Phengite-hosted LILE Enrichment in Eclogite and Related Rocks: Implications for Fluid-Mediated Mass...

32
JOURNAL OF PETROLOGY VOLUME 38 NUMBER 1 PAGES 3–34 1997 Phengite-hosted LILE Enrichment in Eclogite and Related Rocks: Implications for Fluid-Mediated Mass Transfer in Subduction Zones and Arc Magma Genesis SORENA S. SORENSEN 1 *, JEFFREY N. GROSSMAN 2 AND MICHAEL R. PERFIT 3 1 DEPARTMENT OF MINERAL SCIENCES, NHB-119, NATIONAL MUSEUM OF NATURAL HISTORY, WASHINGTON, DC 20560, USA 2 US GEOLOGICAL SURVEY, MS-923, RESTON, VA 22094, USA 3 DEPARTMENT OF GEOLOGY, UNIVERSITY OF FLORIDA, GAINESVILLE, FL 32611, USA RECEIVED ON MARCH 6, 1996 REVISED TYPESCRIPT ACCEPTED ON JULY 29, 1996 Geochemical dierences between island arc basalts (IAB) and ocean- wt % (0·027 per 11 oxygens). Ba in phengite does not covary strongly with either Na or K. Ba contents of phengite increase from floor basalts (mid-ocean ridge basalts; MORB) suggest that the large-ion lithophile elements (LILE) K, Ba, Rb and Cs are probably some blocks to their transition zones or rinds, or from blocks to mobilized in subduction zone fluids and melts. This study documents their veins. Averaged K/Ba ratios for phengite and host samples LILE enrichment of eclogite, amphibolite, and epidote ± garnet define an array which describes other subsamples of the block and blueschist tectonic blocks and related rocks from melanges of two other analyzed blocks. Phengite carries essentially all of the LILE subduction complexes. The samples are from six localities of the in otherwise mafic eclogite, amphibolite, and garnet blueschist blocks Franciscan Complex, California, and related terranes of Oregon that are enriched in these elements compared with MORB. It and Baja California, and from the Samana Metamorphic Complex, evidently tracks a distinctive type of LILE metasomatism that Samana Peninsula, Dominican Republic. Most Franciscan blocks attends both high-T and retrograde subduction zone metamorphism. are MORB-like in their contents of rare earth elements (REE) and An obvious source for the LILE is a fluid in equilibrium with high field strength elements (HFSE); in contrast, most Samana metasedimentary rocks. High-grade semipelitic schists from sub- blocks show an IAB signature of these elements. The whole-rock duction complexes and subductable sediment display LILE values K 2 O contents of both groups range from 1 to 3 wt %; K, Ba, Rb, that resemble those seen in the most LILE-rich blocks. Modeling and Cs are all strongly intercorrelated. Many blocks display K/Ba of Ba and Ti suggests that 1–40 wt % of phengite added to similar to metasomatized transition zones and rinds at their outer MORB can produce their observed LILE enrichment. Thus, the margins. Some transition zones and rinds are enriched in LILE release of LILE from such rocks to fluids or melts in very high-T compared with host blocks; others are relatively depleted in these and -P parts of subduction zones probably depends critically on elements. Some LILE-rich blocks contain ‘early’ coarse-grained the stability and solubility relations of phengite, which is thought muscovite that is aligned in the foliation defined by coarse-grained to be stable at pressures as high as 95–110 kbar at T= omphacite or amphibole grains. Others display ‘late’ muscovite in 750–1050°C. veins and as a partial replacement of garnet; many contain both textural types. The muscovite is phengite that contains ~3·25–3·55 Si per 11 oxygens, and ~0·25–0·50 Mg per 11 oxygens. Lower- Si phengite has a significant paragonite component: Na per 11 KEY WORDS: geochemistry; LT eclogite; mineral chemistry; metasomatism; phengite oxygens ranges to ~0·12. Ba contents of phengite range to over 1 *Corresponding author. Oxford University Press 1997 by guest on June 28, 2015 http://petrology.oxfordjournals.org/ Downloaded from

Transcript of Phengite-hosted LILE Enrichment in Eclogite and Related Rocks: Implications for Fluid-Mediated Mass...

JOURNAL OF PETROLOGY VOLUME 38 NUMBER 1 PAGES 3–34 1997

Phengite-hosted LILE Enrichment inEclogite and Related Rocks: Implications forFluid-Mediated Mass Transfer in SubductionZones and Arc Magma Genesis

SORENA S. SORENSEN1∗, JEFFREY N. GROSSMAN2 ANDMICHAEL R. PERFIT3

1DEPARTMENT OF MINERAL SCIENCES, NHB-119, NATIONAL MUSEUM OF NATURAL HISTORY, WASHINGTON,

DC 20560, USA2US GEOLOGICAL SURVEY, MS-923, RESTON, VA 22094, USA3DEPARTMENT OF GEOLOGY, UNIVERSITY OF FLORIDA, GAINESVILLE, FL 32611, USA

RECEIVED ON MARCH 6, 1996 REVISED TYPESCRIPT ACCEPTED ON JULY 29, 1996

Geochemical differences between island arc basalts (IAB) and ocean- wt % (0·027 per 11 oxygens). Ba in phengite does not covary

strongly with either Na or K. Ba contents of phengite increase fromfloor basalts (mid-ocean ridge basalts; MORB) suggest that the

large-ion lithophile elements (LILE) K, Ba, Rb and Cs are probably some blocks to their transition zones or rinds, or from blocks to

mobilized in subduction zone fluids and melts. This study documents their veins. Averaged K/Ba ratios for phengite and host samples

LILE enrichment of eclogite, amphibolite, and epidote ± garnet define an array which describes other subsamples of the block and

blueschist tectonic blocks and related rocks from melanges of two other analyzed blocks. Phengite carries essentially all of the LILEsubduction complexes. The samples are from six localities of the in otherwise mafic eclogite, amphibolite, and garnet blueschist blocksFranciscan Complex, California, and related terranes of Oregon that are enriched in these elements compared with MORB. Itand Baja California, and from the Samana Metamorphic Complex, evidently tracks a distinctive type of LILE metasomatism thatSamana Peninsula, Dominican Republic. Most Franciscan blocks attends both high-T and retrograde subduction zone metamorphism.are MORB-like in their contents of rare earth elements (REE) and An obvious source for the LILE is a fluid in equilibrium withhigh field strength elements (HFSE); in contrast, most Samana metasedimentary rocks. High-grade semipelitic schists from sub-blocks show an IAB signature of these elements. The whole-rock duction complexes and subductable sediment display LILE valuesK2O contents of both groups range from 1 to 3 wt %; K, Ba, Rb, that resemble those seen in the most LILE-rich blocks. Modelingand Cs are all strongly intercorrelated. Many blocks display K/Ba of Ba and Ti suggests that 1–40 wt % of phengite added tosimilar to metasomatized transition zones and rinds at their outer MORB can produce their observed LILE enrichment. Thus, themargins. Some transition zones and rinds are enriched in LILE release of LILE from such rocks to fluids or melts in very high-Tcompared with host blocks; others are relatively depleted in these and -P parts of subduction zones probably depends critically onelements. Some LILE-rich blocks contain ‘early’ coarse-grained the stability and solubility relations of phengite, which is thoughtmuscovite that is aligned in the foliation defined by coarse-grained to be stable at pressures as high as 95–110 kbar at T=omphacite or amphibole grains. Others display ‘late’ muscovite in 750–1050°C.veins and as a partial replacement of garnet; many contain both

textural types. The muscovite is phengite that contains ~3·25–3·55

Si per 11 oxygens, and ~0·25–0·50 Mg per 11 oxygens. Lower-

Si phengite has a significant paragonite component: Na per 11 KEY WORDS: geochemistry; LT eclogite; mineral chemistry; metasomatism;

phengiteoxygens ranges to ~0·12. Ba contents of phengite range to over 1

∗Corresponding author. Oxford University Press 1997

by guest on June 28, 2015http://petrology.oxfordjournals.org/

Dow

nloaded from

JOURNAL OF PETROLOGY VOLUME 38 NUMBER 1 JANUARY 1997

conditions during subduction zone metamorphism (Do-INTRODUCTIONmanik et al., 1993; compare Moran et al., 1992). Fur-Potassium, Ba, Rb, and Cs typically are enriched inthermore, integrated trace element and isotopic studiesisland arc basalts (IAB), and their differentiates, comparedof Be and B indicate that these elements, which track awith mid-ocean ridge basalts (MORB) (e.g. Kay, 1980;component of subducted sediment, are transferred fromPerfit et al., 1980a; Gill, 1981). Part of this LILE en-subduction zones to source regions of arc magmas onrichment is probably due to an added component derived relatively short time scales (e.g. Tera et al., 1986; Ryan

from subducted sediment or altered portions of the slab; & Langmuir, 1988; Morris et al., 1990; Bebout et al.,therefore, some LILE must be transferred from the 1993).subducted slab to the mantle wedge that overlies the This work integrates whole-rock geochemistry with thesubduction zone at depth (e.g. Gill, 1981; Hawkesworth mineral chemistry of phengite for low-T [LT: clas-et al., 1993). Some workers conclude that melts are the sification of Carswell (1990)] eclogite and related rocksprimary media of mass transfer from the slab to the from two subduction zone metamorphic terranes, andmantle wedge (e.g. Kay, 1980; Kelemen et al., 1990; inteprets the systematics of K, Ba, Rb, and Cs duringSchiano et al., 1995), but others think that fluids may fluid–rock interaction that attended metamorphism overbe the pre-eminent agents (e.g. Tatsumi, 1986, 1989). a range of P–T conditions. Rather paradoxically, LTHowever, most devolatilization of the altered slab and eclogite is a relatively high-T (~450–700°C) metamorphicsubducted sediment occurs at shallower depths than the rock type in subduction complexes (e.g. Schliestedt, 1990).source regions of arc magmas (e.g. Anderson et al., 1976; Although LT eclogite is but a small constituent of sub-Delany & Helgeson, 1978; Bebout, 1991; Moran et al., duction zone metamorphic terranes such as the Fran-1992). Therefore, mechanisms such as corner flow in the ciscan Complex of coastal California (Bailey et al., 1964),mantle wedge are called upon to bring the ‘subducted phase equilibrium and thermal modeling results suggestcomponent’ to the source region of arc magmas (e.g. it is a major constituent of steady-state subducted slabsTatsumi, 1986, 1989; Arculus & Powell, 1986). with little shear heating (e.g. Peacock, 1993).

Rocks that represent source materials of the majority We describe a distinctive style of K–Ba–Rb–Cs al-of arc magmas are only rarely entrained as xenoliths in teration that appears to be acquired during subductionarc volcanic rocks (e.g. Schiano et al., 1995; Ertan & zone metamorphism of LT eclogite, garnet blueschistLeeman, 1996), and they are not exposed in subduction and epidote ± garnet amphibolite facies rocks fromcomplexes. However, high P/T metamorphic terranes Franciscan and related subduction complexes on the westpreserve evidence for subduction-related fluid–rock inter- coast of the USA and Mexico, and from the Samanaaction. Fluid-mediated mass transfer at model depths Peninsula, Dominican Republic. The alteration is man-of 20–35 km within the Catalina Schist, a Cretaceous ifested by phengite, a Si- and Mg-rich muscovite, in thesubduction complex in southern California, is docu- metasomatized rocks, and probably reflects con-mented by trace element and isotopic studies of both tamination by fluids derived from subducted sediment.relatively high-T (>450°C) and low-T (<450°C) sub- We propose that the Ba content of phengite can beduction zone metamorphic rocks (e.g. Bebout & Barton, used as a geochemical tracer of the aforementioned1989, 1993; Sorensen & Grossman, 1989; Bebout et al., K–Ba–Rb–Cs alteration, which appears to accompany1993). Knowing how K, Ba, Rb, and Cs are mobilized both ‘high-grade’ and retrograde metamorphism.under blueschist to eclogite facies conditions (T Because K, Ba, Rb, and Cs reside primarily in phengite~300–700°C, P ~8–15 kbar), can help constrain how in these metamafic rocks, its stability can conceivablythese elements may be transferred from the slab to the exert an important control upon mass transfer of alkalismantle wedge. from the slab to the mantle wedge. An LILE ‘sediment

Alkali and alkaline earth elements commonly are mo- signature’ stored in phengite by subduction zonemetasomatism could be effectively decoupled from otherbilized by fluids in many metamorphic environmentsgeochemical tracers of sediment contamination, notably(e.g. Rose & Burt, 1979; Krogh & Brunfelt, 1981; Alt etSr and Pb.al., 1986; Glazner, 1988; Roddy et al., 1988; Bednarz &

Schmincke, 1989; Barton et al., 1991). Several studieshave described K–Ba–Rb–Cs alteration that ac-companies lawsonite–blueschist facies subduction zone

GEOLOGIC BACKGROUNDmetamorphism (Platt et al., 1976; Moore & Liou, 1979;LT eclogite, epidote ± garnet amphiboliteMoore et al., 1981; Bebout & Barton, 1993; Tenoreand garnet blueschist blocksNortrup & Bebout, 1993). Ion-probe data for minerals

from subduction zone metamorphic rocks indicate that LT eclogite, epidote ± garnet amphibolite, and garnetK, Ba, Rb, and Cs are probably mobilized in fluids along blueschist occur as isolated blocks in matrices of either

metasedimentary rocks (most commonly meta-argillitewith hydrophilic elements such as B under various P–T

4

by guest on June 28, 2015http://petrology.oxfordjournals.org/

Dow

nloaded from

SORENSEN et al. PHENGITE-HOSTED LILE ENRICHMENT

or semipelitic schist) or of meta-ultramafic rocks (most Jurassic to Early Cretaceous subduction complex in Bajacommonly serpentinite or talc schist) in melanges of California (Fig. 1a). Most of the Central Melange Beltcircumpacific and Caribbean subduction complexes consists of metagraywacke and metashale (e.g. Bailey et

(Bailey et al., 1964; Morgan, 1970; Ernst et al., 1970; al., 1964; Cloos, 1983, 1986). The metashale matrixColeman & Lanphere, 1971; Hermes, 1973; Platt, 1975; of the melange was metamorphosed at relatively highFeininger, 1980; Maresch & Abraham, 1980; Moore, pressures and low temperatures (Cloos, 1983; Dalla Torre1986; Draper & Nagle, 1988; Sedlock, 1988; Ross & et al., 1996). Franciscan high-grade blocks have beenSharp, 1988; Baldwin & Harrison, 1989, 1992; Bocchio subjects of extensive petrologic study (Holway, 1904;et al., 1990). Collectively, these rock types are colloquially Switzer, 1945; Borg, 1956; Coleman et al., 1965; Colemanreferred to as ‘high-grade blocks’. Some workers conclude & Lanphere, 1971; Ghent & Coleman, 1973; Brown &that high-grade blocks are remnants of high-T meta- Bradshaw, 1979; Coleman, 1980; Moore, 1984; Moore,morphic aureoles that formed during the early stages of 1986; Sedlock, 1988; Oh & Liou, 1990; Baldwin &subduction (e.g. Platt, 1975; Cloos, 1984, 1985). Others Harrison, 1992; Krogh et al., 1994). They occur primarilyinterpret them as reworked basement that predates the within mud-matrix melange, or as ‘float’. In many loc-initiation of subduction and genesis of the subduction alities of the Franciscan Complex, outcrops in the im-complex (e.g. Coleman & Lanphere, 1971; Moore, 1984; mediate area of float blocks consist of metasedimentaryKunugiza et al., 1986; Sharp & Ross, 1987). rocks (Crawford, 1965; Moore & Blake, 1989). Although

Many high-grade blocks are partially encased in a Cloos (1986) reported that many high-grade blocks ofconcentrically foliated selvage, or ‘rind’, of schistose meta- the Central Melange Belt of the Franciscan Complex areultramafic rock that contains chlorite ± actinolite ± found in a matrix of fine-grained, argillaceous meta-talc ± white mica ± other amphiboles, titanite, rutile, sedimentary rock (metashale), he noted that most alsoREE-epidote, and zircon (Coleman, 1980; Moore, 1984; bear Mg-rich rinds, which indicate a former associationSorensen, 1988; Sorensen & Grossman, 1989, 1993). with meta-ultramafic rock. At Ring Mountain on theThis rind forms by aqueous-fluid-induced metasomatic Tiburon Peninsula, Rice et al. (1976) mapped the high-reaction between the block and ultramafic rocks during grade blocks in a shale-matrix melange unit, which issubduction zone metamorphism (Coleman, 1980; Moore, intercalated with serpentinite. On the Vizcaino Peninsula,1984; Cloos, 1986; Sorensen, 1988). An Mg-rich rind is Cedros Island and the San Benito Islands of Baja Cali-evidence that its host block was in contact with ultramafic fornia, high-grade blocks are found in serpentinite-matrixrocks during some stage of its metamorphic history, even melanges (Moore, 1986; Sedlock, 1988).if the block is not found in a matrix of ultramafic rocks. Samples of 13 high-grade blocks from six localities of

High-grade blocks typically show retrograde high-P the Franciscan Complex and related units were studiedmetamorphic effects such as partial replacement of garnet (Fig. 1a). These are (from north to south): (1) two blocksby lawsonite and chlorite, rimming of barroisitic am- from a unit in southwest Oregon that is correlative withphibole by glaucophane–crossite, and rimming of rutile the Franciscan Complex (ORE-1, ORE-3: Moore &by titanite (e.g. Coleman & Lee, 1963). The presence Blake, 1989); (2) the ‘Junction Schoolhouse’ eclogite ofof a lawsonite–blueschist facies retrograde assemblage the Central Melange Belt at Healdsburg, California (GL-indicates that a high-grade block partially recrystallized 14: Moore & Blake, 1989); (3) a block located near Dosunder hydrous, lower-T (<400–450°C) subduction zone Rios, California (DR: Nelson, 1991, 1995); (4) five blocksmetamorphic conditions during exhumation (Ernst, from Ring Mountain, Tiburon Peninsula, California (T-1988). Some LT eclogite blocks contain cavities that are 90-1, T-90-2, T-90-3, T-90-5 and T-90-6: Ransome,lined with euhedral crystals of sodic clinopyroxene, sodic 1895; Tallifero, 1943; Dudley, 1972; Rice et al., 1976;amphibole, aragonite, and titanite, which indicate high Ingersoll et al., 1984; Oh, 1990); (5) a block on the roadP H2O during subduction zone metamorphism (Cloos, to Mt Hamilton, Diablo Range, California (GL-16 and1986). The following sections provide a brief outline of MH-90: Cloos, 1986; Moore & Blake, 1989; Nelson,the geologic context of the high-grade blocks analyzed 1991, 1995; Giaramita & Sorensen, 1994); (6) blocks offor this study, and describe the samples. epidote amphibolite, epidote blueschist and pumpellyite-

bearing blueschist (the last is not a high-grade assemblage)from the Jurassic–Early Cretaceous subduction complexSetting and samples from the Franciscan

Complex, California, USA, and related exposed on Cedros Island and East San Benito Island,terranes of Oregon, USA, and Baja Baja California, Mexico (RF, 687186, 68718: Sedlock,California, Mexico 1988; Baldwin & Harrison, 1989, 1992). Locality (5) was

bulldozed in the summer of 1990 during road work; theHigh-grade blocks of blueschist, amphibolite, and eclogiteblock existed then only as partly buried, tumbled slabs.are found in the Central Melange Belt of the Franciscan

Sets of samples from six individual blocks and theirComplex, California Coast Ranges, in Franciscan-cor-relative rock units in southwestern Oregon, and in a rinds are represented in the ‘Franciscan’ sample group.

