Geochemical Constraints on Possible Subduction Components in

Geochemical Constraints on PossibleSubduction Components in Lavas of Mayonand Taal Volcanoes, Southern Luzon,Philippines

P. R. CASTILLO1* AND C. G. NEWHALL2

1SCRIPPS INSTITUTION OF OCEANOGRAPHY, UNIVERSITY OF CALIFORNIA, SAN DIEGO, LA JOLLA, CA 92093-0212, USA

2USGS, MAILSTOP 351310, UNIVERSITY OF WASHINGTON, SEATTLE, WA 98195-1310, USA

RECEIVED AUGUST 1, 2002; ACCEPTED NOVEMBER 14, 2003

Mayon is the most active volcano along the east margin of southern

Luzon, Philippines. Petrographic and major element data indicate

that Mayon has produced a basaltic to andesitic lava series by

fractional crystallization and magma mixing. Trace element data

indicate that the parental basalts came from a heterogeneous mantle

source. The unmodified composition of the mantle wedge is similar to

that beneath the Indian Ocean. To this mantle was added a

subduction component consisting of melt from subducted pelagic

sediment and aqueous fluid dehydrated from the subducted basaltic

crust. Lavas from the highly active Taal Volcano on the west margin

of southern Luzon are compositionally more variable than Mayon

lavas. Taal lavas also originated from a mantle wedge metasoma-

tized by aqueous fluid dehydrated from the subducted basaltic crust

and melt plus fluid derived from the subducted terrigenous sediment.

More sediment is involved in the generation of Taal lavas. Lead

isotopes argue against crustal contamination. Some heterogeneity of

the unmodified mantle wedge and differences in whether the sediment

signature is transferred into the lava source through an aqueous fluid

or melt phase are needed to explain the regional compositional

variation of Philippine arc lavas.

KEY WORDS: Mayon Volcano; Philippines; sediment melt; subduction

component; Taal Volcano

INTRODUCTION

Most magmas generated along convergent marginscome from a mantle wedge that has been enriched by asubduction component consisting primarily of volatiles

(e.g. H2O, CO2) and large ion lithophile elements(LILE; e.g. Cs, Ba, Rb, Sr) (Kay, 1980; Gill, 1981;Hawkesworth et al., 1991). The subduction componentoriginates from the subducted oceanic lithosphere andoverlying sediments, but the type, composition, and pro-portion of these source materials are highly variable.Thus, the degree of geochemical enrichment exhibitedby volcanic arc lavas is also highly variable. Constrainingthe nature and specific source of the subduction com-ponent in volcanic arcs is a prime objective of manyEarth scientists, for it provides the key to a better under-standing of mass balance problems and other magmaticprocesses along convergent margins, as well as formationof mantle heterogeneities and sources of intraplate mag-mas (e.g. Kogiso et al., 1997b; Kamber & Collerson,2000). Progress in constraining the nature and source ofthe subduction component is advancing (e.g. Elliot et al.,1997; Plank & Langmuir, 1998; Class et al., 2000; Defant& Kepezhinskas, 2001; Hochstaedter et al., 2001), butthere is still a lack of consensus.In this paper, we present a detailed major element,

trace element, and Sr---Nd---Pb isotopic investigation ofMayon Volcano, a highly active arc volcanic center alongthe east margin of the southern portion of Luzon, Philippines(Fig. 1). The primary objective of our study is to constrainthe nature of the subduction component in Mayon lavas.We extend our investigation to constrain the nature of thesubduction component in the lavas from Taal Volcano,the most active center along the west margin of southernLuzon (Fig. 1). Previous studies (e.g. Knittel & Defant,

JOURNAL OF PETROLOGY VOLUME 45 NUMBER 6 PAGES 1089–1108 2004 DOI: 10.1093/petrology/egh005

*Corresponding author. Telephone: (858) 534-0383. Fax: (858) 822-

4945. E-mail: [email protected]

Journal of Petrology 45(6) # Oxford University Press 2004; all rights

reserved

Downloaded from https://academic.oup.com/petrology/article-abstract/45/6/1089/1542471by gueston 12 February 2018

1988; Defant et al., 1989; Forster et al., 1990; Mukasa et al.,1994) have indicated that a subduction component wasrecently added to the mantle wedge beneath TaalVolcano. Using the combined results for the two most

highly active volcanoes along the east and west marginsof southern Luzon, we attempt to provide a betterunderstanding of magma generation along oceanicconvergent margins.

