern South Carolina: Insights into the developmentof an evolving island arc. In: Avalonian and Related Peri-Gondwanan Terranes of the Circum North Atlantic. Special Paper 304, (eds Nance, R. D. & M. D. Thompson), p. 237– 256. Geological Society of America, Boulder.

Dennis, A. J., Shervais, J. W. & Secor, D. T., 2000. Newberry,South Carolina Eclogite. Structural Setting and Style of Occurrence. In: A Compendium of Field Trips of SouthCarolina Geology, (eds C. Abate & A. B. Maybin), 29–38.South Carolina Geological Survey, Columbia

Dennis, A. J. & Wright, J. E., 1995. Mississippian (ca. 326–323Ma) U-Pb crystallization ages for two granitoids in Spartan-burg and Union Counties, South Carolina. In: Geology of theWestern Part of the Carolina Terrane, (ed. Dennis, A. J.) SouthCarolina Geology 38, 23–28.

Dennis, A. J. & Wright, J. E., 1997. The Carolina terrane innorthwestern South Carolina, USA: Late precambrian-Cam-

brian deformation and metamorphism in a peri-Gondwanaoceanic arc. Tectonics, 16, 460–473.

Ellis, D. J. & Green, D. H., 1979. An experimental study of theeffect of Ca upon garnet-clinopyroxene Fe-Mg exchangeequilibria. Contributions to Mineralogy and Petrology, 71,13–22.

Elvevold, S. & Gilotti, J. A., 2000. Pressure-temperature evolutionof retrogressed kyanite eclogites, Weinschenk Island, North-East Greenland Caledonides. Lithos, 53, 127–147.

England, P. C. & Thompson, A. B., 1984. Pressure-temperaturetime paths of regional metamorphism; I, Heat transfer duringthe evolution of regions of thickened continental crust. Journal of Petrology, 25, 894–928.

Ernst, W. G., Maruyama, S. & Wallis, S. R., 1997. Bouyancydriven rapid exhumation of ultra-high pressure metamor-phosed continental crust. Proceedings of the National Academyof Science, 94, 9532–9537.

Farrar, S. S., 1984. The Goochland granulite terrane: Remobi-lized Grenville basement in the eastern Virginia Piedmont.In: The Grenville Event in the Appalachians and Related Topics.Special Paper 194. (ed. Bartholomew, M. J.), pp. 215–227.Geological Society of America, Boulder.

Fullagar, P. D., 1981. Summary of Rb-Sr whole rock ages forSouth Carolina. South Carolina Geology, 25, 29–32.

Galan, G. & Marcos, A., 2000. The metamorphic evolution of the pressure mafic granulites of the Bacariza formation(Cabo Ortegal complex), Hercynian belt, NW Spain. Lithos,54, 139–171.

Graham, C. M. & Powell, R., 1984. A garnet-hornblendegeothermometer; calibration, testing, and application to thePelona Schist, Southern California. Journal of MetamorphicGeology, 2, 13–31.

Halik, R. S., 1983. Characterization of the Charlotte belt–

Carolina slate belt boundary in the Silverstreet and Dennyquadrangles, South Carolina: A stratigraphic, structural andmetamorphic study. MSc Thesis University of South Carolina,Columbia.

Hauck, M. L., 1984. Geologic characterization of the Charlottebelt in South Carolina. MSc Thesis University of SouthCarolina, Columbia.

Heatherington, A. L., Mueller, P. A. & Nutman, A. P., 1996.Neoproterozoic magmatism in the Suwannee terrane:Implications for terrane correlations. In: Avalonian and Related Peri-Gondwanan Terranes of the Circum-North Atlan-tic, Special Paper 304 (eds Nance, R. D. & Thompson, M.),pp. 219–236. Geological Society of America, Boulder.

7 8 J . W . S H E R V A I S E T A L .


15/16


16/16

southern Appalachians. Contributions to Mineralogy and Petrology, 121, 171–184.

Samson, S., Palmer, A. R., Robison, R. A. & Secor, D. T, 1990.Biogeographical significance of Cambrian trilobites from theCarolina slate belt. Geological Society of America Bulletin,102, 1459–1470.

Samson, S. D. & Secor, D. T., 2000. New U-Pb geochronologicalevidence for a Silurian magmatic event in central South

Carolina. Geological Society of America Abstracts withPrograms, 32/4.Secor, D. T., 1988. Geology of the eastern Piedmont in central

South Carolina. In: D. T. Secor, ed. Southeastern Geological Excursions Geological Society of America, SoutheasternSection, Field Trip Guidebook for 1988. 204–225.

Secor, D. T., Dennis, A. J., Maher, H. D. & Sacks, P. E., 1989.Geology and terrane analysis of the eastern Piedmont, SouthCarolina and Georgia. In: Dallmeyer, R. D., ed. Tectonos-tratigraphic Expression of Terrane Accretion in the SouthernAppalachian Orogen: a Geotraverse Excursion, the Universityof Georgia, I G C , p. 233. Guidebook Series, 8 ⁄ 25–8 ⁄ 91 and9 ⁄ 1–9 ⁄ 17.

Secor, D. T., Peck, L. S., Pitcher, D. M. et al., 1982. Geology of the area of induced seismic activity at Monticello Reservoir,South Carolina. Journal of Geophysical Research, 87, 6945– 6957.