5

by guest on June 28, 2015http://petrology.oxfordjournals.org/

Dow

nloaded from

JOURNAL OF PETROLOGY VOLUME 38 NUMBER 1 JANUARY 1997

Fig. 1.

These are: ORE-1, ORE-3, GL-14, T-90-1, T-90-2, and the Fe2+–Mg, garnet–clinopyroxene geothermometer(Ellis & Green, 1979). The results are (T, P min): (a) forGL-16/MH-90 (Tables 1 and 2). Other subsamplingT-90-1B, 691 ± 84°C, 10·4 ± 2·1 kbar; (b) for T-90-was guided by heterogeneity within individual blocks.5, 593 ± 63°C, 9·0 ± 1·3 kbar; (c) for MH-90-1,Accordingly, two blocks that consist of centimeter-scale671 ± 77°C, 10·5 ± 1·5 kbar. Corrections to theinterlayers of eclogite and garnet blueschist (DR and T-garnet–clinopyroxene geothermometer proposed by90-3), and two other blocks that consist of interlayeredKrogh (1988) or by Pattison & Newton (1989) lower theeclogite and amphibolite (T-90-1 and MH-90) wereT estimates by ~50°C.subsampled by layer. Samples that were not collected

by the first author were obtained from the followingresearchers: ORE and GL samples, Dr Diane E. Moore,

Setting and samples from the SamanaUS Geological Survey; RF, 68718, 687186 samples, DrMetamorphic Complex, Samana Peninsula,Suzanne Baldwin, University of Arizona; DR sample, DrDominican RepublicMark Cloos, University of Texas [for this sample, com-

pare Domanik et al. (1993)]. Blocks of eclogite and high-grade blueschist are found inGiaramita & Sorensen (1994) estimated P and T for melanges of Cuba and the north coast of the Dominican

eclogite samples T-90-1B, T-90-5 and MH-90-1, using Republic (Nagle, 1974; Perfit & McCulloch, 1982; Perfitet al., 1982; Joyce, 1985, 1991; Draper & Lewis, 1991).the jadeite content of clinopyroxene (Ghent, 1989) and

6

by guest on June 28, 2015http://petrology.oxfordjournals.org/

Dow

nloaded from

SORENSEN et al. PHENGITE-HOSTED LILE ENRICHMENT

Fig. 1. Index maps of California and Mexico (a) and the Samana Peninsula, Dominican Republic (b). CA, California; NV, Nevada; OR,Oregon; BC, Baja California; BCS, Baja California Sur; CB, Coos Bay, OR; E, Eureka, CA; SF, San Francisco, CA; M, Monterey, CA; SAF,

San Andreas Fault. Other initials designate sample localities.

Lower-grade blueschist has been dredged from the inner conditions. The eclogite blocks contain epidote and sodicamphibole. Except for a few sodic amphibole veins, thewalls of the Puerto Rico trench (Perfit et al., 1980b). On

the Samana Peninsula of the Dominican Republic, high- textural relations of the epidote blueschist facies mineralsare not obviously retrograde, and no indications of ret-grade blocks occur in a zone [the Punta Balandra Zone

of Joyce (1991)] within the Santa Barbara Schist, which rograde greenschist facies metamorphism of eclogite arepresent (compare Joyce, 1991). Mean temperatures foris one of three units of the Samana Metamorphic Com-

plex of Joyce (1991). Here, high-grade blocks have been the eclogite samples, estimated using the Fe2+–Mg gar-net–clinopyroxene geothermometer and the jadeite con-emplaced into a section dominated by metacarbonate

rocks. Most blocks with diameters >0·5 m are found in tents of clinopyroxene, are in the range T= 502–601°Cat P min = 8·2–9·9 kbar (Ellis & Green, 1979; Ghent,lenses of talc ± Mg-chlorite schist that locally contain

fuchsite, which were presumably derived from ultramafic 1989; Giaramita & Sorensen, 1994). The P–T estimatesfor eclogite, the mineral assemblages and textures ofrock. These lenses range to tens of meters wide. They

locally crosscut the regionally developed foliation of sur- the metasedimentary host rocks, and structural ob-servations that indicate relatively late emplacement ofrounding marbles, micaceous marbles, calcareous schists

and semipelitic schists (compare Joyce, 1991). the meta-ultramafic lenses into host rocks suggest thatmetamorphism of the high-grade blocks did not occur inThe mineral assemblage of the semipelitic schist in-

dicates regional, greenschist facies P–T conditions: it situ at the regional P–T conditions (compare Joyce,1991).consists of actinolite, albite, quartz, clinozoisite, chlorite,

white mica, calcite, and titanite. Inclusions of lawsonite Eclogite blocks were sampled east of the town ofSamana near Punta Balandra, on the south coast of theare present only in albite and clinozoisite porphyroblasts

of semipelitic schist (compare Joyce, 1991). This texture Samana Peninsula (Fig. 1b). Eclogite occurs as: (1) bou-dins Ζ0·5 m in their longest dimension in a singlesuggests that earlier, high P/T, metamorphism in the

metasedimentary host rocks of the eclogite-bearing outcrop of coarse-grained calcite marble along the coastroad near Punta Balandra, (2) inclusions 0·5–3 m inultramafic rocks was overprinted by greenschist facies

7

by guest on June 28, 2015http://petrology.oxfordjournals.org/

Dow

nloaded from

JOURNAL OF PETROLOGY VOLUME 38 NUMBER 1 JANUARY 1997

Tab

le1:

Min

eral

asse

mbl

ages

ofhi

gh-g

rade

bloc

ks,

tran

sition

zone

s,an

dri

nds

Sam

ple

grt

chl

epi

law

alb

amp

qtz

cpx

tte

rut

mu

sst

pm

gt

hm

tzi

rap

tp

yto

thco

mm

ents

Blo

ck

sa

nd

alt

ere

db

lock

s

Sam

ana,

D.R

.

SS

85-2

2x

rsr

—r

na

—x

rx

sr—

?r

x—

?—

ree

afte

rg

rt

SS

84-2

4A1

x—

cz&

zo—

—n

c&ca

xx

x—

——

?r

——

?—

SS

84-2

4A2

xr

cz&

zo—

—ca

xx

x—

—sr

——

x—

——

SS

84-2

4Bx

rx

——

na

xx

xcs

x—

?r

xx

?—

ree

epi

core

s

84-2

4/85

GR

EE

Nx

rzo+

r—

plg

ca—

xx

—r

—?

rx

—?

rcal

84-2

4/85

WH

ITE

xr

zo+

r—

plg

ca—

—x

iep

ir

——

——

——

rcal

SS

85-2

7Dx

rcz+

r—

—n

c&ca

igt

—sr

xx

sr—

—x

——

——

SS

85-2

7Ex

rre

e+r

——

ncz

tna

xx

srx

sr—

?r

——

?—

iree

nc

R75

0x

——

——

—ig

tx

srcs

—r

——

——

——

icp

xin

tte

R54

8x

r—

——

na

—x

srcs

r—

——

——

—rc

alic

px

intt

e

R75

1x

—zo

——

na&

cax

xx

xx

——

——

——

——

SA

M-0

9x

rx

—r

na

——

srcs

sr—

—r

——

xrc

al—

SA

M-1

2Ex

rx

——

na

xx

srcs

sr—

—r

——

——

Fran

cisc

anC

mpx

.

OR

E-1

-10

xr,

vsr

——

na

—sr

srx

x—

xr

x—

?—

OR

E-3

-4—

rx

—sr

na

xr

x—

sr—

—r

x—

——

DR

x—

srr

—n

aig

tx

srcs

srr

—r

x—

xtu

rit

ur

ing

rt

DR

-EC

L1x

—x

v—

na

—x

xig

tsr

rgt

—x

——

——

law

sve

in

DR

-EC

L2x

—x

r—

na

igt

xsr

cssr

rgt

—r

——

x—

DR

-EC

L3x

—x

r+v

—n

a—

xx

igt

sr—

—r

xx

x—

law

smu

sap

tv

DR

-BS

T1

x—

srr

rn

aig

tx

srcs

srr

—r

x—

x—

DR

-BS

T2

x—

xr

—n

aig

tx

srcs

srr

—r

—x

x—

GL-

14-2

xr

r—

—rn

aig

tx

rx

r—

——

x—

——

T-90

-1A

xsr

x—

—h

bzt

na

—sr

srx

sr—

——

——

——

rut

exfr

grt

?

T-90

-1B

xsr

gt

x—

—h

b,n

a—

xsr

xsr

—?

r—

—?

——

T-90

-1B

EC

Lx

srg

tx

——

nc,

na

—x

srx

sr—

?r

x—

?—

chl+

ms

xcfo

l

T-90

-1B

AM

PH

xsr

gt

x—

—h

bzt

na

—tr

srx

sr—

?r

x—

?—

T-90

-1C

xsr

gt

x—

—h

bzt

na

—x

srx

srg

t—

?r

——

?—

iep

inat

tnm

us

T-90

-2A

xsr

gt

x—

—h

bzt

na

—tr

srx

x—

—r

xx

x—

mcv

T-90

-3A

BS

Trc

hl

sr—

——

na

—x

x—

x—

——

—x

——

T-90

-3A

EC

Lx

srg

tx

r—

na

igt

xx

—sr

gt

——

r—

x?

——

T-90

-3B

xsr

gt

xrg

t—

na

igt

xx

csx

——

rx

xx

——

T-90

-5C

xsr

gt

rgt

——

na

igt

xsr

xx

——

——

——

—ru

texg

t?m

cv

T-90

-6x

rgt

x—

——

—x

x—

x—

—r

——

x—

mu

s+ep

ive

in

GL-

16-8

xr

——

—n

aig

tx

rx

x—

——

——

——

MH

-90-

1AE

CL

xr,

srg

t—

—tr

ncz

tna

—x

srx

x—

——

——

——

8

by guest on June 28, 2015http://petrology.oxfordjournals.org/

Dow

nloaded from

SORENSEN et al. PHENGITE-HOSTED LILE ENRICHMENT

Sam

ple

grt

chl

epi

law

alb

amp

qtz

cpx

tte

rut

mu

sst

pm

gt

hm

tzi

rap

tp

yto

thco

mm

ents

MH

-90-

1AA

MP

Hx

r,sr

gt

ree

——

hb

ztn

cztn

a—

trsr

xx

——

——

——

——

MH

-90-

3—

r—

——

na

xx

x—

x—

——

——

——

MH

-90-

11C

xr,

srg

t—

——

ncz

tna

—x

xx

x—

——

——

——

MH

-90-

12A

xr,

srg

t—

—tr

na

—x

srx

srg

tr

——

——

——

RF

xsr

x—

plg

ca—

—sr

x—

—x

—x

xx

——

6871

86—

——

x—

na

——

x—

xx

——

xx

——

6871

8—

—x

——

cazt

na

——

srx

x—

—r

——

—p

um

pn

ot

anH

GB

Tra

nsit

ion

zo

ne

s

Fran

cisc

anC

mpx

.

OR

E-1

-5—

x—

——

na

——

srx

x—

——

—x

——

OR

E-3

-5—

——

——

na

——

srx

x—

——

—x

——

Rin

ds

Sam

ana,

D.R

.

SS

84-2

4C—

mg

ree

——

ca—

—sr

x—

——

——

——

cal+

tlc

SS

84-2

4D—

xre

e—

—ca

x—

srx

x—

——

xx

—ca

l—

SS

85-2

7B1-

IR—

xre

e—

xca

ztn

a—

—sr

xx

——

rx

x—

——

SS

85-2

7B2-

OR

—m

gre

e—

—ca

ztn

a—

—r

xx

——

——

——

——

Fran

cisc

anC

mpx

.

OR

E-1

-2—

xre

e—

—ca

ztn

ax

—sr

rxl

——

——

x—

——

ina

intt

n

OR

E-3

-2—

x—

——

cazt

na

——

——

——

——

——

——

GL-

14-1

—x

——

—ca

ztn

a—

—r

xx

——

—x

——

——

T-90

-1D

—m

g—

——

ca—

—sr

xx

——

—x

x—

——

T-90

-2B

—m

g—

——

hb

ztn

a—

—x

xx

——

——

——

——

GL-

16-1

—x

——

—ca

ztn

a—

—r

xx

——

—x

——

——

GL-

16-4

—x

——

—ca

ztn

a—

—r

xx

——

—x

——

——

MH

-90-

8—

xre

e—

—h

bzt

ncz

tna

—x

srx

x—

——

——

——

MH

-90-

9—

——

——

hb

ztn

cztn

a—

xx

xx

——

——

——

——

alb

,alb

ite;

amp

,am

ph

ibo

le;a

pt,

apat

ite;

ca,c

alci

cam

ph

ibo

le;c

al,c

alci

te;c

hl,

chlo

rite

;cp

x,cl

ino

pyr

oxe

ne;

cs,r

uti

leo

ccu

rso

nly

asin

clu

sio

ns

inth

eco

res

of

tita

nit

eg

rain

s;cz

,clin

ozo

isit

e;ep

i,ep

ido

te;e

xfr,

exso

lved

fro

m;g

rt,g

arn

et;h

bzt

na,

ho

rnb

len

de

iszo

ned

toso

dic

amp

hib

ole

;hb

ztn

cztn

a,h

orn

ble

nd

eis

zon

edto

sod

ic–c

alci

can

dth

ento

sod

icam

ph

ibo

le;H

GB

,hig

h-g

rad

eb

lock

;hm

t,h

emat

ite;

icp

x,in

clu

sio

ns

ofc

lino

pyr

oxe

ne;

iep

i,o

ccu

rsas

incl

usi

on

sin

epid

ote

;iep

inat

tnm

us,

incl

usi

on

so

fep

ido

te,

sod

icam

ph

ibo

le,

and

tita

nit

ear

ep

rese

nt

inm

usc

ovi

te;

igt,

occ

urs

on

lyas

incl

usi

on

sin

gar

net

;in

a,in

clu

sio

ns

of

sod

icam

ph

ibo

le;

iree

nc,

rare

-ear

th-

epid

ote

incl

usi

on

so

ccu

rin

sod

icam

ph

ibo

le;l

aw,l

awso

nit

e;m

cv,m

usc

ovi

te+

chlo

rite

vein

;mg

,mag

nes

ian

;mg

t,m

agn

etit

e,m

us;

mu

sco

vite

,na;

sod

icam

ph

ibo

le;

nc,

sod

ic–c

alci

cam

ph

ibo

le;

ncz

tca,

sod

ic–c

alci

cam

ph

ibo

leis

zon

edto

calc

icam

ph

ibo

le;

ncz

tna,

sod

ic–c

alci

cam

ph

ibo

leis

zon

edto

sod

icam

ph

ibo

le;

oth

,o

ther

;p

lg,

pla

gio

clas

e;p

um

p,

pu

mp

elly

ite;

pyt

,p

yrit

e;q

tz,

qu

artz

;re

e,ra

re-e

arth

-ele

men

t-b

eari

ng

epid

ote

;r,

min

eral

isa

retr

og

rad

ep

has

e;rc

al,

calc

ite

occ

urs

asa

retr

og

rad

em

iner

al;r

gt,

min

eral

occ

urs

asa

rep

lace

men

to

fg

arn

et;r

na,

retr

og

rad

eso

dic

amp

hib

ole

;ru

t,ru

tile

;sr,

som

eo

fth

em

iner

alis

are

tro

gra

de

ph

ase;

srg

t,so

me

of

this

min

eral

occ

urs

asa

rep

lace

men

to

fg

arn

et;s

tp,s

tilp

no

mel

ane;

tlc,

talc

;tte

,tit

anit

e;tu

r,to

urm

alin

e;v,

occ

urs

asa

vein

min

eral

;x,m

iner

alp

rese

nt;

xc,

cro

ss-c

ut;

zir,

zirc

on

;zo

,zo

isit

e.

9

by guest on June 28, 2015http://petrology.oxfordjournals.org/

Dow

nloaded from

JOURNAL OF PETROLOGY VOLUME 38 NUMBER 1 JANUARY 1997

diameter, with or without rinds, in semi-concordant lenses National Museum of Natural History (DMS-NMNH).The INAA analyses were performed at the US Geologicalof talc and chlorite schist which locally cut the layering-

parallel foliation of sequences of metasedimentary rocks, Survey, Reston, Virginia. Details of the analytical pro-cedures and additional references for these techniquesand (3) loose boulders of [1 m in diameter, found in

canyons or on beaches between the ‘marble locality’ and have been cited by Sorensen & Grossman (1989).Electron microprobe analyses of muscovite were per-the town of Samana. The loose blocks, many of which

display rinds, are presumably derived from outcrops formed with the ARL-SEMQ microprobe at DMS-NMNH, using mineral and synthetic standards developedsimilar to setting (2). Most of the small eclogite lenses

and boudins in the marble outcrop [setting (1)] show by Jarosewich et al. (1978). Counting time for eachelement was 20 s. For each grain, a minimum of threeextensive cataclasis and alteration of clinopyroxene, re-

placement of clinopyroxene and garnet by carbonate, analyses were performed adjacent to each other.Throughout each analysis, the sample was moved slightlyand development of retrograde epidote blueschist facies

assemblages. Samples SAM-9 and SAM-12E are from to minimize volatilization effects upon K; significantlylower total counts for K were observed if this was notthis locality.