Luzon

Mindanao

Palaw

an

Celebes Sea

SuluSea

PhilippineSea

SouthChinaSea

115 E 120 E 125 E

5 N

10 N

15 N

20 N

500 km0

Man

ila

Tr e

nc

h

Ph

i l i pp

ine

Tr e

nc

h

Borneo

Cotaba to Trench

Sulu TrenchN

egro

s Tr

ench

Leyte

(a)

(b)LuzonIsland

Min

doro

Island

MayonVolcano

Manila

TaalVolcano

SouthChinaSea

sediments

SuluSea

sediments

o

o o o

o

o

o

Fig. 1. (a) Generalized map of the Philippine archipelago showing oppositely dipping subduction zones. Gray box labeled South China Seasediments represents the approximate sites of South China Sea sediment cores RC14-90, RC17-59, RC17-156, and VM19-119 (McDermott et al.,1993); gray box labeled Sulu Sea sediment represents the approximate locations of Ocean Drilling Project Sites 769, 771 and 768 (Brass et al., 1991;Solidum, 2002). (b) Enlarged map of southern Luzon [see box in (a) for location] showing the locations of Mayon and Taal volcanoes (~) and othervolcanoes (~) along the east and west margins of the archipelago.

JOURNAL OF PETROLOGY VOLUME 45 NUMBER 6 JUNE 2004

1090Downloaded from https://academic.oup.com/petrology/article-abstract/45/6/1089/1542471by gueston 12 February 2018

GEOLOGICAL SETTING AND

SAMPLES

The east margin of the Philippine archipelago is linedwith volcanic centers that are associated with thePhilippine Trench where the Philippine Sea Plate, witha modest cover of pelagic sediment, is being subductedtoward the west (Fig. 1a). These volcanic centers belongto the east Philippine arc system and are grouped region-ally into the Bicol Arc at the northern end of thePhilippine Trench (e.g. Newhall, 1979; Knittel et al.,1988), volcanoes on Leyte Island in the center (Sajonaet al., 1994), and the East Mindanao Arc at the southernend (Sajona et al., 1997). Mayon is a highly active strato-volcano that belongs to the Bicol Arc on the east marginof southern Luzon (Fig. 1b). Although its first recordedactivity was in AD 1616, Mayon has probably been erupt-ing since the Pliocene and has continued to erupt inter-mittently up to the present. The Bicol Arc is underlain byTertiary---Quaternary sedimentary and volcanic rocksand pre-Tertiary schists, gneisses, and ultramafics.The samples analyzed in this study are a subset of

Mayon lavas previously analyzed for petrography, min-eralogy, and bulk major element chemistry by Newhall(1977, 1979) plus two newer samples from the 1984 and1993 eruptions. The samples are historic lavas eruptedfrom the main vent except for two older, primitive basalts(Lignon1 and Pac8) from two parasitic cones near thebase of the volcano. All samples were analyzed for theirmajor and trace element contents. Strontium, Nd and Pbisotopic ratios were determined on a smaller number ofsamples selected on the basis of petrographic data ofNewhall (1977, 1979) and new major and trace elementresults.

ANALYTICAL PROCEDURES

Major oxide and Ba, Co, Cr, Nb, Ni, Rb, Sr, V, Y, andZr contents were determined by X-ray fluorescence(XRF) on a wavelength-dispersive Phillips instrument atthe Scripps Institution of Oceanography (SIO). Deter-mination of major element oxides was conducted onfused disks (0�5 g sample to 2�5 g LiBO2 flux) followingthe method of Norrish & Hutton (1969). Trace elementswere measured using pressed powder pellets (3 g sampleto 1 g methyl cellulose) following the method of Norrish &Chappell (1977). Rare earth element (REE), U, Th andPb determinations were performed by inductivelycoupled plasma mass spectrometry (ICP-MS) on aFinnigan Element 2 high-resolution ICP-MS instrument.Some of the Taal samples previously analyzed for Sr,Nd and Pb isotope ratios by Mukasa et al. (1994) wereanalyzed for REE, U, Th, and Pb concentrations. ForICP-MS analyses, rock powders were digested using theprocedure described by Janney & Castillo (1996) and