Secor, D. T., Samson, S. L., Snoke, A. W. & Palmer, A. R.,1983. Confirmation of the Carolina slate belt as an exoticterrane. Science, 221, 649–650.

Secor, D. T., Snoke, A. W., Bramlett, K. W., Costello, O. P. &Kimbrell, O. P., 1986. Character of the Alleghanian orogeneyin the southern Appalachians. Part I. Alleghanian deforma-tion in the eastern Piedmont of South Carolina. Geological Society of America Bulletin, 97, 1314–1328.

Shervais, J. W., 1982. Ti-V plots and the petrogenesis of modernand ophiolitic lavas. Earth and Planetary Science Letters, 59,101–118.

Shervais, J. W., Shelley, S. A. & Secor, D. T., 1996.Geochemistry of volcanic rocks of the Carolina and Augustaterranes in central South Carolina: An exotic rifted volcanicarc?. In: Avalonian and Related Per-Gondwanan Terranes of theCircum-North Atlantic, Special Paper 304. (eds Nance, R. D.& M. D. Thompson), pp. 237–256. Geological Society of America, Boulder.

Spear, F. S., 1993. Metamorphic Phase Equilibria and Pressure-Temperature Time Paths. Monograph 1, Mineralogical Societyof America, Washington DC.

Spear, F. S. & Kohn, M. A., 2001. GTB: Program Geothermo-barometry, Version 2.1. February 2001. computer program.http://ees2.geo.rpi.edu/MetaPetaRen/Frame_software.html.

van Staal, C. R., Dewey, J. F., MacNiocaill, C. & McKerrow,W. S., 1998. The Cambrian-Silurian tectonic evolution of thenorthern Appalachians and British Caledonides: history of acomplex, west and southwest Pacific-type segment of Iapetus.In: Lyell: the Past Is the Key to the Present. Special

Publication, 143, (eds Blundell, D. J. & Scott, A. C.),pp. 199–242. Geological Society, London.Stewart, K. G., Adams, M. G. & Trupe, C. H., 1997.

Paleozoic structural evolution of the Blue Ridge thrustcomplex, western North Carolina. In: Paleozoic Structure,Metamorphism, and Tectonics of the Blue Ridge of WesternNorth Carolina, Field Trip Guide, (eds Stewart, K. G., Adams,M. G. & Trupe, C. H.), pp. 21–31. Carolina GeologicalSociety, Banner Elk.

West, T. E., 1998. Structural analysis of the Carolina–InnerPiedmont terrane boundary: Implication for the age andkinematics of the central Piedmont suture, a terrane boundarythat records Paleozoic Laurentia–Gondwana interactions.Tectonics, 17, 379–394.

Will, T. M. & Schma ¨ dicke, E., 2001. A first find of retrogressedeclogites in the Odenwald crystalline complex, mid-GermanCrystalline Rise, Germany: evidence for a so far unrecognized

high-pressure metamorphism in the central Variscides. Lithos,59, 109–125.

Willard, R. A. & Adams, M. G., 1994. Newly discoveredeclogite in the southern Appalachian orogen, northwesternNorth Carolina. Earth and Planetary Science Letters, 123,61–70.

Williams, H. & Hatcher, R. D. 1983. Appalachian suspectterranes. In: Contributions to the Tectonics and Geophysics of Mountain Chains. Memoir 158, (eds Hatcher, R. D., Williams,H. & Zietz, I.), pp. 33–53. Geological Society of America,Boulder.

Wortman, G. L., Samson, S. D. & Hibbard, J. P., 1996.Discrimination of the Milton belt and Carolina terrane in thesouthern Appalachians: a Nd isotopic approach. Journal of Geology, 104, 239–247.

Wortman, G. L., Samson, S. D. & Hibbard, J. P., 2000. PreciseU-Pb zircon constraints on the earliest magmatic history of the Carolina terrane. Journal of Geology, 108.

Received 26 October 2001; revision accepted 8 July 2002.

AP P ENDI X : MET HO DS

All minerals were analyzed on the Cameca SX-50 electron micro-probe at the University of South Carolina using natural and syn-thetic mineral standards from the Smithsonian Institution;operating conditions were typically 20 KV at 25 nA. Data reductionwas carried out using the Cameca implementation of the phi-rho-zalgorithm (Pouchou & Pichoir, 1991). Large area X-ray composi-tion mapping was carried out in stage mode using four fixedwavelength spectrometers in conjunction with the backscattered

electron (BSE) diodes. Typical maps are 512 · 512 pixels with a stepsize 10 lm anddwell time of 200 milliseconds, for a total area scannedof 5.0 · 5.0 mm and an analysis time of about 14 h per map.

Omphacite was reconstructed from diopside-plagioclase symplec-tite by acquiring a series of BSE images on different diopside grains,which were then segmented in the Cameca image analysis software torecover the proportion of plagioclase in diopside. Results for allgrains imaged were around 20% modal plagioclase. The inferredomphacite composition was reconstructed by converting the averagecomposition of plagioclase in symplectite to a mineral formulaand subtracting one mole of SiO2 to create jadeite. The jadeiteformula was converted back to weight% oxide and mixed with theaverage composition of diopside in 20 : 80 proportions to create

omphacite.Calculation of Fe3+/Fe2+ by stoichiometry for thermobarometry

followed method of Spear (1993).

8 0 J . W . S H E R V A I S E T A L .

Documents

JMG_ShervDenn2003