Four eclogite blocks were studied in detail: (1) SS85- done.22, a block of ~0·5 m diameter, from a pod of talc–chloriteschist ~10 m wide within a sequence of marble andfoliated micaceous marble; (2) SS84-24, a rind-bearing

PETROLOGY AND GEOCHEMISTRYblock ~2 m in diameter, found in an arroyo; (3) SS84-24/85, a second block similar in size and from the same Mineral assemblages of high-grade blockslocality as (2), but which shows extensive calcic alteration; The mineral assemblages of Franciscan and Samana(4) SS85-27, a rind-bearing block of interlayered blue- high-grade blocks are typical of eclogite, high-grade blue-schist and eclogite 3 m in diameter, from the beach near schist, and garnet+ epidote-bearing amphibolite (Tablethe locality of (2) and (3). Transition zones and rinds 1). Sodic amphibole and omphacitic clinopyroxene com-from blocks SS84-24, SS84-24/85, and SS85-27 were monly coexist in high-T mineral assemblages of blue-also collected and analyzed. The rind around block SS85- schist, and in eclogite. High-grade blocks typically show27 was subsampled as inner (SS85-27B1-IR) and outer complex textural relationships that reflect varying degrees(SS85-27B2-OR) rind. Giaramita & Sorensen (1994) re- of retrogression under epidote blueschist or lawsoniteported the following P–T estimates for three of these blueschist facies P–T conditions (Table 1). The samplesblocks: for SS84-22, P min = 8·2 ± 0·8 kbar, T = 502 are classified as blocks, altered blocks, transition zones,± 42°C; for SS84-24, P min= 9·6± 1·1 kbar, T= 597 and rinds (Table 1) based on field relations and thin± 59°C; for SS85-27, P min= 9·9± 1·0 kbar, T= 601 section study. Altered blocks are regions of a block’s± 55°C. Samples R548, R750, and R751 were collected interior that show unusually large amounts of sulfidefrom the Samana Peninsula by Fred Nagle of the Uni- minerals, or of minerals which typically occur as retro-versity of Miami, and sent to the third author for analysis. grade minerals (e.g. chlorite). Transition zones are

areas that are mineralogically distinct from the block’sinterior, and which separate blocks from rinds (e.g.Moore, 1984; Moore, 1986; Sorensen, 1988; Sorensen

ANALYTICAL METHODS & Grossman, 1993).Most whole-rock samples of blocks and rinds were ana-lyzed for major, minor, and trace elements by both X-rayfluorescence (XRF) and instrumental neutron activation

Geochemistry of high-grade blocksanalysis (INAA). Samples R548, R750, R751, SAM-09,Major and minor elementsand SAM-12E were analyzed by the third author at the

Australian National University using XRF and spark- Most Franciscan and Samana high-grade blocks resemblesource mass spectrometric methods. All samples weighed basalts sampled from the Pacific Ocean floor in theirbetween 0·05 and 2 kg. Sample sizes reflect the grain SiO2, Al2O3, TiO2, FeO∗, MgO, and CaO contentssize and amount of material available; subsamples of (Tables 2 and 3, Fig. 2). Values for major elements arerinds and of interlayered lithologies were smaller than plotted against SiO2 because it is the most abundantthose of entire rinds and of the host rocks. Splits were constituent in all of these rocks, and it tends to show aground to <100 mesh in an Al-ceramic SPEX shatterbox. restricted range of values in fresh ocean-floor basalts (Fig.Pure quartz blanks ground in this mill lack significant 2). The recalculated data for blocks are compared withcontamination of any reported element. The XRF ana- a 3r bracket about the mean of 2109 electron microprobelyses and FeO and loss on ignition (LOI) determinations analyses of ocean-floor glasses (including highly frac-

tionated basalts, but excluding andesitic compositions)were carried out in the Department of Mineral Sciences,

10

by guest on June 28, 2015http://petrology.oxfordjournals.org/

Dow

nloaded from

SORENSEN et al. PHENGITE-HOSTED LILE ENRICHMENT

Tab

le2:

Who

le-r

ock

maj

or,

min

oran

dtr

ace

elem

ent

abun

danc

esof

high

-gra

debl

ocks

,tr

ansi

tion

zone

san

dri

nds

Sam

ana,

D.R

.,b

lock

san

dal

tere

db

lock

sFr

anci

scan

blo

cks

and

alte

red

blo

cks

SS

85-2

2S

S84

-24A

SS

84-2

4B84

-24/

8584

-24/

85S

S85

-27D

SS

85-2

7EO

RE

-1-1

0O

RE

-3-4

DR

-WR

DR

-EC

L1D

R-E

CL2

DR

-EC

L3

Gre

enW

hit

e

SiO

2(w

t%

)48

·37

48·4

250

·64

44·2

338

·44

51·4

650

·71

47·9

951

·02

47·6

447

·03

48·5

649

·42

TiO

21·

190·

590·

320·

670·

070·

700·

641·

411·

892·

242·

171·

772·

07A

l 2O

315

·36

19·3

312

·54

22·5

730

·64

16·2

517

·62

14·8

713

·84

14·3

215

·10

13·2

314

·18

Fe2O

34·

142·

382·

912·

792·

162·

043·

054·

095·

645·

664·

144·

855·

12Fe

O6·

525·

068·

592·

810·

147·

165·

306·

806·

547·

237·

615·

586·

77M

nO

0·20

0·19

0·21

0·11

0·03

0·17

0·16

0·19

0·16

0·28

0·36

0·20

0·24

Mg

O5·

434·

1310

·72

3·73

0·28

8·76

5·88

8·23

6·70

5·17

4·87

5·42

5·17

CaO

12·4

517

·18

10·1

817

·86

23·7

24·

659·

2310

·30

7·48

11·3

711

·22

12·7

212

·32

Na 2

O3·

961·

282·

301·

990·

344·

163·

452·

554·

343·

703·

023·

903·

85K

2O0·

870·

030·

720·

140·

091·

972·

510·

960·

950·

821·

831·

320·

84P

2O5

0·39

0·13

0·07

0·11

0·15

0·09

0·13

0·11

0·15

0·26

0·20

0·23

0·22

LOI

0·59

0·87

0·98

2·34

2·73

2·24

1·55

2·54

1·70

1·06

1·24

1·32

1·15

Tota

l99

·47

99·5

910

0·18

99·3

598

·79

99·6

510

0·23

100·

0410

0·41

99·7

598

·79

99·1

010

1·35

(p.p

.m.)

Sc

3840

2732

9·4

3834

3844

4244

4042

Cr

8627

211

4726

209

171

260

169

240

255

217

224

Co

3025

5420

0·8

3628

4240

4531

3543

Ni

6636

112

366

8165

107

5790

7461

77C

u81

4067

20<5

4443

2624

4762

6459

Zn

7179

112

414·

187

6786

8711

894

105

116

As

1·6

<0·5

<0·5

0·77

0·59

<0·9

<0·9

1·1

<0·7

<0·8

<0·9

<2<0

·8R

b22

<516

<5<5

4863

2021

1943

3020

Sr

430

1490

160

1660

2240

131

450

156

139

334

126

200

336

Y28

119

188

2018

3242

4741

3944

Zr

167

8441

105

8753

6910

012

114

913

611

613

8N

b7

<5<5

<5<5

99

8<5

76

1316

Sb

0·21

0·23

0·16

0·27

0·28

<0·1

30·

130·

130·

170·

250·

180·

24<0

·23

Cs

1·10

<0·3

0·72

0·28

0·09

2·54

3·44

0·83

0·38

0·61

1·10

0·84

0·64

Ba

450

<20

283

109

7589

018

5030

123

644

086

062

044

0La

13·9

3·31

5·2

5·7

2·0

5·7

8·9

4·3

4·7

5·6

5·6

4·4

5·0

Ce

30·9

8·0

10·7

12·3

3·9

12·9

17·0

11·9

13·7

15·7

14·9

12·9

14·5

Nd

20·5

5·7

5·7

9·6

<49·

210

·610

·211

·513

·013

·510

·013

·3S

m5·

61·

831·

382·

671·

232·

152·

713·

884·

85·

44·

64·

285·

0E

u1·

440·

660·

480·

970·

660·

670·

861·

171·

571·

651·

361·

401·

54T

b0·

700·

310·

230·

500·

280·

400·

450·

821·

121·

221·

000·

991·

11Y

b3·

261·

281·

052·

000·

511·

811·

873·

104·

05·

04·

74·

14·

8Lu

0·52

0·21

0·14

0·27

0·06

0·29

0·28

0·43

0·54

0·68

0·67

0·58

0·69

Hf

4·3

1·10

0·92

1·76

0·30

1·58

1·57

2·43

3·22

3·74

3·69

3·14

3·59

Ta0·

25<0

·10·

090·

11<0

·03

<0·2

0·24

0·24

0·20

0·22

0·31

0·22

0·31

Th

2·05

0·48

1·30

0·64

<0·0

60·

972·

250·

28<0

·40·

29<0

·15

0·32

0·26

U0·

890·

400·

700·

990·

410·

400·

690·

310·

63<0

·4<0

·3<0

·3<0

·4A

u(p

.p.b

.)12

1014

103

<13

77

616

1013

<9

11

by guest on June 28, 2015http://petrology.oxfordjournals.org/

Dow

nloaded from

JOURNAL OF PETROLOGY VOLUME 38 NUMBER 1 JANUARY 1997

Tab

le2:

cont

inue

d

Fran

cisc

anb

lock

san

dal

tere

db

lock

s

DR

-BS

T1

DR

-BS

T2

GL-

14-2

T-90

-1A

T-90

-1B

T-90

-1B

T-90

-1B

T-90

-1C

T-90

-2A

T-90

-3A

T-90

-3A

T-90

-3A

EC

LA

MP

HB

ST

EC

L

SiO

2(w

t%

)46

·82

45·5

249

·61

43·4

241

·83

42·7

243

·54

44·1

446

·13

48·3

149

·05

47·4

0Ti

O2

2·23

2·29

1·20

0·98

0·99

0·97

1·04

0·96

1·27

2·13

2·23

2·28

Al 2

O3

13·7

814

·03

13·4

117

·31

17·1

917

·00

17·9

716

·99

16·8

215

·65

15·8

616

·49

Fe2O

35·

254·

873·

724·

104·

344·

584·

424·

082·

903·

602·

974·

29Fe

O7·

788·

326·

189·

218·

578·

308·

759·

185·

526·

397·

215·

45M

nO

0·25

0·27

0·16

0·27

0·29

0·30

0·32

0·30

0·14

0·21

0·18

0·22

Mg

O5·

174·

797·

028·

427·

557·

647·

688·

318·

926·

137·

415·

19C

aO10

·96

10·7

912

·71

11·3

511

·62

11·8

011

·85

11·1

011

·73

10·0

67·

0111

·79

Na 2

O3·

783·

484·

872·

672·

502·

602·

432·

812·

263·

693·

703·

36K

2O0·

520·

620·

160·

530·

590·

460·

540·

561·

391·

702·

441·

73P

2O5

0·27

0·26

0·09

0·06

0·05

0·08

0·06

0·07

0·17

0·38

0·33

0·42

LOI

0·96

0·99

0·90

1·78

2·44

2·71

2·00

2·37

2·63

1·92

3·05

1·56

Tota

l97

·77

96·2

310

0·03

100·

1097

·96

99·1

610

0·60

100·

8799

·88

100·

1710

1·44

100·

18

(p.p

.m.)

Sc

4244

3957

5755

5656

3530

2836

Cr

245

248

177

163

130

142

140

128

400

197

176

194

Co

5250

3141

3837

3638

4237

4428

Ni

9795

6064

6755

5257

162

9098

59C

u32

35<5

7358

7131

2052

7051

14Z

n13

513

853

241

173

164

178

170

7099

129

71A

s<0

·8<0

·8<1

·1<1

·2<0

·8<0

·9<0

·8<0

·9<0

·51·

5<0

·71·

4R

b11

15<5

89

910

1232

4367

44S

r41

539

561

030

941

542

639

932

129

223

610

931

8Y

4950

2940

4746

4943

2332

2934

Zr

156

160

9594

105

9810

898

9016

217

316

4N

b5

10<5

<5<5

<5<5

<58

2229

24S

b0·

220·

19<0

·30·

130·

200·

200·

150·

210·

170·

240·

22<0

·2C

s0·

410·

470·

36<0

·3<0

·3<0

·3<0

·30·

340·

340·

530·

790·

50B

a26

732

914

613

720

722

518

918

860

081

011

4083

0La

5·7

5·7

4·2

7·0

8·3

8·2

8·7

7·8

10·3

21·4

22·9

21·2

Ce

16·0

16·1

11·1

15·8

18·6

17·9

19·2

17·5

21·5

4449

43N

d15

·712

·68·

010

·312

·414

·414

·213

·111

·523

·125

·122

·0S

m5·

45·

52·

884·

34·

94·

85·

24·

73·

475·

96·

26·

2E

u1·

741·

801·

021·

471·

611·

551·

681·

591·

161·

721·

731·

95T

b1·

251·

330·

651·

101·

151·

121·

211·

090·

630·

980·

931·

07Y

b5·

05·

22·

804·

85·

65·

56·

55·

52·

502·

972·

763·

64Lu

0·72

0·76

0·40

0·70

0·83

0·81

0·95

0·84

0·36

0·43

0·40

0·49

Hf

3·83

3·92

2·01

2·78

2·91

3·02

3·28

2·99

2·31

4·1

4·7

4·3

Ta0·

280·

300·

250·

310·

290·

280·

260·

281·

022·

112·

232·

14T

h0·

31<0

·20·

361·

101·

271·

241·

351·

321·

192·

823·

212·

54U

<0·3

<0·3

0·19

1·02

1·16

1·24

1·35

1·06

0·36

0·54

0·70

0·47

Au

(p.p

.b.)

12<9

<1<6

1158

14<6

67

<9<7

12

by guest on June 28, 2015http://petrology.oxfordjournals.org/

Dow

nloaded from

SORENSEN et al. PHENGITE-HOSTED LILE ENRICHMENT

Fran

cisc

anb

lock

san

dal

tere

db

lock

s

T-90

-3B

T-90

-5C

T-90

-6G

L-16

-8M

H-9

0-1A

MH

-90-

1AM

H-9

0-1A

MH

-90-

3M

H-9

0-11

CM

H-9

0-12

AR

F68

7186

6871

8

AM

PH

EC

L

SiO

2(w

t%

)47

·75

49·2

447

·57

53·5

650

·34

47·2

652

·97

55·1

150

·97

53·0

744

·76

49·6

548

·81

TiO

21·

702·

011·

220·

931·

441·

581·

021·

241·

480·

693·

000·

811·

93A

l 2O

315

·21

13·0

815

·71

9·44

11·5

613

·89

10·7

910

·17

11·5

59·

6713

·94

21·1

813

·06

Fe2O

33·

384·

553·

314·

093·

312·

513·

783·

922·

334·

361·

523·

935·

12Fe

O7·

949·

845·

435·

007·

259·

874·

954·

858·

224·

4315

·10

5·19

8·01

Mn

O0·

210·

290·

170·

270·

290·

320·

230·

190·

290·

220·

310·

150·

21M

gO

7·04

6·43

6·45

7·53

9·43

10·9

28·

048·

429·

438·

116·

964·

066·

95C

aO8·

3911

·30

15·8

112

·60

9·39

6·44

11·0

98·

899·

4412

·15

9·07

8·50

8·73

Na 2

O4·

034·

123·

485·

544·

213·

075·

254·

743·

295·

633·

083·

023·

24K

2O1·

090·

290·

310·

221·

081·

671·

150·

771·

070·

450·

402·

581·

43P

2O5

0·22

0·10

0·12

0·08

0·05

0·01

0·07

0·06

0·07

0·08

0·32

0·16

0·23

LOI

2·30

0·02

0·92

0·77

1·49

2·49

1·15

2·07

1·46

1·13

1·25

0·37

2·53

Tota

l99

·26

101·

2710

0·50

100·

0399

·84

100·

0110

0·49

100·

4399

·60

99·9

999

·71

99·6

010

0·25

(p.p

.m.)

Sc

4347

4238

4149

3534

4438

5037

44C

r22

510

339

226

929

725

325

018

931

727

762

420

88C

o42

3635

2033

5221

2245

2346

4143

Ni

9669

133

9411

199

125

121

101

9462

104

59C

u10

46

5341

5060

6695

4140

171

3321

Zn

133

127

5512

717

319

014

814

217

412

714

512

910

1A

s1·

1<1

·21·

3<1

·1<0

·9<0

·7<1

·2<0

·8<0

·7<1

<0·9

<0·8

<0·8

Rb

2914

87

2844

3226

3915

<556

33S

r18

224

227

3021

3328

<20

2147

152

234

81Y

3357

2619

2126

1923

2312

6122

45Z

r10

413

078

5159

6949

5457

3817

051

109

Nb

<5<5

<58

65

<5<5

<5<5

8<5

10S

b0·

230·

160·

200·

170·

13<0

·20·

120·

15<0

·2<0

·2<0

·20·

30<0

·3C

s0·

300·

25<0

·20·

220·

350·

430·

390·

470·

600·

23<0

·32·

870·

76B

a57

016

919

319

683

014

2010

7066

011

0042

011

132

210

1La

7·4

5·2

3·85

3·36

4·7

5·3

2·88

4·1

4·9

3·44

6·9

2·02

3·99

Ce

17·1

15·0

9·7

7·6

10·0

11·9

5·5

8·5

10·2

6·8

19·0

4·1

10·9

Nd

11·6

13·2

8·3

6·3

9·2

10·5

4·9

9·3

9·8

5·3

16·8

4·8

9·2

Sm

4·3

4·9

2·98

2·90

2·95

3·23

2·52

3·99

3·70

1·99

6·66

1·73

4·3

Eu

1·33

1·48

0·94

0·97

0·87

0·87

0·88

1·26

1·12

0·54

2·12

0·71

1·40

Tb

0·91

1·56

0·69

0·57

0·45

0·48

0·49

0·78

0·64

0·27

1·60

0·50

1·14

Yb

3·60

6·1

2·91

2·49

3·68

4·4

2·52

1·99

3·44

1·69

6·4

2·02

4·7

Lu0·

530·

890·

400·

390·

560·

650·

370·

260·

520·

230·

890·

300·

66H

f2·

983·

932·

131·

342·

052·

621·

421·

612·

091·

184·

81·

122·

90Ta

0·58

0·49

0·27

0·27

0·32

0·24

0·28

0·44

0·33

0·16

0·41

<0·1

30·

27T

h0·

770·

430·

32<0

·20·

320·

33<0

·30·

280·

29<0

·30·

49<0

·20·

38U

0·38

0·27

0·35

0·30

0·35

0·39

<0·4

0·32

0·28

<0·4

1·23

0·38

1·39

Au

(p.p

.b.)

<6<9

<5<1

<10

8<5

<10

<6<6

<1<1

<1

13

by guest on June 28, 2015http://petrology.oxfordjournals.org/

Dow

nloaded from

JOURNAL OF PETROLOGY VOLUME 38 NUMBER 1 JANUARY 1997

Tab

le2:

cont

inue

d

Fran

.tr

.zo

nes

Sam

ana

rin

ds

Fran

cisc

anri

nd

s

OR

E-1

-5O

RE

-3-5

SS

84-2

4CS

S84

-24D

SS

85-2

7-B

1-S

S85

-27-

B2-

OR

E-1

-2O

RE

-3-2

GL-

14-1

T-90

-1D

T-90

-2B

GL-

16-1

GL-

16-4

MH

-90-

8M

H-9

0-9

IRO

R

SiO

2(w

t%

)46

·28

54·3

351

·07

40·6

153

·00

54·9

951

·75

51·5

442

·97

51·7

151

·94

54·5

151

·62

50·4

148

·22

TiO

22·

321·

980·

380·

650·

580·

440·

210·

060·

570·

280·

560·

890·

330·

721·

52A

l 2O

315

·40

12·2

58·

3415

·19

12·7

47·

445·

124·

9211

·04

7·52

9·52

7·52

7·74

10·2

712

·75

Fe2O

33·

633·

931·

431·

682·

991·

831·

791·

823·

161·

331·

511·

361·

461·

731·

92Fe

O9·

366·

706·

487·

645·

415·

426·

406·

4110

·38

6·14

6·36

5·90

6·74

7·94

9·98

Mn

O0·

160·

110·

130·

140·

130·

170·

290·

320·

260·

170·

190·

190·

150·

170·

22M

gO

10·5

68·

1917

·38

17·3

612

·32

16·3

520

·56

20·9

919

·36

17·7

716

·11

15·4

316

·89

14·2

911

·37

CaO

1·96

2·13

9·12

6·38

4·09

6·81

9·12

9·50

5·58

9·95

8·43

8·31

10·0

28·

786·

13N

a 2O

2·99

5·36

1·62

0·51

3·52

3·22

1·19

0·93

1·42

1·95

2·50

1·56

2·12

2·38

3·15

K2O

2·53

2·22

0·84

2·00

2·17

0·57

0·06

0·09

0·11

0·54

0·92

2·01

0·70

1·31

1·67

P2O

50·

060·

100·

040·

060·

110·

050·

060·

060·

040·

080·

070·

050·

100·

090·

02LO

I4·

341·

912·

997·

672·

992·

393·

433·

395·

622·

122·

262·

341·

791·

833·

43To

tal

99·5

999

·21

99·8

299

·89

100·

0599

·68

99·9

810

0·03

100·

5199

·56

100·

3710

0·07

99·6

699

·92

100·

38

(p.p

.m.)

Sc

6645

2730

3428

7·4

1136

2230

1917

2846

Cr

130

281

1050

283

570

960

1810

1840

1340

1080

1570

1370

890

690

297

Co

4041

5256

4347

6466

7248

5846

4959

53N

i66

8764

020

933

067

013

6012

0082

061

077

069

073

056

013

8C

u41

<350

2160

18<5

149

<514

33<5

610

1Z

n81

106

114

123

104

7982

8511

911

470

153

121

153

161

As

<1<0

·9<0

·51·

1<0

·9<0

·6<0

·3<0

·5<1

<0·5

<0·5

<0·8

<0·8

<0·5

<0·7

Rb

6456

1845

5815

<5<5

<513

2046

1533

46S

r<9

<529

8128

<20

<20

<20

<20

28<2

024

2425

121

Y48

396

1514

136

711

1516

1711

1225

Zr

161

121

3266

5851

23<1

035

2861

4228

3110

7N

b<5

15<5

9<5

<5<5

<5<5

<5<5

17<5

<5<5

Sb

<0·4

0·18

<0·0

7<0

·08

0·15

<0·0

8<0

·06

<0·0

6<0

·1<0

·08

<0·2

<0·0

8<0

·08

<0·0

9<0

·1C

s1·

940·

880·

951·

942·

530·

68<0

·18

<0·2

<0·2

<0·2

<0·2

0·38

0·22

0·28

0·46

Ba

1890

940

254

940

1480

314

<20

<20

<20

269

440

1170

400

1030

1360

La8·

45·

41·

825·

36·

54·

20·

820·

510·

852·

637·

72·

250·

872·

825·

0C

e20

·115

·65·

212

·115

·212

·22·

89<1

·9<3

5·5

15·5

5·2

<1·8

7·1

11·7

Nd

15·7

11·5

<66·

38·

86·

4<1

·6<1

·7<2

·43·

608·

3<3

<46·

311

·9S

m5·

14·

40·

781·

832·

581·

260·

540·

290·

731·

402·

391·

110·

641·

824·

1E

u1·

811·

290·

280·

540·

680·

290·

160·

150·

270·

460·

600·

410·

250·

561·

27T

b1·

061·

030·

200·

280·

410·

270·

150·

140·

260·

310·

470·

280·

250·

350·

74Y

b4·

44·

30·

470·

811·

451·

360·

540·

701·

291·

471·

741·

800·

931·

213·

14Lu

0·62

0·57

0·07

0·12

0·21

0·18

0·09

0·11

0·18

0·20

0·23

0·25

0·11

0·16

0·43

Hf

4·3

3·24

0·78

1·49

1·58

1·43

0·45

0·13

0·84

0·75

1·76

1·05

0·74

1·21

2·46

Ta0·

360·

280·

110·

120·

160·

120·

11<0

·06

0·14

0·21

0·37

0·30

<0·0

7<0

·30·

28T

h1·

790·

410·

461·

241·

461·

010·

30<0

·09

0·30

0·50

2·51

0·32

0·30

0·29

0·24

U0·

86<0

·60·

210·

570·

700·

38<0

·3<0

·2<0

·3<0

·20·

31<0

·3<0

·3<0

·20·

54A

u(p

.p.b

.)<8

<7<4

<3<9

<4<3

<3<1

<46

<1<1

<510

14

by guest on June 28, 2015http://petrology.oxfordjournals.org/

Dow

nloaded from

SORENSEN et al. PHENGITE-HOSTED LILE ENRICHMENT

from the Pacific Ocean (source of data: Smithsonian K2O (Melson et al., 1976). Some Franciscan greenstoneblocks are rich in K2O compared with the Pacific ocean-Institution Ocean Floor Glass Database; Melson et al.,

1976). The 3r ocean-floor glass field in Fig. 2 represents floor glass data (Fig. 2); this feature suggests both theyand the suites of high-grade blocks have been affectedextreme values for compositions of volcanic glasses from

the Pacific ocean basin. The field is intended to proxy by K alteration.for the likely compositions of unaltered protoliths ofFranciscan high-grade blocks. (This is not an appropriate Minor and trace elements, exclusive of K, Ba, Rb, and Csassumption for the Samana high-grade blocks; see below.)

The abundances and ratios of ‘immobile’ minor andFields of major element concentrations for 35 whole-trace elements for Franciscan high-grade blocks suggestrock analyses of greenstone blocks from the Franciscanthat most are derived from MORB-like protoliths; forComplex (Shervais & Kimbrough, 1987; MacPherson etSamana high-grade blocks, a comparable dataset sup-

al., 1990) show large overlap with the 3r ocean-floorports the suggestion of Perfit & McCulloch (1982) andglass field and with data for high-grade blocks. These Perfit et al. (1982) that these rocks primarily are meta-

greenstone blocks are very low-grade metabasaltic rocks morphosed island arc volcanic rocks, or sediments derivedthat provide an analog for protoliths altered at low-T from them. ‘Spider-diagrams’ of whole-rock minor andmetamorphic conditions, either on the seafloor or in the trace elements normalized to MORB (Fig. 3a; Pearce,subduction complex. Most of these greenstone samples 1982) illustrate that most Franciscan high-grade blocksdisplay mineral assemblages found in ‘classic’ spilite (e.g. are little fractionated in REE, high field strength elementsCoombs, 1974). (HFSE), and Sc compared with the normalizing values

Two other reference suites reflect unusual or un- of Pearce (1982), and span an ‘average’ MORB com-common protolith types that nonetheless could be sub- position. In addition, most plots of REE for Franciscanducted, or are represented by subducted materials. One high-grade blocks have slightly light (L) REE-depletedis composed of highly Fe–Ti-enriched ocean-floor gabbro, or flat REE patterns at abundances of ~10–30× chon-and the other, of rodingite. A field of data for Fe–Ti- drite (Fig. 3d). In contrast, most Samana high-gradeenriched gabbro is defined in Fig. 2 by values from blocks are depleted in HFSE (Fig. 3b) and enriched inpublished analyses of such rocks from ODP core 735B, LREE compared with MORB (Fig. 3e). Most samplesdrilled on Leg 118 (Shipboard Scientific Party, 1989) display distinctly LREE-enriched patterns at REE abund-and from samples dredged from the mid-Atlantic ridge ances of ~10–50 × chondrite for LREE, ~5–20 ×near 24°N (Miyashiro & Shido, 1980), blueschist and chondrite for HREE. The protoliths of Franciscan green-eclogite from Alaska (Barker et al., 1994), and unpublished stone blocks are mostly MORB, but include a few ex-data of S. S. Sorensen & J. N. Grossman (1994) for amples of IAB, and possible off-axis seamount basalteclogite facies Fe–Ti metagabbro boudins from Monviso (MacPherson et al., 1990). One Franciscan greenstoneand Gruppo di Voltri, western Alps; all 81 analyses were locality is a semi-intact seamount (MacPherson, 1983).recalculated on an anhydrous basis. The only sample The scatter in the data for Franciscan greenstone blocksthat plots consistently in the field defined by the Fe–Ti in Fig. 3c and 3f thus reflects a wide range of possiblereference suite is RF, from Baja California (Table 2, Fig. protoliths.2). The rodingite reference suite represents a type of calcic A plot of TiO2 abundance vs chondrite-normalizedalteration seen in basalt dikes emplaced into serpentinite La/Sm highlights the contrast in geochemistry betweenwithin ophiolitic terranes. The rodingite field is defined Franciscan and Samana high-grade blocks (Fig. 4). Nearlyby 19 analyses of eclogite facies metarodingites from all of the Franciscan rocks display TiO2 values >1 wt %,Cima di Gagnone, Lepontine Alps (Evans et al., 1981). and show little evidence for LREE enrichment (LaCN/Two samples from block SS84/85, from the Samana SmCN < 1), whereas the Samana rocks display TiO2Peninsula, are as rich in Ca and Al as are metarodingites, values <1 wt % and are LREE enriched. The formerbut also distinctly poorer in both MgO and FeO∗ con- characteristics are typical of MORB, the latter, of islandtents. This block contains large modal amounts of calcite, arc volcanic rocks (Perfit et al., 1980a; BVSP, 1981). Alltremolite, and zoisite, and also displays veins and pods but two analyses of Franciscan high-grade blocks lieof these minerals. It may have undergone metasomatic within fields defined by data for Franciscan greenstoneexchange with carbonate rocks during recrystallization. blocks on this plot, which supports the assumption that

In contrast to results for the other major elements, the two populations sampled a similar protolith as-many of the high-grade blocks are enriched in Na2O, semblage (i.e. subducted ocean-floor rocks).and particularly in K2O, compared with Pacific ocean-floor glasses; 14 analyses plot at extreme K2O values

Systematics of K, Ba, Rb, and Cscompared with ‘average’ Pacific MORB values of ~0·11wt % (Fig. 2; BVSP, 1981). Indeed, even the relatively In addition to K, the elements Ba, Rb and Cs are

enriched in many samples of both Franciscan anduncommon E-type MORB typically contains <0·6 wt %

15

by guest on June 28, 2015http://petrology.oxfordjournals.org/

Dow

nloaded from

JOURNAL OF PETROLOGY VOLUME 38 NUMBER 1 JANUARY 1997

Fig. 2. Major element data for high-grade blocks, normalized on an anhydrous basis. Localities: Χ, Franciscan Complex, California andOregon; Ν, Κ, Samana Peninsula; Ο, Baja California. Κ, samples analyzed at the Australian National University (Table 3). Fields of unplottedanalyses: ofg, ocean-floor glass data (continuous line); mrd, metarodingite data (dot–dashed line); feti, Fe–Ti gabbro data (dotted line); gst,

Franciscan greenstone data (dashed line). Data sources are cited in the text.

16

by guest on June 28, 2015http://petrology.oxfordjournals.org/

Dow

nloaded from

SORENSEN et al. PHENGITE-HOSTED LILE ENRICHMENT

Table 3: Major, minor and trace element abundances in high-grade blocks,

Samana Peninsula, Dominican Republic

R548 R750 R751 SAM9 SAM12E

(wt %)

SiO2 53·21 58·04 53·06 51·59 53·35

TiO2 1·24 0·54 0·53 2·02 0·91

Al2O3 11·04 14·88 15·77 12·80 17·80

Fe2O3∗ 7·80 7·71 9·43 10·78 9·02

MnO 0·13 0·00 0·00 0·10 0·22

MgO 7·33 7·57 8·99 8·30 6·45

CaO 14·7 6·26 7·30 6·76 10·24

Na2O 5·41 3·45 5·23 4·45 3·00

K2O 0·02 2·32 0·64 2·28 0·28

P2O5 0·00 0·00 0·00 0·23 0·18

Total 100·88 100·77 100·95 99·31 101·45

(p.p.m.)

Rb 0·1 62·3 10·4 70 8

Sr 143 171 333 83 525

Y 38 32 14 40 30

Zr 416 101 55 167 179

Nb 13 4 1·7 5 3·3

Cs 0·01 3·73 1·09 3·3 0·57

Ba 2·8 659 122 1230 84

La 65 16·9 5·3 7·51 27·2

Ce 136 36 14·3 22·1 61·2

Pr 16·5 5·02 1·98 3·7 8·7

Nd 64 21·1 9·34 19·9 40·6

Sm 11·8 4·56 2·25 6·5 8·29

Eu 2·51 1·14 0·69 2·17 2·45

Tb 1·19 0·63 0·38 1·35 0·91

Dy 6·55 3·73 2·48 7·82 4·95

Ho 1·36 0·74 0·5 1·58 0·93

Er 3·85 2·05 1·48 4·44 2·36

Yb 3·42 2·18 1·3 4·26 2·38

Pb 5 3·13 48 4·7 16·7

Th 13·9 3·33 0·97 0·27 2·29

U 5·15 1·11 0·52 0·39 0·49

Samana high-grade blocks compared with MORB (Figs ranges (Fig. 5a,b). In addition, Cs increases with Ba inboth groups of high-grade blocks, although values for2 and 5). Because this feature is also seen in the Franciscan

greenstone suite (Figs 2 and 5; Shervais & Kimbrough, Samana blocks define a trend of lower Ba/Cs than domost of the data for Franciscan ones (Fig. 5c). In contrast,1987; MacPherson et al., 1990), it might merely reflect

isochemical high-grade metamorphism of protoliths increases of Ba versus K, Rb, and Cs are, for the mostpart, not seen in the data for greenstone blocks (Fig. 5d–f ).altered at low-grade conditions. However, the inter-

element systematics of K, Ba, Rb, and Cs in the high- These observations suggest that: (1) LILE enrichment wascontrolled by different processes in the high-grade blocksgrade blocks differ from those seen in the greenstone

suite (Fig. 5). In Franciscan and Samana high-grade compared with Franciscan greenstone blocks, and (2) theenrichment process was probably the same for both theblocks, K, Ba and Rb are strongly intercorrelated; fur-

thermore, data from both localities show overlapping Franciscan and Samana high-grade blocks.

17

by guest on June 28, 2015http://petrology.oxfordjournals.org/

Dow

nloaded from

JOURNAL OF PETROLOGY VOLUME 38 NUMBER 1 JANUARY 1997

Fig. 3. Trace element data for high-grade and greenstone blocks. In (a)–(c), analyses are normalized to the suggested values of Pearce (1982)for average MORB; in (d)–(f ), REE analyses are normalized to the values of Anders & Ebihara (1982) for C1 chondritic abundances, multiplied

by a factor of 1·31. Data for greenstone blocks from MacPherson et al. (1990).

Variations of major, minor and trace elements withininterlayered blueschist–eclogite and amphibolite–eclogite blocks

Blocks DR and T-90-3 consist of interlayered garnetblueschist and eclogite, and blocks T-90-1 and MH-90(GL-16) consist of interlayered garnet amphibolite andeclogite. To evaluate the contributions of each rock typeto the whole-rock composition, as well as the partitioningof elements between rock types, analyses (Table 2) wereperformed on two adjacent, 2–3 cm thick slices of eachblock. The first slice was analyzed as a bulk sample. Thesecond slice was separated along boundaries betweeninterlayered blueschist–eclogite or amphibolite–eclogite,and each rock type was analyzed separately. Data forthe subsamples were normalized to the bulk sample.Additional samples of blocks T-90-1 and MH-90 rep-resent heavily retrograded regions of the blocks, otherFig. 4. Values for TiO2 (normalized on an anhydrous basis) versusdiscrete layers of amphibolite and eclogite, and rindsthe chondrite-normalized La/Sm of high-grade blocks. Symbols are as

in Fig. 2. The fields labeled ‘Greenstones’ represent data for Franciscan around the blocks. Normalized values for subsamples thatgreenstone blocks (MacPherson et al., 1990). The continuous line range within 20% of the bulk-rock value are interpreted toencloses a field for rocks interpreted by MacPherson et al. (1990) to

be isochemical within the resolution of the analyticalreflect MORB and IAB protoliths; the dashed line delineates a fieldfor off-axis seamount basalts. methods and means of separation. If element values for

18

by guest on June 28, 2015http://petrology.oxfordjournals.org/

Dow

nloaded from

SORENSEN et al. PHENGITE-HOSTED LILE ENRICHMENT

suggest that the distribution of K, Ba, Rb, and Cs withinlayered blueschist–eclogite blocks is independent of rocktype.

Amphibolite and eclogite layers in block T-90-1 areessentially isochemical; only Cu and Au are enriched ineclogite compared with amphibolite. Sample T-90-1C,a strongly retrograded sample of interlayered eclogiteand amphibolite, is enriched in LILE compared with lessoverprinted samples. This links some LILE enrichmentto the development of a retrograde blueschist assemblage.Rind sample T-90-1D is enriched in SiO2 and in elementsprobably gleaned from ultramafic rock (MgO, Ni, Cr,and Co) compared with the bulk sample of interlayeredeclogite and amphibolite. It contains much less Al, Sc,Fe, Mn, Zn, HFSE, REE, and K, but much more Baand Rb than does the host block. Similar LILE-rich rindsare found around blocks from the Shuksan MetamorphicComplex of Misch (1966) and the Catalina Schist (So-rensen & Grossman, 1993).

Two types of amphibolite layers are present in theMount Hamilton block (MH-90 and GL-16 samples).Both are poorer in Fe2O3 and richer in FeO, K, Ba, Rb,Cs, Sc, and Co than eclogite layers. The first type ofamphibolite (MH-90-8, -9; GL-16-4; Table 2) displays a‘rind-like’ composition, in that it is rich in Mg, Ni, andCr compared with both the bulk sample (MH-90-1A)Fig. 5. K, Ba, Rb, and Cs systematics of high-grade [(a)–(c)] and

greenstone [(d)–(f )] blocks. Symbols are as in Fig. 2, with the following and eclogite layers within the block. Middle-REE, Ti,additions in (d)–(f ): Ε, analyses of Franciscan greenstone from Mac- Hf, and Zr are either very enriched or greatly depletedPherson et al. (1990);Μ, analyses of Franciscan greenstone from Shervais

in rind-like amphibolite compared with eclogite layers in& Kimbrough (1987). _, a high-K, palagonitized ocean-floor basaltthe block. Rind-like amphibolite samples are made upcomposite from DSDP Hole 417A, analyzed by Staudigel et al. (1981,

1995, and unpublished data, 1994). The dashed boxes labeled OFB almost entirely of hornblende; they lack garnet, Carepresent extreme values for ocean-floor basalts (BVSP, 1981). The epidote, or lawsonite. The presence of Al-rich hornblendelines show constant ratios for the trace elements, values for which are

instead of actinolite suggests that this is a type of high-indicated in (a)–(c).T metasomatic rock (compare Sorensen, 1988). A secondtype of amphibolite (MH-90-1A AMPH; Table 2) con-

subsampled lithologies do not lie within the ±20% tains both garnet and epidote in addition to hornblende.‘envelope’, one layer must be richer in that element than These layers are not particularly Ni and Cr rich, andthe bulk sample, and the other poorer in it for mass contain more Mn and Zn than do either the eclogite or thebalance to be maintained. Of 36 major, minor and trace rind-like amphibolite layers. Compared with interlayeredelements that were compared in lithologies of the four eclogite, garnet–epidote amphibolite is lower in both itssubsampled blocks, elements that could not be mass- Ca and Na contents. The eclogite and amphibolite layersbalanced within 20% are fewer than three per block. of MH-90–GL-16 appear to be much less isochemical

Blueschist and eclogite layers in blocks DR and T-90- in composition than counterparts in the T-90-1 block,3 display strong fractionations between alkali elements. but both show evidence for LILE enrichment linked toIn the Dos Rios block, eclogite layers (DR-E1,-E2, and retrograde and metasomatic assemblages.-E3) are richer in K2O, Ba, Rb, and Cs (as well as inCu and La) compared with adjacent, Sr-rich blueschist

Block-to-rind increases of K, Ba, Rb, and Cslayers (DR-B1,-B2; Table 2). In contrast, a blueschistlayer in block T-90-3 (T-90-3A BST) is richer in K2O, In addition to the observation that rinds can be much

richer in LILE than host blocks, four of the eight sets ofBa, Rb, and Cs (as well as in Zn, Co, and U) than anadjacent, Sr-rich eclogite layer (T-90-3A ECL; Table 2). blocks and rind samples show progressive enrichment in

K, Rb, Ba, and Cs from block to transition zone or rindA second blueschist layer in block T-90-3 (T-90-3B) isrich in Sc and FeO, and poor in HFSE, REE, alkalis (Table 2). The rind around block SS84-24 (samples SS84-

24C, -D) contains about five times the K, Ba, Rb, andand alkaline earth elements, Th, and U compared withblueschist layer T-90-3A. Taken together, these data Cs seen in the host block. The blueschist transition zones

19

by guest on June 28, 2015http://petrology.oxfordjournals.org/

Dow

nloaded from

JOURNAL OF PETROLOGY VOLUME 38 NUMBER 1 JANUARY 1997

between blocks ORE-1 and ORE-3 and their rinds R2+ cation sum (Mg + Fe2+). If one calculates all iron(samples ORE-1-5 and ORE-3-5, respectively) show sim- as Fe3+, the data straddle a line that represents theilar LILE enrichments. The low-T rind (GL-16-1) around phengite substitution. Indeed, Mg p.f.u. is generallythe Mount Hamilton block is enriched in K–Ba–R–Cs within 0·1 cation of an Fe2+-free muscovite–phengitecompared with eclogite layers GL-16-8 and MH-90-1A join (Fig. 7a).ECL. In addition, rind-like amphibolite (GL-16-4, MH- To explain the deviation of data from the continuous90-8, MH-90-9), garnet–epidote amphibolite (MH-90- line in Fig. 7a (which is expressed by the dashed, least-1A AMPH) and retrograded eclogite–amphibolite (MH- squares line through the data), some Fe2+ is probably90-11C) all are richer in LILE compared with the eclogite present in some phengite analyses. However, the oc-layers of this block. The transition zones and rinds formed tahedral cation sums based on the ‘all Fe3+’ assumptionfrom regions of the block that were probably close to or indicate that most Fe is probably Fe3+. To obtain mean-at its exterior surface at the time of metasomatism. This ingful octahedral cation sums, AlVI is calculated by as-suggests an infiltrative mechanism for alkali enrichment signing sufficient Al to fill the tetrahedral sites, andduring rind formation at both low- and high-T conditions. subtracting this value from total Al p.f.u.; Ti plus a

vacancy are assumed to replace two M2+ cations andmaintain local charge balance. The sum of octahedralcations (AlVI+ Fe3+ + 2Ti+Mg) is 1·98–2·02 for 168Textures of muscovite from high-gradeof 210 analyses and 1·97–2·03 for 188 analyses; the totalblocksrange of values is 1·95–2·05 (Table 4). In contrast,Muscovite textures suggest that it is a primary as well asformula calculations based on ‘all Fe2+’ yield octahedrala retrograde mineral in many LILE-rich, high-gradecation sums that range between 2·00 and 2·11, withblocks (Fig. 6). Thin sections of eclogite, blueschist andonly 24 values between 2·00 and 2·03. Because theseamphibolite contain coarse-grained muscovite that iscalculations indicate that ‘all Fe3+’ formulae result inaligned with the foliation defined by coarse-grained om-a large number of stoichiometric analyses, all furtherphacite or hornblende, sodic–calcic amphibole, or coarse-discussions will be based on this assumption.grained sodic amphibole (Fig. 6a,b). Coarse-grained mus-

Interlayer-site sums (K + Na + Ba, p.f.u.) for 210covite occurs as pressure shadows around garnet grainsanalyses range from ~0·92 to 1·03, with most values(Fig. 6a). In sample ORE-3-5, coarse muscovite grainsbetween 0·97 and 1·02 (Table 4). Data for phengiteare broken and recrystallized around fold hinges (Fig.from the Dos Rios eclogite–blueschist block show lower6c). Muscovite grains at the margins of ‘early’ hornblendeinterlayer-site sums (average ~0·93) than do any otherin a garnet amphibolite layer (sample MH-90-1A) ap-samples. Analyses of phengite from the Dos Rios (DR)parently co-crystallized with barroisitic hornblende andblock also show high average Si (~3·4), Mg and Feare overgrown and rimmed by ‘late’ sodic amphibolecontents, and low Na contents (<0·05 p.f.u.) compared(Fig. 6d). On the other hand, muscovite is also partwith phengite from the other blueschist and eclogiteof the mineral assemblage of many rinds, which areblocks (Table 4, Fig. 7). Perhaps unanalyzed componentsretrograde features (Table 1; Moore, 1984; Nelson, 1995).such as Li, B, N, or H2O account for the poor stoi-Textures such as partial replacement of garnet by mus-chiometry of Dos Rios phengite analyses. However, thesecovite + chlorite are easily assigned to a relativelyanalyses yield higher analytical totals than those fromlate stage of recrystallization, especially if muscovite isother samples (Table 4), which would argue against thisintergrown with retrograde assemblages of epidote-grouphypothesis.minerals, lawsonite, and sodic amphibole (Table 1, Fig.

6e). Veins of muscovite± chlorite that crosscut layeringare also likely to be relatively late-stage or even retrograde Na and Ba substitution in phengitefeatures (Fig. 6f ).

The Na content of phengite (the paragonite component)generally anticorrelates with its Si content (Figs 7 and 8;e.g. Guidotti et al., 1994). Paragonite is not observed in

Mineral chemistry of muscovite these samples, but all of them contain either omphaciteThe phengite formula or sodic amphibole, which would also buffer the Na

content of phengite. The Si content of muscovite isThe muscovite found in blocks, transition zones, andthought to increase with pressure (e.g. Massone &rinds alike shows significant tschermakitic substitutionSchreyer, 1987). Although Na substitution in phengite is(Mg,Fe2+Si6AlVIAlIV) and is thus phengitic (Table 4,inhibited by crystal-chemical effects (e.g. Guidotti, 1984;Fig. 7a). In formulae calculated on an ‘all ferrous’ basis,Evans & Patrick, 1987) the paragonite component ofFe2+ ranges from ~0·2 to 0·5 cations per 11 oxygenphengite in an Na-buffered assemblage increases withformula unit (p.f.u.), and many samples require more Si

cations than are available in order to charge-balance the temperature at P < 20 kbar [Guidotti et al. (1994) and

20

by guest on June 28, 2015http://petrology.oxfordjournals.org/

Dow

nloaded from

SORENSEN et al. PHENGITE-HOSTED LILE ENRICHMENT

Fig. 6. Plane-polarized light photomicrographs illustrating textures of ‘early’ [(a)–(d)] and ‘late’ [(e), (f )] phengite in five samples of high-gradeblocks and a transition zone. In all views, the width of the field is 2·3 mm. In (a), phengite occurs in a pressure shadow around garnet withinan eclogite layer in layered eclogite–blueschist block DR. In (b), phengite is aligned within the foliation of layered eclogite–blueschist block T-90-3A, along with elongate omphacitic clinopyroxene, sodic amphibole, and epidote grains. In (c), a coarse-grained, phengite layer in schistosetransition zone ORE-3-5 is folded and recrystallized around the fold hinge. (d) shows phengite at grain edges of calcic amphibole from garnetamphibolite of the Mt Hamilton block (MH-90). Both the Ca-amphibole and phengite are rimmed by sodic amphibole. (e) shows phengite andchlorite replacement of garnet in eclogite block GL-14-1 (the locality at Junction Schoolhouse, Healdsburg, CA). (f ) shows a coarse-grainedphengite vein that cuts block T-90-2A. gt, garnet; cp, clinopyroxene; ep, epidote; mu, phengite; nam, sodic amphibole; cam, calcic amphibole;

tt, titanite; ch, chlorite.

21

by guest on June 28, 2015http://petrology.oxfordjournals.org/

Dow

nloaded from

JOURNAL OF PETROLOGY VOLUME 38 NUMBER 1 JANUARY 1997

Tab

le4:

Ele

ctro

nm

icro

prob

ean

alys

esof

mus

covi

te

No

.N

o.

SiO

2Ti

O2

Al 2

O3

FeO

Mg

ON

a 2O

K2O

BaO

Tota

lC

atio

ns

per

11-o

xyg

enfo

rmu

la

ang

r

Sam

ple

Si

Al

IVA

lV

ITi

Mg

Fe3+

KN

aB

a

SS

85-2

7E12

4co

re50

·00·

2026

·61·

873·

360·

6810

·20·

4793

·43·

390·

611·

520·

010

0·34

0·11

0·88

0·09

00·

013

31

core

49·1

0·21

26·9

1·99

3·19

0·69

9·8

0·43

92·3

3·37

0·63

1·54

0·01

10·

330·

110·

860·

092

0·01

2{ 3

1ri

m49

·80·

1325

·52·

813·

520·

4810

·10·

7493

·13·

410·

591·

460·

007

0·36

0·16

0·88

0·06

40·

020

31

core

49·3

0·25

27·1

2·58

3·19

0·72

10·1

1·27

94·5

3·34

0·66

1·50

0·01

30·

320·

150·

870·

094

0·03

4{ 3

1ri

m50

·00·

1626

·02·

643·

390·

4510

·30·

7893

·73·

400·

601·

480·

008

0·34

0·15

0·89

0·05

90·

021

OR

E-3

-415

5co

re50

·00·

2727

·13·

323·

250·

6710

·60·

1595

·43·

340·

661·

470·

014

0·32

0·19

0·90

0·08

70·

004

{ 155

rim

49·9

0·27

27·1

3·33

3·30

0·63

10·6

0·22

95·4

3·33

0·67

1·47

0·01

40·

330·

190·

900·

082

0·00

614

5co

re49

·90·

2326

·53·

373·

390·

5110

·80·

2595

·03·

350·

651·

450·

012

0·34

0·19

0·93

0·06

60·

007

{ 145

rim

50·1

0·20

26·1

3·41

3·50

0·45

10·8

0·32

94·9

3·37

0·63

1·44

0·01

00·

350·

190·

930·

059

0·00

8D

R-B

233

11co

re50

·80·

1926

·43·

723·

920·

3310

·40·

5496

·33·

360·

641·

420·

009

0·39

0·21

0·88

0·04

20·

014

{ 3311

rim

51·7

0·11

24·5

4·05

4·38

0·17

10·5

0·63

96·0

3·43

0·57

1·35

0·00

50·

430·

230·

890·

022

0·01

6D

R-E

351

17co

re51

·40·

2025

·93·

604·

260·

3010

·60·

4896

·73·

390·

611·

400·

010

0·42

0·20

0·89

0·03

80·

012

{ 4917

rim

52·1

0·15

24·1

4·06

4·58

0·15

10·7

0·64

96·5

3·45

0·55

1·33

0·00

70·

450·

220·

900·

019

0·01

7T-

90-2

Am

atri

x12

4co

re50

·50·

1226

·92·

293·

960·

3810

·70·

4995

·33·

370·

631·

480·

006

0·39

0·13

0·91

0·04

90·

013

{ 134

rim

50·6

0·10

26·0

2·52

4·21

0·29

10·9

0·43

95·1

3·39

0·61

1·44

0·00

50·

420·

140·

930·

038

0·01

1T-

90-2

Ash

ear

zon

e18

6co

re49

·30·

2228

·51·

863·

410·

7810

·30·

6495

·03·

300·

701·

550·

011

0·34

0·10

0·88

0·10

10·

017

{ 186

rim

49·9

0·18

27·7

2·03

3·62

0·57

10·5

0·61

95·1

3·34

0·66

1·52

0·00

90·

360·

110·

900·

074

0·01

6T-

90-2

Ave

in21

6co

re50

·10·

2128

·42·

063·

540·

7210

·40·

6096

·03·

320·

681·

530·

010

0·35

0·11

0·88

0·09

20·

016

{ 216

rim

51·0

0·11

26·1

2·46

4·18

0·37

10·4

0·55

95·2

3·40

0·60

1·46

0·00

60·

420·

140·

890·

048

0·01

4T-

90-3

AE

CL

LgG

r3

1co

re49

·10·

3926

·62·

713·

060·

6810

·40·

4793

·43·

350·

651·

490·

020

0·31

0·15

0·91

0·09

00·

013

{ 31

rim

50·4

0·06

23·8

4·21

3·18

0·05

10·6

1·12

93·4

3·46

0·54

1·39

0·00

30·

330·

240·

930·

007

0·03

0T-

90-3

AE

CL

62

core

49·3

0·40

26·3

2·84

3·12

0·66

10·3

0·56

93·5

3·36

0·64

1·47

0·02

10·

320·

160·

900·

087

0·01

5{ 9

3ri

m48

·80·

4526

·72·

773·

070·

6310

·30·

5593

·33·

340·

661·

490·

023

0·31

0·16

0·90

0·08

30·

015

T-90

-3A

BS

TS

mG

r3

1co

re48

·80·

5327

·12·

842·

910·

8210

·00·

4493

·43·

320·

681·

500·

027

0·30

0·16

0·87

0·10

80·

012

{ 31

rim

48·7

0·65

27·5

2·92

2·90

0·89

10·0

0·53

94·1

3·30

0·70

1·49

0·03

30·

290·

170·

860·

117

0·01

4T-

90-3

AB

ST

93

core

49·2

0·50

27·3

2·81

2·88

0·84

10·0

0·49

94·0

3·33

0·67

1·50

0·02

50·

290·

160·

860·

110

0·01

3{ 9

3ri

m50

·20·

1825

·52·

963·

430·

3910

·60·

5893

·83·

410·

591·

450·

009

0·35

0·17

0·92

0·05

10·

015

MH

-90-

1AE

CL

31

core

49·8

0·28

24·7

2·66

3·73

0·34

10·5

1·45

93·5

3·42

0·58

1·42

0·01

40·

380·

150·

920·

045

0·03

9{ 3

1ri

m50

·50·

2623

·02·

833·

850·

2810

·40·

7391

·93·

500·

501·

380·

014

0·40

0·16

0·92

0·03

80·

020

93

core

50·1

0·26

23·8

3·09

3·89

0·28

10·7

0·80

92·9

3·45

0·55

1·38

0·01

30·

400·

180·

940·

037

0·02

2{ 9

3ri

m50

·30·

2624

·03·

283·

720·

2610

·60·

8893

·33·

450·

551·

390·

013

0·38

0·19

0·93

0·03

50·

024

MH

-90-

1AA

MP

H3

1co

re49

·10·

3425

·03·

003·

650·

3810

·70·

8693

·03·

380·

621·

410·

018

0·37

0·17

0·94

0·05

10·

023

{ 31

rim

50·8

0·14

23·1

2·99

4·07

0·11

10·6

1·33

93·1

3·50

0·50

1·37

0·00

70·

420·

170·

930·

015

0·03

69

3co

re49

·90·

1723

·63·

424·

010·

1010

·61·

3493

·13·

440·

561·

360·

009

0·41

0·20

0·93

0·01

30·

036

{ 93

rim

50·9

0·10

22·2

3·45

4·66

0·06

10·7

0·83

92·9

3·51

0·49

1·31

0·00

50·

480·

200·

940·

008

0·02

2O

RE

-3-5

186

core

49·4

0·20

25·3

3·40

3·29

0·39

10·6

0·70

93·3

3·39

0·61

1·43

0·01

00·

340·

200·

930·

052

0·01

9{ 15

5ri

m50

·00·

1524

·73·

233·

560·

3310

·60·

6693

·23·

420·

581·

420·

008

0·36

0·19

0·93

0·04

40·

018

T-90

-2B

279

core

50·2

0·24

27·7

1·81

3·60

0·75

10·4

0·60

95·3

3·35

0·65

1·52

0·01

20·

360·

100·

880·

097

0·01

6{ 26

9ri

m50

·40·

2127

·21·

863·

700·

6710

·40·

6195

·13·

370·

631·

510·

011

0·37

0·10

0·89

0·08

70·

016

MH

-90-

915

5co

re50

·00·

2824

·32·

953·

840·

2610

·61·

0693

·33·

430·

571·

400·

014

0·39

0·17

0·93

0·03

50·

029

{ 155

rim

50·5

0·22

23·5

3·09

4·12

0·18

10·7

1·02

93·3

3·47

0·53

1·37

0·01

10·

420·

180·

940·

024

0·02

7

22

by guest on June 28, 2015http://petrology.oxfordjournals.org/

Dow

nloaded from

SORENSEN et al. PHENGITE-HOSTED LILE ENRICHMENT

Fig. 7. Mineral chemistry of phengite. All iron is recalculated as ferric iron. Plotted are: Mg vs Si content (a), Na vs K (b) and Ba vs K (c).The continuous line in (a) represents an ideal Tschermak’s substitution (MgSi6AlVIAlIV); the dashed line is a least-squares fit of the data. Thecontinuous line in (b) indicates the limit of interlayer site occupancy in a stoichiometric muscovite. The symbols represent analyses from individual

blocks: Β· , SS85-27; Υ, T-90-3; Α, MH-90; Φ· , DR; Η, T-90-02; ∗, ORE-3.

references cited therein]. Some lower-Si phengite grains subduction zone metamorphic conditions (T =450–600°C; Guidotti et al., 1994).from these samples show a significant paragonite com-

ponent: Na p.f.u. ranges to ~0·12 in samples T-90-2, T- In most cases, texturally ‘late’ and retrograde phengitegrains tend to be less paragonitic than texturally ‘early’90-3, and ORE-3; this value corresponds to X ms =

~0·86. In an Na-buffered mineral assemblage (one that ones. Phengite from the Dos Rios blueschist–eclogiteblock and amphibolite–eclogite blocks T-90-2 and MH-contains sodic amphibole± omphacite) this value of X ms

is compatible with formation at moderate- to high-T 90 contains less Na at comparable Si values than does

23

by guest on June 28, 2015http://petrology.oxfordjournals.org/

Dow

nloaded from

JOURNAL OF PETROLOGY VOLUME 38 NUMBER 1 JANUARY 1997

phengite from blueschist–eclogite block T-90-3, eclogite Samana Metamorphic Complex lie along these dis-85-27, and blueschist block ORE-3 (Figs 7 and 8). In tribution lines, which indicates that phengite controls thethe last two samples, textural evidence suggests that whole-rock ratio of these elements (Fig. 11). The K/Bathe low-Na phengite is a late crystallizing phase. Some of two phengite samples from the Tiburon locality of theanalyses of phengite from the retrograde transition zone Franciscan Complex well describes this ratio in most ofaround the ORE-3 block and the rind around the MH- other whole-rock samples from this area as well (Fig.90 block are richer in Si and poorer in Na than those 11b). Phengite from one sample of the (Franciscan)from the host block. In contrast, phengite from the rind Mount Hamilton block displays the same ratio as mostaround block T-90-2, and phengite in veins that cut this whole-rock samples of other parts of the block, ir-block both show higher Na contents at similar Si contents respective of lithology. One sample from Samana (R548,compared with texturally late fine-grained phengite in Table 3; not shown in Fig. 11) displays very low Ba andthe matrix of the block; in this case, only the ‘latest’ K2O contents. It contains only small modal amountsretrograde muscovite is Na poor. of retrograde phengite and chlorite. This sample may

The Ba contents of phengite do not strongly correlate represent a ‘devolatilized’ rock that has lost LILE alongwith either Na or Si contents, or with the apparent timing with its hydrous minerals. Taken together, these ob-of crystallization (Figs 7–9). For example, block ORE-3 servations suggest that phengite is the principal hostcontains phengite with less Ba than counterparts in its for both K and Ba in most K-rich high-grade blocks.retrograde transition zone. Both the high- and low-Na Furthermore, the good correlation of whole-rock Ba, Rb,phengite in veins and in the rind around block T-90-2 and Cs values (Fig. 5) suggests that the two last elementscontain more Ba than does late-crystallizing retrograde also reside primarily in phengite.phengite in the host block (Fig. 9). Finally, in the MH-90 block, phengite from both garnet amphibolite andhigh-T, rind-like amphibolite contains more Ba than doesphengite from eclogite. The apparent lack of crystal-chemical or textural controls of Ba contents indicated by

DISCUSSIONsuch relationships suggest that if Si and Na substitutionin phengite track metamorphic P and T (respectively), Sources and sinks for K, Ba, Rb, and Cs inthen Ba probably tracks the length scales of fluid access subduction complexesto blocks and the relative timing of their alkali meta- Five principal petrologic and geochemical features aresomatism. evident from the data presented above:

(1) Some otherwise MORB-like high-grade blocks fromthe Franciscan Complex and arc-like high-grade blocksZoning within phengite grainsfrom the Samana Peninsula contain relatively largeIn many samples, cores of phengite grains tend to displayamounts of K, Ba, Rb, and Cs. However, aside frommore Na and Ti and less Si and Mg than do their rimsthis LILE enrichment, metamorphism appears to have(Table 4, Figs 7 and 10). Na and Ti appear to varybeen largely isochemical.systematically; core-to-rim zoning of individual grains is

(2) The systematics of K, Ba, Rb, and Cs distinguishroughly collinear with a trend defined by the array ofthe geochemical signatures of LILE-rich high-gradevalues for a given sample (Fig. 10). Zoning of Ti isblocks from those of low-grade Franciscan greenstoneprobably a P–T effect, because the substitutions of phen-blocks, hydrothermally altered MORB, and palagonitizedgite and paragonite components in muscovite reflectocean-floor basalt.metamorphic P–T conditions, and Ti is well correlated

(3) Both Franciscan and Samana high-grade blockswith Na content. In contrast, the Ba contents of phengitecontain ‘early’ high-Ba phengite in their high-grade blue-do not appear to vary systematically from core to rimschist and eclogite mineral assemblages. ‘Late’ high-Bawithin grains from a single thin section (Table 4). Thisphengite crystallized during retrograde metamorphismsuggests that Ba substitution in phengite is not stronglyof the blocks and rind-forming metasomatism.controlled by the P–T conditions of crystallization, or is

(4) Phengite is the host mineral for K and Ba (andmuch influenced by crystal-chemical effects.probably that for Rb and Cs) in the high-grade blocks.

(5) Because the substitution of Ba in phengite does notPhengite compositions and the mass balance of whole-rock K appear to depend on Na or Si contents, it probablyand Ba results from variations in fluid composition and dis-

tribution rather than the P–T conditions of phengiteThe K and Ba contents of phengite define four arrayscrystallization. Can these features be ascribed to theof constant K/Ba ratio in Fig. 11a–c. Whole-rock K/Ba

ratios of most samples from the Franciscan Complex and protoliths of these rocks?

24

by guest on June 28, 2015http://petrology.oxfordjournals.org/

Dow

nloaded from

SORENSEN et al. PHENGITE-HOSTED LILE ENRICHMENT

Fig. 8. Paragonite contents of phengite. All iron is recalculated as ferric iron. Analyses are grouped by host rock type, as noted in the figure.Symbols are as in Fig. 7. TZ, transition zone between block and rind.

The high-grade blocks contain far more K, Ba, Rb, these K-, Ba-, Rb-, and Cs-rich volcanic rocks also tendto be very rich in Sr, which is not a characteristic shownand Cs than do most unaltered ocean-floor rocks. Based

on comparison with the compositions of Pacific ocean- by many of the LILE-rich high-grade blocks. Based ontheir REE, HFSE, and Sr contents, it seems unlikelyfloor glasses (Melson et al., 1976) and low-grade green-

stone blocks from the Franciscan Complex (Shervais & that the otherwise MORB-like Franciscan and arc-likeSamana blocks all had alkaline igneous protoliths. TheKimbrough, 1987; MacPherson et al., 1990), the protoliths

for most high-grade blocks were probably variably altered LILE signature of enriched high-grade blocks shouldtherefore not be ascribed to igneous processes.N- and E-MORB, IAB, arc-derived sedimentary detritus,

and a minor component of off-axis seamount basalt. If the LILE-rich signature of the high-grade blocks isnot of igneous origin, how is it related to metamorphicDespite significant differences among unaltered MORB

types, they are not particularly rich in K2O, Ba, Rb, and environments or metasomatic processes? The ‘end-mem-ber’ possibilities are: (1) LILE-rich high-grade blocksCs. Relatively uncommon alkaline basaltic rocks found

in some arcs show immobile element characteristics sim- were isochemically metamorphosed from protoliths thathad been altered on the seafloor before subduction, orilar to those of calc-alkaline or tholeiitic volcanic rocks

from the same arc (e.g. Gill & Whelan, 1989; Bloomer at zeolite to prehnite–pumpellyite facies conditions duringsubduction, and (2) K, Ba, Rb, and Cs were addedet al., 1989; Lin et al., 1989; Luhr et al., 1989). However,

25

by guest on June 28, 2015http://petrology.oxfordjournals.org/

Dow

nloaded from

JOURNAL OF PETROLOGY VOLUME 38 NUMBER 1 JANUARY 1997

Fig. 9. Paragonite vs Ba contents of phengite. All iron is recalculated as ferric iron. Analyses are grouped by host rock type, as noted in thefigure. Symbols are as in Fig. 7.

to high-grade blocks by one or more fluid-mediated seen in the high-grade blocks took place on the seafloor,its presence requires that not only did we randomlymetasomatic processes that attended blueschist to LT

eclogite facies metamorphism. sample large numbers of high-grade blocks that werederived from what should be a rare type of protolithWe favor the latter explanation for the following

reasons. If in situ hydrothermal alteration commonly along hundreds of kilometers of strike of the FranciscanComplex (as well as the Samana Peninsula), but also thataffects ocean-floor rocks, one should find evidence of it

in subduction zone metamorphic rocks (e.g. Sorensen, the source of these blocks was only the uppermost fewhundred meters of the slab. This seems geologically1986; MacPherson et al., 1990). However, although Na

alteration (spilitization) of ocean-floor rocks is commonly implausible. Furthermore, the LILE systematics of high-K, palagonitized ocean-floor basalt differ from those ofobserved, pronounced K enrichment is relatively rare in

suites of altered pillow basalts (e.g. Humphris & Thomp- the high-grade blocks (Fig. 5). Although K, Rb, and Csare greatly enriched in the palagonitized rocks comparedson, 1978a,1978b; Staudigel et al., 1981; Alt et al., 1986;

Ridley et al., 1994). Moreover, such K enrichment ac- with fresh MORB, Ba is not enriched to the same degree(Fig. 5).companies palagonitization, an extremely low-T phe-

nomenon that affects only the topmost few hundred The chemical signature of seafloor alteration is presentin some low-grade Franciscan greenstone blocks (Fig.meters of ocean-floor rocks (e.g. Staudigel et al., 1981;

Alt et al., 1986; Ridley et al., 1994). If the LILE enrichment 5d–f ). However, only 6 of 43 samples have (anhydrous)

26

by guest on June 28, 2015http://petrology.oxfordjournals.org/

Dow

nloaded from

SORENSEN et al. PHENGITE-HOSTED LILE ENRICHMENT

Fig. 10. Paragonite and Ti contents of phengite. All iron is recalculatedas ferric iron. Symbols are as in Fig. 7. Arrows connect analyses ofcores (filled symbol) and rims (open symbol) of single grains. Theabbreviation ‘sm’ indicates an analysis of a relatively small grain of

phengite in a sample that also contains much larger ones.

K2O contents >1·5 wt %, and all of these contain <600p.p.m. Ba (Shervais & Kimbrough, 1987; MacPhersonet al., 1990). Like ocean-floor basalts, this populationof prehnite–pumpellyite to lawsonite blueschist faciesgreenstone blocks seems to show that K–Ba–Rb–Csenrichment is far less common than is Na enrichment(i.e. classic spilitization). In addition, the Franciscangreenstone blocks that are rich in K compared withMORB also more closely resemble palagonitized ocean-floor basalt in LILE systematics (Fig. 5).

K–Ba–Rb–Cs enrichment of high-grade blocks, trans-ition zones, or rinds can be attributed to crystallizationof phengite under garnet blueschist to LT eclogite faciesconditions, low-T blueschist retrograde metamorphism,and rind-forming metasomatism. Our data indicate thatat least some LILE enrichment of high-grade blockscan be directly linked to a type of subduction zonemetasomatism, rather than isochemical high P–T meta-morphism of previously altered basalts. These ob-servations do not preclude a multi-stage process in whichpalagonitized MORB would form phengite at low- to Fig. 11. Ba and K contents of phengite and bulk rocks. Continuous

lines are constant ratios of K/Ba within phengite. Symbols are as inhigh-T subduction zone metamorphic conditions, withFig. 2. (a) shows data for blocks, altered blocks, transition zones andphengite subsequently becoming enriched in Ba by ret-rinds from Samana and four of the six Franciscan localities. (b) shows

rograde subduction zone fluids. However, this scenario data for the Tiburon locality (T-90) of the Franciscan Complex, and(c) shows data for the Mount Hamilton block (MH-90 and GL-16) ofalso seems unlikely on geologic grounds, especially in

the Franciscan Complex. (See text for further explanation.)cases such as block T-90-2, in which retrograde phengiteis poorer in Ba than Na-rich, higher-T grains.

If high-grade blocks were enriched in K, Ba, Rb, and subduction complexes typically contain abundant meta-Cs by subduction zone metasomatism, what are the most shale or meta-argillite, which may have K2O contentslikely sources of these elements in subduction complexes? up to 3·3 wt % (Bailey et al., 1964). High-grade semipeliticThe K/Ba, Ba/Rb, and Ba/Cs ratios of high-grade metasedimentary rocks from the Catalina Schist and theblocks are not consistent with a fluid source in equilibrium Shuksan Metamorphic Complex of Misch (1966) contain

up to 2·3 wt % K2O, 2150 p.p.m. Ba, 75 p.p.m. Rb,with fresh or altered mafic or ultramafic rocks. However,

27

by guest on June 28, 2015http://petrology.oxfordjournals.org/

Dow

nloaded from

JOURNAL OF PETROLOGY VOLUME 38 NUMBER 1 JANUARY 1997

and 2·7 p.p.m. Cs (Sorensen & Grossman, 1989, 1993).Greenschist facies metasedimentary rocks dredged fromthe Puerto Rico trench ( just east of the Samana Pen-insula) contain up to 3·0 wt % K2O, 1290 p.p.m. Ba,48 p.p.m. Rb, and 1.6 p.p.m. Cs (Perfit et al., 1980b,unpublished data). Plank & Langmuir (1993) reportedbulk compositions for sediment sections near eight activetrenches. These contain 0·67–2·5 wt % K2O, 205–3250p.p.m. Ba, 20–88 p.p.m. Rb, and 1·5–4·9 p.p.m. Cs. Allof these ‘sediment’ values are remarkably similar to themaximum contents of these elements in the high-gradeblocks. Geochemical modeling shows that subducted andmetamorphosed sediment is an attractive source for K,Ba, Rb, and Cs in subduction zone fluids (e.g. Bebout &Barton, 1993). Furthermore, Sr–Nd isotopic systematicsof both Samana and Franciscan eclogites and their meta-somatic rinds provide additional evidence that a sedimentcomponent can be transferred to mafic rocks duringrind-forming metasomatism (Perfit & McCulloch, 1982;Nelson, 1991, 1995).

The behavior of Ba and TiO2 in the high-grade blocksseems to be analogous to that observed elsewhere for Band Be in that Ba appears to be very mobile and TiO2

relatively immobile. Boron and Be act as monitors ofsubduction-related metamorphism because they are frac-tionated during hydrous-fluid–rock interaction (Beboutet al., 1993). Because B is significantly more compatiblein a fluid phase than Be, metasomatic fluids should imparthigh B/Be in rocks altered by them. However, B is soeasily mobilized in low-T fluids that a geochemical ‘B-signal’ of a sediment component should become less Fig. 12. Ba/TiO2, Ba and K2O systematics in Franciscan high-graderobust at high-T conditions of metamorphism (Moran et blocks (Φ) and high-P metamorphic rocks from the Samana–Puerto

Rico Trench complex (Ε). (a) shows Ba/TiO2 vs Ba, and compares theal., 1992; Bebout et al., 1993). Micas are likely hosts formetamorphic rocks with fields for MORB and island arc volcanic rocksB as well as LILE in subducted rocks (Domanik et al.,(IAV). The MORB field includes some enriched (E-MORB) types from1993). Because Ba is evidently not as hydrophilic as B the northern East Pacific Rise [Perfit et al. (1994) and unpublished data,

under low-T conditions (You et al., 1996), it may be a 1994]. Inset (MTQ) in the IAV field shows values for Martinique (Dav-idson, 1986). Mixing curves between average MORB and phengite is asuitable geochemical monitor of fluid–rock interactionscontinuous line withΑ at 1%, 5%, and then 10% increments. (b) com-at higher-T conditions, under which most sediment- pares Ba and K2O data for metamorphic rocks (labeled as in Fig. 13a)

derived B should have already left the system. and for IAV. The latter data are as used and referenced by Donnelly &Rogers (1980), Perfit et al. (1980), Woodhead (1989), McCulloch &To evaluate whether the LILE enrichments in theGamble (1991) and McDermott et al. (1993). In each of these plots, datahigh-grade blocks result from open-system metasomaticfor the Samana–Puerto Rico suite include values for lower-grade, high-addition, we modeled their Ba and TiO2 contents using P metamorphic rocks (Perfit et al., 1980b).

simple mixing between two end-members (MORB: Ba=15 p.p.m., TiO2=1·5 wt %; average phengite: Ba=6793p.p.m., TiO2=0·24 wt %). The high-grade blocks lie 1991; Sorensen & Grossman, 1993). Mass balance con-along a trend very similar to that of the MORB–IAV siderations indicate that the LILE in these enriched garnetarray (Fig. 12a). Both the observed range of values and amphibolite blocks must reside in phengite (Sorensen &the trend of Ba/TiO2 seen in the blocks and rinds can Grossman, 1989; Sorensen, 1991). The Catalina phengitebe produced by adding 1–40 wt % phengite to a protolith contains about 3·3 Si, 0·3 Mg, and 0·09 Na p.f.u.of unaltered MORB. If altered MORB or IAB-like with all Fetotal = Fe3+ (Sorensen, 1989), a compositionprotoliths are considered, less of the phengite component resembling some reported here (Table 4). The Catalinawill effect the same result. blocks show a variety of metasomatic effects that are

attributed to fluid–rock interaction at T = 640–750°C,Phengite-hosted enrichment of K, Ba, Rb, and Cs hasalso been reported for some garnet amphibolite blocks P= 8–12 kbar (e.g. Sorensen, 1988; Sorensen & Gross-

man, 1989; Bebout & Barton, 1993). Oxygen isotopefrom the Catalina Schist of southern California (Sorensen,

28

by guest on June 28, 2015http://petrology.oxfordjournals.org/

Dow

nloaded from

SORENSEN et al. PHENGITE-HOSTED LILE ENRICHMENT

data for the Catalina amphibolite blocks indicate that for oceanic island arc basalts of the Aleutian, Marianas,the high-T metamorphism was attended by fluids in New Britain, Vanuatu, and Kermadec volcanic arcs thatequilibrium with metasedimentary rocks (Bebout & Bar- range from ~10 to 30. The average Ba content of theseton, 1989). Unlike the high-grade blocks discussed in suites is ~200 p.p.m., and Rb averages 10·7 p.p.m. Thethis paper, Catalina garnet amphibolite blocks lack a Ba/Rb ratios of Franciscan and Samana high-gradeblueschist facies overprint. Thus, fluid–rock interactions blocks also range from ~10–30 (Figs 5 and 13), althoughwhich produce phengite veins during subduction zone the blocks contain much more Ba and Rb than most arcmetamorphism can take place at temperatures even volcanic rocks.higher than those estimated for the Franciscan and Ben Othman et al. (1989) pointed out that because Ba/Samana high-grade blocks. Rb of MORB could resemble that of subducted sediment,

a comparison of Ba/Rb and Cs/Rb is a more reliableindicator of a sediment component in arc magmas. AllSamana and some Franciscan high-grade blocks lie withinImplications for the chemistry and genesisa field of Ba/Rb vs Cs/Rb that is also occupied byof arc magmasarc volcanic rocks (Fig. 13). The LILE signature of

The flux of Ba in trenches and arcs contamination by sediment-derived fluids during highMcCulloch & Gamble (1991) noted that the estimated P–T metamorphism is also distinct from that whichglobal flux of Ba into subduction zones is not balanced accompanies seafloor alteration (Fig. 13). However, theby that out of volcanic arcs. Perhaps part of the apparent Francisan samples that plot in the ‘sediment-con-‘mass balance’ problem of Ba in volcanic arcs results taminated’ field of Ben Othman et al. (1989) show im-from the presence of phengite in deeply subducted, mobile element characteristics of MORB, not arc volcanicmetamorphosed and altered basaltic rocks. On the other rocks. Subduction-related alkali alteration of metabasitehand, Plank & Langmuir (1993) have shown that local can evidently produce an ‘arc-like’ alkali signature thatoutput of Ba in specific arcs is correlated with local is decoupled from REE and HFSE abundances in thesediment input. The high-grade blocks studied here tend rock. Furthermore, the data of You et al. (1996) showto show higher Ba/K2O than do arc volcanic rocks (Fig. that increases in LILE ratios such as Ba/Rb and Cs/Rb12b). Perhaps the apparent global retention of Ba in may result simply from fractionation in moderate- tosubduction zones reflects the presence of deeply sub- high-T hydrothermal fluids of complex chemistry.ducted phengite, a phase that might serve to both store In summary, this paper shows that metasomatism andBa and buffer the local variations in Ba contents of arc consequent phengite crystallization in mafic rocks couldproducts that reflect varying sediment input. Schmidt play an important role in storing K, Ba, Rb, and Cs(1996) and Domanik & Holloway (1997) have reported derived from fluids that had previously equilibrated withnew experimental data on phengite that suggest it is sediments at moderate depths within subduction zones.stable to T=1000–1050°C, at P=95–110 kbar. As noted This may have significant ramifications for forearc vol-by these workers, phengite can potentially store LILE in canism. However, if phengite is stable at great depthsa hydrous phase at depths as great as 300 km, provided within subduction zones, this mineral is also critical tothat it is not soluble in lower-P fluids or melts. Knowing our understanding of element partitioning via de-how LILE partition into K-hollandite (the high-P break- hydration and melting beneath arcs and the com-down product of phengite) versus fluids or melts is neces- positional variations of IAB. In particular, phengitesary to understand how these LILE may contaminate breakdown might account for the association of high-Kthe upper mantle during subduction. basalts that also have unusually high contents of Ba, Rb,

and Cs, over deep segments of subduction zones, andAre sediment-derived LILE transferred to phengite via the hydrous, LILE-enriched glass inclusions found in ametasomatism? few metasomatized arc xenoliths (Schiano et al., 1995).The processes and sources of the characteristic LILEenrichment seen in IAB remain somewhat enigmatic.Our data support the enrichment of K, Ba, Rb, and Cs(particularly with respect to the REE and HFSE) owing to

CONCLUSIONSsubduction-related metasomatism. This process produces(1) Some high-grade blocks from Franciscan and Samanaeclogite that is greatly enriched in LILE compared withmelanges have been enriched in the LILE K, Ba, Rb, andMORB. In arc volcanic rocks, high Ba/Rb and Cs/RbCs via metasomatic processes that accompany subductionare thought to reflect a component of subducted sedimentzone metamorphism. Some of this enrichment ac-supplied to source regions of arc magmas (e.g. Bencompanies rind-formation, which is a well-known, sub-Othman et al., 1989; McCulloch & Gamble, 1991).

McCulloch & Gamble (1991) reported average Ba/Rb duction-related metasomatic process.

29

by guest on June 28, 2015http://petrology.oxfordjournals.org/

Dow

nloaded from

JOURNAL OF PETROLOGY VOLUME 38 NUMBER 1 JANUARY 1997

observed in IAV, which are commonly attributed to acomponent of subducted sediment, may also criticallydepend on the stability and solubility relations of phengite.Fluids that deposit phengite could not only redistributea ‘sediment signature’ of LILE within a variety of rocktypes during subduction zone metamorphism, but po-tentially store them at great depths within the eclogiticdescending slab.

ACKNOWLEDGEMENTSS.S.S. thanks the Sprague and Becker Funds of theSmithsonian Institution for support of this research, andis grateful to G. Draper for supporting her field work inthe Dominican Republic, and introducing her to thegeology of Samana. Field work by M.R.P. in Hispaniolawas supported by NSF INT-9107784, and a grant fromthe University of Florida Division of Sponsored Research.M.R.P. thanks W. Wertz for help in the field, and S. R.Taylor and P. Oswald-Sealy for laboratory support atANU. We all thank those who reviewed this manuscript:C. G. Cunningham and B. A. Morgan, III, USGS; W.G. Melson, Smithsonian Institution; and B. Harte andJ. Ryan, for the Journal of Petrology. Every one of themcontributed many valuable comments that aided us inrevising the paper.

Note added in proofWe were recently referred to an excellent summary of

Fig. 13. Ratios of Ba/Rb vs Cs/Rb for high-grade blocks, transition the crystal chemistry of Ba-rich micas (Harlow, 1995),zones, and rinds (a) and for Franciscan greenstone blocks (b). The data which reported the effective substitution mechanismare plotted on a diagram presented by Ben Othman et al. (1989). Field

Ba(Mg,Fe2+)6KAlVI for barian phengites from jadeititelabeled ‘MORB’ indicates the range of these values observed in ocean-and albitite blocks in serpentinite, Motagua Fault Zone,floor basalts. The tick-marked lines show mixing effects of sediment

contamination, calculated by Ben Othman et al. (1989). Symbols are Guatemala. Harlow (1995) concluded that this exchangeas in Figs 2 and 5. Ratios calculated from an analysis of a composite indicates growth under increasing P/T, and also re-sample of highly palagonitized MORB from DSDP Hole 417A (Stau-

marked on the common association of barian phengitedigel et al., 1981, 1995, and unpublished data, 1994) are indicated bya star. with metasomatized rock units. His SEM images reveal

barian phengite grains that display spectacular rhythmiczoning in their Ba contents, which range up to 7·26 w %BaO (0·4 Ba p.f.u.).(2) Phengite, an Si-, Mg-, and Fe-rich muscovite sta-

bilized during high P–T metamorphism, carries es-sentially all the K, Ba, Rb, and Cs in LILE-rich high-grade blocks.

REFERENCES(3) Some Franciscan high-grade blocks have Ba/TiO2,Alt, J. C., Honnorez, J., Laverne, C. & Emmermann, R., 1986.Ba/Rb, and Cs/Rb characteristics that resemble the

Hydrothermal alteration of a 1 km section through the upper oceanic‘sediment signature’ reported for island arc basalts. Incrust, Deep Sea Drilling Project Hole 504B: mineralogy, chemistrysuch blocks, these ratios are controlled by phengite. Theand the evolution of seawater–basalt interactions. Journal of Geophysicalsediment signature of the high-grade blocks appears toResearch 91, 10309–10335.

have been acquired via metasomatism that accompanied Anders, E. A. & Ebihara, M., 1982. Solar-system abundances of thesubduction zone metamorphism. elements. Geochimica et Cosmochimica Acta 46, 2363–2380.

(4) The release of the LILE K, Ba, Rb, and Cs from Anderson, R. N., Uyeda, S. & Miyashiro, A., 1976. Geophysical andgeochemical constraints at converging plate boundaries—Part I:subducted materials to the mantle wedge in the ratios

30

by guest on June 28, 2015http://petrology.oxfordjournals.org/

Dow

nloaded from

SORENSEN et al. PHENGITE-HOSTED LILE ENRICHMENT

Dehydration in the downgoing slab. Geophysical Journal of the Royal Cloos, M., 1985. Thermal evolution of convergent plate margins:thermal modeling and re-evaluation of isotopic Ar-ages for blueschistsAstronomical Society 44, 333–357.

Arculus, R. J. & Powell, R., 1986. Source component mixing in the in the Franciscan Complex of California. Tectonics 4, 421–433.Cloos, M., 1986. Blueschists in the Franciscan Complex of California:regions of arc magma generation. Journal of Geophysical Research 91,

5913–5926. petrotectonic constraints on uplift mechanisms. Geological Society of

America, Memoir 164, 77–93.Bailey, E. H., Irwin, W. P. & Jones, D. L., 1964. Franciscan andrelated rocks, and their significance in the geology of western Coleman, R. G., 1980. Tectonic inclusions in serpentinites. Archives des

Sciences, Societe de Physique et d’Histoire Naturelle de Geneve 33, 89–102.California. California Division of Mines and Geology Bulletin 183.Baldwin, S. L. & Harrison, T. M., 1989. Geochronology of blueschists Coleman, R. G. & Lanphere, M. A., 1971. Distribution and age of

high-grade blueschists, associated eclogites, and amphibolites fromfrom west–central Baja California and the timing of uplift in sub-duction complexes. Journal of Geology 97, 149–163. Oregon and California. Geological Society of America Bulletin 82, 2397–

2412.Baldwin, S. L. & Harrison, T. M., 1992. The P–T–t history ofblocks in serpentinite-matrix melange, west–central Baja California. Coleman, R. G. & Lee, D. E., 1963. Glaucophane bearing metamorphic

rocks of the Cazadero area, California. Journal of Petrology 4, 260–301.Geological Society of America Bulletin 104, 18–31.Barker, F. (with contributions from 14 others), 1994. Some accreted Coleman, R. G., Lee, D. E., Beatty, L. B. & Brannock, W. W., 1965.

Eclogites and eclogites: their similarities and differences. Geologicalvolcanic rocks of Alaska and their elemental abundances. In: Plafker,G. & Berg, H.C. (eds) The Geology of Alaska, The Geology of North Society of America Bulletin 76, 483–508.

Coombs, D. S., 1974. On the mineral facies of spilitic rocks and theirAmerica, Vol. G-1. Boulder, CO: Geological Society of America, pp.558–560. genesis. In: Amstutz, G.C. (ed.) Spilites and Spilitic Rocks. New York:

Springer Verlag, pp. 372–385.Barton, M. D., Ilchik, R. P. & Marikos, M. A., 1991. Metasomatism.In: Kerrick, D. M. (ed.) Contact Metamorphism. Mineralogical Society of Crawford, W. A., 1965. Studies in Franciscan metamorphism near

Jenner, California. Ph.D. Dissertation, University of California,America, Reviews in Mineralogy 26, 321–345.Bebout, G. E., 1991. Field-based evidence for devolatilization in sub- Berkeley.

Dalla Torre, M., De Capitani, C., Frey, M., Underwood, M. B., Mullis,duction zones: implications for arc magmatism. Science 251, 413–416.Bebout, G. E. & Barton, M. D., 1989. Fluid flow and metasomatism J. & Cox, R., 1996. Very low-temperature metamorphism of shales

from the Diablo Range, Franciscan Complex, California: new con-in a subduction zone hydrothermal system: Catalina Schist terrane,California. Geology 17, 976–980. straints on the exhumation path. Geological Society of America Bulletin

108, 578–601.Bebout, G. E. & Barton, M. D., 1993. Metasomatism during subduction:products and possible paths in the Catalina Schist, California. Davidson, J. P., 1986. Isotopic and trace element constraints on the

petrogenesis of subduction-related lavas from Martinique, LesserChemical Geology 108, 61–92.Bebout, G. E., Ryan, J. G. & Leeman, W. P., 1993. B–Be systematics Antilles. Journal of Geophysical Research 91(B6), 5943–5962.

Delany, J. M. & Helgeson, H. C., 1978. Calculation of the thermo-in subduction-related metamorphic rocks: characterization of thesubducted component. Geochimica et Cosmochimica Acta 57, 2227–2237. dynamic consequences of dehydration in subducting oceanic crust

to 100 kb and >800°C. American Journal of Science 278, 638–686.Bednarz, U. & Schmincke, H.-U., 1989. Mass transfer during sub-seafloor alteration of the upper Troodos crust (Cyprus). Contributions Domanik, K. J. & Holloway, J. R., 1997. Stability and composition of

phengitic muscovite and associated phases from 5·5 to 11 GPa:to Mineralogy and Petrology 102, 93–101.Ben Othman, D., White, W. M. & Patchett, J., 1989. The geochemistry implications for deeply subducted sediments. Geochimica et Cosmo-

chimica Acta (in press).of marine sediments, island arc magma genesis, and crust–mantlerecycling. Earth and Planetary Science Letters 94, 1–21. Domanik, K. J., Hervig, R. L. & Peacock, S. M., 1993. Beryllium

and boron in subduction zone minerals: an ion microprobe study.Bloomer, S. H., Stern, R. J., Fisk, E. & Geschwind, C. H., 1989.Shoshonitic volcanism in the northern Mariana arc 1. Mineralogic Geochimica et Cosmochimica Acta 57, 4997–5010.

Donnelly, T. & Rogers, J., 1980. Igneous series in island arcs: theand major and trace element characteristics. Journal of Geophysical

Research 94, 4469–4496. northeastern Caribbean compared with worldwide island-arc as-semblages. Bulletin Volcanologique 43, 347–382.Bocchio, R., De Capitani, L., Liborio, G., Maresch, W. V. & Mottana,

A., 1990. The eclogite-bearing series of Isla Margarita, Venezuela: Draper, G. & Lewis, J. F., 1991. Metamorphic belts in centralHispaniola. Geological Society of America, Special Papers 262, 29–45.geochemistry of metabasic lithologies in the La Rinconada and Juan

Griego Groups. Lithos 25, 55–69. Draper, G. & Nagle, F., 1988. Geologic setting and characteristics ofthe blueschist and eclogite bearing melanges in northern Hispaniola.Borg, I. Y., 1956. Glaucophane schists and eclogites near Healdsburg,

California. Geological Society of America Bulletin 67, 1563–1584. In: Transactions of the 11th Caribbean Geological Conference, Bridgetown,

Barbados. Energy Division of the Ministry of Finance of Barbados 33, 1–33.Brown, E. H. & Bradshaw, J. Y., 1979. Phase relations of pyroxene andamphibole in greenstone, blueschist, and eclogite of the Franciscan Dudley, P. O., 1972. Comments on the distribution and age of high-

grade blueschists, associated eclogites, and amphibolites from theComplex, California. Contributions to Mineralogy and Petrology 71, 67–83.BVSP (Basaltic Volcanism Study Project), 1981. Basaltic Volcanism of the Tiburon Peninsula, California. Geological Society of America Bulletin 83,

3497–3500.Terrestrial Planets. Houston, TX: Lunar and Planetary Institute.Carswell, D. A., 1990. Eclogites and the eclogite facies: definitions and Ellis, D. J. & Green, D. H., 1979. An experimental study of the

effect of Ca upon garnet–clinopyroxene Fe–Mg exchange equilibria.classifications. In: Carswell, D. A. (ed.) Eclogite Facies Rocks. Glasgow:Blackie; New York: Chapman and Hall. Contributions to Mineralogy and Petrology 71, 12–22.

Ernst, W. G., 1988. Tectonic history of subduction zones inferred fromCloos, M., 1983. Comparative study of melange matrix and metashalesfrom the Franciscan subduction complex with the basal Great Valley retrograde blueschist P–T paths. Geology 16, 1081–1084.

Ernst, W. G., Seki, Y., Onuki, H. & Gilbert, M. C., 1970. A comparativeSequence, California. Journal of Geology 91, 291–306.Cloos, M., 1984. Flow melanges and the structural evolution of ac- study of low-grade metamorphism in the California Coast Ranges

and the Outer Metamorphic Belt of Japan. Geological Society of America,cretionary wedges. Geological Society of America, Special Papers 198,71–80. Memoir 124, 276 pp.

31

by guest on June 28, 2015http://petrology.oxfordjournals.org/

Dow

nloaded from

JOURNAL OF PETROLOGY VOLUME 38 NUMBER 1 JANUARY 1997

Ertan, I. E. & Leeman, W. P., 1996. Metasomatism of Cascades subarc Joyce, J., 1985. High pressure–low temperature metamorphism andmantle: evidence from a rare phlogopite orthopyroxenite xenolith. the tectonic evolution of the Samana Peninsula, Dominican Republic,Geology 24, 451–454. Greater Antilles. Ph.D. Dissertation, Northwestern University, Ev-

Evans, B. E. & Patrick, B., 1987. Phengite 3-T in high-pressure anston, IL, 250 pp.metamorphosed granitic orthogneiss, Seward Peninsula, Alaska. Joyce, J., 1991. Blueschist metamorphism and deformation on theCanadian Mineralogist 25, 141–158. Samana Peninsula; a record of subduction and collision in the

Evans, B. E., Trommsdorf, V. & Goles, G. G., 1981. Geochemistry of Greater Antilles. Geological Society of America, Special Papers 262, 47–76.high-grade eclogites and metarodingites from the Central Alps. Kay, R. W., 1980. Volcanic arc magmas: implications of a melting–Contributions to Mineralogy and Petrology 76, 301–311. mixing model for element recycling in the crust–upper mantle

Feininger, T., 1980. Eclogite and related high-pressure regional meta- system. Journal of Geology 88, 497–522.morphic rocks from the Andes of Ecuador. Journal of Petrology 21, Kelemen, P. B., Johnson, K. T. M., Kinzler, R. J. & Irving, A. J.,107–140. 1990. High field strength element depletions in arc basalts due to

Ghent, E. D., 1989. A review of chemical zoning in eclogite garnets. mantle–magma interaction. Nature 345, 521–524.In: Smith, D.C. (ed.) Eclogites and Eclogite-facies Rocks. Developments in Krogh, E. R., 1988. The garnet–clinopyroxene Fe–Mg geo-Petrology 12. Amsterdam: Elsevier, pp. 207–231. thermometer—a reinterpretation of existing experimental data. Con-

Ghent, E. D. & Coleman, R. G., 1973. Eclogites from southwestern tributions to Mineralogy and Petrology 99, 44–48.Oregon. Geological Society of America Bulletin 84, 2471–2488. Krogh, E. R. & Brunfelt, A. O., 1981. REE, Cs, Rb, Sr and Ba

Giaramita, M. J. & Sorensen, S. S., 1994. Primary fluids in low- in glaucophane-bearing eclogites and associated rocks, Sunnfjord,temperature eclogites: evidence from two subduction complexes western Norway. Chemical Geology 33, 295–305.(Dominican Republic, and California, USA). Contributions to Mineralogy Krogh, E. J., Oh, C. W. & Liou, J. G., 1994. Polyphase and anticlockwiseand Petrology 117, 279–292. P–T evolution for Franciscan eclogites and blueschists from Jenner,

Gill, J. B., 1981. Orogenic Andesites and Plate Tectonics. Berlin: Springer California, USA. Journal of Metamorphic Geology 12, 121–134.Verlag, 390 pp. Kunugiza, K., Takasu, A. & Banno, S., 1986. The origin and meta-

Gill, J. B. & Whelan, P., 1989. Early rifting of an oceanic island arc morphic history of the ultramafic and metagabbro bodies in the(Fiji) produced shoshonitic to tholeiitic basalts. Journal of Geophysical Sanbagawa metamorphic belt. Geological Society of America, MemoirResearch 94, 4561–4578. 164, 375–385.

Glazner, A. F., 1988. Stratigraphy, structure, and potassic alterationLin, P. N., Stern, R. J. & Bloomer, S. H., 1989. Shoshonitic volcanism

of Miocene volcanic rocks in the Sleeping Beauty area, Cadyin the northern Mariana arc 2. Large-ion lithophile and rare earth

Mountains, central Mojave Desert, California. Geological Society ofelement abundances: evidence for the source of incompatible element

America Bulletin 100, 424–435.enrichments in intraoceanic arcs. Journal of Geophysical Research 94(B4),

Guidotti, C. V., 1984. Micas in Metamorphic Rocks. Mineralogical Society of4497–4514.

America, Reviews in Mineralogy 13, 357–456.Luhr, J. F., Allan, J. F., Carmichael, I. S. E., Nelson, S. A. & Hasenaka,Guidotti, C. V., Sassi, F. P., Sassi, R. & Blencoe, J., 1994. The effects

T., 1989. Primitive calc-alkaline and alkaline rock types from theof ferromagnesian components on the paragonite–muscovite solvus:western Mexican Volcanic Belt. Journal of Geophysical Research 94(B4),a semiquantitative analysis based on chemical data for natural4515–4530.paragonite–muscovite pairs. Journal of Metamorphic Geology 12, 779–

MacPherson, G. J., 1983. The Snow Mountain volcanic complex: an788.on-land seamount in the Franciscan terrain, California. Journal ofHarlow, G. E., 1995. Crystal chemistry of barian enrichment in micasGeology 91, 73–92.from metasomatized inclusions in serpentinite, Motagua Fault Zone,

MacPherson, G. J., Phipps, S. P. & Grossman, J. N., 1990. DiverseGuatemala. European Journal of Mineralogy 7, 775–789.sources for igneous blocks in Franciscan melanges, California CoastHawkesworth, C. J., Gallagher, J. M., Hergt, J. M. & McDermott, F.,Ranges. Journal of Geology 98, 845–862.1993. Mantle and slab contributions in arc magmas. Annual Reviews

Maresch, W. V. & Abraham, J., 1980. Petrography, mineralogy, andof the Earth and Planetary Sciences 21, 175–204.metamorphic evolution of an eclogite from the island of Margarita,Hermes, O. D., 1973. Paragenetic relationships in an amphiboliticVenezuela. Journal of Petrology 22, 337–362.tectonic block in the Franciscan terrain, Panoche Pass, California.

Massone, H.-J. & Schreyer, W., 1987. Phengite geobarometry basedJournal of Petrology 14, 1–32.on the limiting assemblage with K-feldspar, phlogopite, and quartz.Holway, R. S., 1904. Eclogites in California. Journal of Geology 12,Contributions to Mineralogy and Petrology 96, 212–224.344–358.

McCulloch, M. T. & Gamble, J. A., 1991. Geochemical and geo-Humphris, S. E. & Thompson, G., 1978a. Hydrothermal alterationdynamical constraints on subduction zone magmatism. Earth andof oceanic basalts by seawater. Geochimica et Cosmochimica Acta 42,Planetary Science Letters 102, 358–374.107–125.

McDermott, F., Defant, M. J., Hawkesworth, R. C., Maury, R. C. &Humphris, S. E. & Thompson, G., 1978b. Trace element mobilityJoran, J. L., 1993. Isotopic and trace element evidence for threeduring hydrothermal alteration of oceanic basalts. Geochimica et Cosmo-component mixing on the genesis of the North Luzon arc lavaschimica Acta 42, 127–136.(Philippines). Contributions to Mineralogy and Petrology 113, 9–23.Ingersoll, R. V., Schweickert, R. A., Kleist, J. R., Graham, S. A. &

Melson, W. G., Vallier, T. L., Wright, T. L., Byerly, G. & Nelen, J.,Cowan, D. S., 1984. Field guide to the Mesozoic–Cenozoic con-1976. Chemical diversity of abyssal volcanic glass erupted alongvergent margin of northern California: revised roadlogs. In: Lintz,Pacific, Atlantic, and Indian Ocean sea-floor spreading centers. TheJ. (ed.) Western Geological Excursions, Vol. 4, prepared for the 1984geophysics of the Pacific Ocean basin and its margin. GeophysicalAnnual Meetings of the Geological Society of America and AffiliatedMonograph, American Geophysical Union 19, 351–368.Societies, Reno, NV. Dept. of Geological Sciences, Mackay School

Misch, P., 1966. Tectonic evolution of the northern Cascades ofof Mines, University of Nevada, Reno, NV: pp. 304–353.Washington State. In: Tectonic History and Mineral Deposits of the WesternJarosewich, E., Nelen, J. A. & Norberg, J. A., 1978. Reference samples

for electron microprobe analysis. Geostandards Newsletter 4, 43–47. Cordillera in British Columbia and Neighbouring Parts of the United States.

32

by guest on June 28, 2015http://petrology.oxfordjournals.org/

Dow

nloaded from

SORENSEN et al. PHENGITE-HOSTED LILE ENRICHMENT

Canadian Institute of Mineralogy and Metallurgy Symposium, 1964 Annual Perfit, M. R., Nagle, F. & Bowin, C. O., 1982. Petrology and geo-Meeting, Special Volume 8, 109–111. chemistry of eclogites and blueschists from Hispaniola. Terra Cognita

Miyashiro, A. & Shido, F., 1980. Differentiation of gabbros in the Mid- 2, Abstract W7.Atlantic Ridge near 24°N. Geochemical Journal 14, 145–154. Perfit, M. R., Fornari, D. J., Bender, J. F., Langmuir, C. & Haymon,

Moore, D. E., 1984. Metamorphic history of a high-grade blueschist R. M., 1994. Small-scale spatial and temporal variations in mid-exotic block from the Franciscan Complex, California. Journal of ocean ridge crest magmatic processes. Geology 22, 375–379.Petrology 25, 126–150. Plank, T. & Langmuir, C. H., 1993. Tracing trace elements from

Moore, D. E. & Blake, M. C., Jr, 1989. New evidence for polyphase sediment input to volcanic output at subduction zones. Nature 362,metamorphism of glaucophane schist and eclogite blocks in the 739–742.Franciscan Complex, California and Oregon. Journal of Metamorphic Platt, J. B., Liou, J. G. & Page, B. M., 1976. Franciscan blueschistGeology 7, 192–211. facies metaconglomerates, Diablo Range, California. Geological Society

Moore, D. E. & Liou, J. G., 1979. Chessboard-twinned albite from of America Bulletin 87, 581–591.Franciscan metaconglomerates of the Diablo Range, California. Platt, J. P., 1975. Metamorphic and deformational processes in theAmerican Mineralogist 64, 329–336. Franciscan Complex, California: some insights from the Catalina

Moore, D. E., Liou, J. G. & King, B., 1981. Chemical modifications Schist terrane. Geological Society of America Bulletin 86, 1337–1347.accompanying blueschist facies metamorphism of Franciscan con- Ransome, F. L., 1895. On lawsonite, a new rock-forming mineral fromglomerates, Diablo Range, California. Chemical Geology 33, 237–263. the Tiburon Peninsula, Marin County, California. California University

Moore, T. E., 1986. Petrology and tectonic implications of the blue- Department of Geology Bulletin 1, 193–234.schist-bearing Puerto Nuevo melange complex, Vizcaino Peninsula, Rice, S. J., Smith, T. C. & Strand, R. G., 1976. Geology for planning,Baja California Sur, Mexico. Geological Society of America, Memoir 164, central and southeastern Marin County, California. California Division43–58. of Mines and Geology Open File Report OFR 76-2.

Moran, A. E., Sisson, V. B. & Leeman, W. P., 1992. Boron depletion Ridley, W. I., Perfit, M. R., Jonasson, R. & Smith, M. F., 1994.during progressive metamorphism: implications for subduction zone Hydrothermal alteration in oceanic ridge volcanics: a detailed studyprocesses. Earth and Planetary Science Letters 111, 331–349. at the Galapagos Fossil Hydrothermal Field. Geochimica et Cosmochimica

Morgan, B. A., III, 1970. Petrology and mineralogy of eclogite andActa 58, 2477–2494.

garnet amphibolite from Puerto Cabello, Venezuela. Journal of Pet- Roddy, M. S., Reynolds, S. J., Smith, B. M. & Ruiz, J., 1988. K-rology 11, 101–145.

metasomatism and detachment-related mineralization, HarcuvarMorris, J. D., Leeman, W. P. & Tera, F., 1990. The subducted

Mountains, Arizona. Geological Society of America Bulletin 100, 1627–component in island arc lavas: constraints from Be isotopes and

1639.B–Be systematics. Nature 344, 31–36.

Rose, A. W. & Burt, D. M., 1979. Hydrothermal alteration. In: Barnes,Nagle, F., 1974. Blueschist, eclogite, paired metamorphic belts, and

H.L. (ed.) Geochemistry of Hydrothermal Ore Deposits, 2nd edn. New York:the early tectonic history of Hispaniola. Geological Society of America

Wiley–Interscience, pp. 173–234.Bulletin 85, 1461–1466.

Ross, J. A. & Sharp, W. D., 1988. The effects of sub-blocking tem-Nelson, B. K., 1991. Sediment-derived fluids in subduction zones:perature metamorphism on K/Ar systematics of hornblendes: 40Ar/isotopic evidence from veins in blueschist and eclogite of the Fran-39Ar dating of polymetamorphic garnet amphibolite from the Fran-ciscan Complex, California. Geology 19, 1033–1037.ciscan Complex, California. Contributions to Mineralogy and PetrologyNelson, B. K., 1995. Fluid flow in subduction zones: evidence from100, 213–221.Nd- and Sr-isotope variations in metabasalts of the Franciscan

Ryan, J. G. & Langmuir, C. H., 1988. Beryllium systematics in youngComplex, California. Contributions to Mineralogy and Petrology 119,volcanic rocks: implications from 10Be. Geochimica et Cosmochimica Acta247–262.552, 237–244.Oh, C. W., 1990. The metamorphic evolution of the high-grade blocks

Schiano, P., Clocchiatti, N., Shimizu, N., Maury, R. C., Jochum,in the Franciscan Complex, California, U.S.A. Ph.D. Dissertation,K. P. & Hoffman, A. W., 1995. Hydrous, silica-rich melts in theStanford University, CA.sub-arc mantle and their relationship with erupted arc magmas.Oh, C. W. & Liou, J. G., 1990. Metamorphic evolution of two differentNature 377, 595–600.eclogites in the Franciscan Complex. Lithos 25, 41–53.

Schliestedt, M., 1990. Occurrence and stability conditions of low-Pattison, D. R. M. & Newton, R. C., 1989. Reversed experimentaltemperature eclogites. In: Carswell, D. A. (ed.) Eclogite Facies Rocks.calibration of the garnet–clinopyroxene Fe–Mg exchange geo-Glasgow: Blackie; New York: Chapman and Hall, pp. 160–179.thermometer. Contributions to Mineralogy and Petrology 101, 87–103.

Schmidt, M. W., 1996. Experimental constraints on recycling of potas-Peacock, S. M., 1993. The importance of blueschist to eclogite de-sium from subducted oceanic crust. Science 272, 1927–1930.hydration reactions in subducting oceanic crust. Geological Society of

Sedlock, R. L., 1988. Metamorphic petrology of a high-pressure, low-America Bulletin 105, 684–694.temperature subduction complex in west–central Baja California,Pearce, J. A., 1982. Trace element characteristics of lavas from de-Mexico. Journal of Metamorphic Geology 6, 205–233.structive plate boundaries. In: Thorpe, R. S. (ed.) Andesites. New

Sharp, W. D. & Ross, J. A., 1987. Are high-grade ‘knockers’ inYork: John Wiley, pp. 525–548.accretionary wedges samples of forearc basement? Evidence fromPerfit, M. R. & McCulloch, M., 1982. Nd- and Sr-isotope geochemistrythe Franciscan Complex, California. Geological Society of America,of eclogites and blueschists from the Hispaniola–Puerto Rico sub-Abstracts with Programs 19, 480.duction zone. Terra Cognita 2, Abstract I-8.

Shervais, J. W. & Kimbrough, D. L., 1987. Alkaline and transitionalPerfit, M. R., Gust, D. A., Bence, A. E., Arculus, R. J. & Taylor, S.subalkaline metabasalts in the Franciscan Complex melange, Cali-R., 1980a. Chemical characteristics of island-arc basalts: implicationsfornia. Geological Society of America, Special Papers 215, 165–182.for mantle sources. Chemical Geology 30, 227–256.

Shipboard Scientific Party 1989. Site 735. In: Robinson, P. T. & VonPerfit, M. R., Heezen, B. C., Rawson, M. & Donnelly, T. W., 1980b.Herzen, R. (eds) Proceedings of the Ocean Drilling Program, Initial Reports,Chemistry, origin and tectonic significance of metamorphic rocks

from the Puerto Rico Trench. Marine Geology 34, 125–156. 118. College Station, TX: Ocean Drilling Program, pp. 89–212.

33

by guest on June 28, 2015http://petrology.oxfordjournals.org/

Dow

nloaded from

JOURNAL OF PETROLOGY VOLUME 38 NUMBER 1 JANUARY 1997

Sorensen, S. S., 1986. Petrologic and geochemical comparison of the of altered oceanic crust: DSDP/ODP Sites 417/418. Earth and

blueschist and greenschist units of the Catalina Schist terrane, Planetary Science Letters 130, 169–185.southern California. Geological Society of America, Memoir 164, 59–75. Switzer, G., 1945. Eclogite from the California glaucophane schists.

Sorensen, S. S., 1988. Petrology of amphibolite-facies mafic and ultra- American Journal of Science 243, 1–8.mafic rocks from the Catalina Schist, southern California: meta- Tallifero, N. L., 1943. The Franciscan–Knoxville problem. Bulletin,somatism and migmatization in a subduction zone metamorphic American Association of Petroleum Geologists 27, 109–219.setting. Journal of Metamorphic Geology 6, 405–435. Tatsumi, Y., 1986. Formation of the volcanic front in subduction zones.

Sorensen, S. S., 1989. Trace element metasomatic effects of the initiation Geophysical Research Letters 1, 717–720.of subduction? Catalina Schist, southern California. Geological Society Tatsumi, Y., 1989. Migration of fluid phases and genesis of basaltof America, Abstracts with Programs 21(6), A216. magmas in subduction zones. Journal of Geophysical Research 94, 4697–

Sorensen, S. S., 1991. Petrogenetic significance of zoned allanite in 4707.garnet amphibolites from a paleo-subduction zone: Catalina Schist, Tenore-Nortrup, J. & Bebout, G. E., 1993. Metasomatism of gabbroicsouthern California. American Mineralogist 76, 589–601.

and dioritic cobbles in blueschist facies metaconglomerates: sourcesSorensen, S. S. & Grossman, J. N., 1989. Enrichment of trace elements

and sinks for high-P/T metamorphic fluids. EOS, Transactions of thein garnet amphibolites from a paleo-subduction zone: Catalina Schist,American Geophysical Union 74, 331.southern California. Geochimica et Cosmochimica Acta 53, 3155–3177.

Tera, F., Brown, L., Morris, J., Sacks, I. S., Klein, J. & Middleton, R.,Sorensen, S. S. & Grossman, J. N., 1993. Accessory minerals and1986. Sediment incorporation in island arc magmas: inferences fromsubduction zone metasomatism: a geochemical comparison of two10Be. Geochimica et Cosmochimica Acta 50, 535–550.melanges (Washington and California, U.S.A.). Chemical Geology 110,

Woodhead, J. D., 1989. Geochemistry of the Mariana Arc (Western269–297.Pacific); source composition and processes. Chemical Geology 76, 1–24.Staudigel, H., Hart, S. R. & Richardson, S. H., 1981. Alteration of

You, C.-F., Castillo, P. R., Giesko, J. M., Chan, L. H. & Spivak, A.the oceanic crust: processes and timing. Earth and Planetary ScienceJ., 1996. Trace element behavior in hydrothermal experiments:Letters 52, 311–327.implications for fluid processes at shallow depths in subductionStaudigel, H., Davies, G. R., Hart, S. R., Marchant, K. M. & Smith,

B. M., 1995. Large scale isotopic Sr, Nd and O isotopic anatomy zones. Earth and Planetary Science Letters 140, 41–52.

34

by guest on June 28, 2015http://petrology.oxfordjournals.org/

Dow

nloaded from