then diluted in a 2�5% HNO3 solution containing 1 ppb115In as an internal standard. The accuracy and precisionof the major and trace element analyses were monitoredby repeated analysis of known rock standards and arereported as notes under Table 1.Almost all the Sr, Nd, and Pb isotope measurements

by thermal ionization mass spectrometry (TIMS) werecarried out at the Department of Terrestrial Magnetism(DTM) of the Carnegie Institution of Washington usingwell-established procedures (e.g. Walker et al., 1989;Castillo et al., 1991). About 200mg of rock powders weredissolved in Teflon beakers and then passed throughsmall HBr ion exchange columns to collect Pb. Theresidues from Pb extraction were collected and dried,and less than half (�75mg) of each was then passedthrough primary cation exchange columns to collect Srand the REE. Finally, Nd was separated from the restof the REE by passing the REE cuts through smallEDTA ion exchange columns. Pb and Sr isotope mea-surements were made on a five-collector VG 354 thermalionization mass spectrometer. Nd isotope measurementwas carried out in oxide form using a home-built, 15 inchradius mass spectrometer. A few isotope measure-ments were carried out at SIO. The sample prepara-tion procedure used at SIO is similar to that describedby Lugmair & Galer (1992) and Janney & Castillo(1996). Fractionation corrections used in both labora-tories are similar. Details of fractionation correctionsand analytical uncertainties are presented as notesunder Table 2.

RESULTS

Major element chemistry

Mayon lavas range from basalt to andesite, but most arebasaltic andesites (Table 1 and Fig. 2). The lavas belongto the medium-K, calc-alkaline arc series except for oneprimitive basalt (Pac8) that is in the high-K series. Mayonlavas overlap with those from Leyte and East Mindanaosegments of the east Philippines arc system to the south ofMayon Volcano. Some of the andesites and dacites fromthe southern arc segments belong to the low-K lava seriesand were thought by Sajona et al. (1994, 1997) to be‘adakites’ formed by partial melting of the subductedPhilippine Sea basaltic crust (PSBC) during initiation ofsubduction along the eastern margin of the Philippines.As a whole, Mayon lavas are less alkalic than lavas fromTaal Volcano, which belongs to the west Philippine arc atthe west margin of southern Luzon (Fig. 1).Except for the two primitive basalts, Mayon lavas fall

within a narrow compositional range of 54---58 wt %SiO2 (Fig. 2). This narrow bulk compositional range isreflected in a relatively constant mineral assemblage.

CASTILLO AND NEWHALL LAVAS OF MAYON AND TAAL


Mayon lavas are all porphyritic with 25---50%phenocrysts (mostly 35---45%) typically consisting ofplagioclase, augite, hypersthene, olivine, titaniferousmagnetite, and sparse hornblende set in a matrix of thesame minerals plus glass and accessory apatite. Despite

the narrow compositional range and the fact that thelavas are porphyritic, and thus were not erupted atnear-liquidus conditions, the major element oxides ofMayon lavas define linear or curvilinear trends in Harkerdiagrams with little scatter (Fig. 3).

Table 1: Major and trace element composition of Mayon lavas

1814 1881---1882 1885---1887(A) 1885---1887(B) 1885---1887(D) 1885---1887(E)

Major elements (wt %)

SiO2 54.45 52.69 51.95 52.88 53.00 52.56

TiO2 0.82 0.74 0.61 0.78 0.77 0.77

Al2O3 18.03 18.81 19.05 18.70 18.65 18.64

Fe2O3Total 8.83 8.92 8.69 8.87 8.96 8.96

MnO 0.19 0.17 0.16 0.18 0.19 0.17

MgO 4.10 4.78 4.60 4.67 4.62 4.60

CaO 8.38 9.30 9.52 9.24 9.38 9.37

Na2O 3.44 3.29 3.42 3.25 3.40 3.28

K2O 1.02 0.88 1.01 0.90 0.89 0.87

P2O5 0.29 0.22 0.24 0.25 0.26

Total 99.54 99.80 99.01 99.70 100.09 99.47

LOI 0.09 0.08 0.13 0.22

Trace elements (ppm)

Ba 347 342 298 300 293

Co 34 41 13 14 12

Cr 2 8 2 0

Nb 9 8 7 6 6

Ni 7 12 9 9 6

Rb 21 19 15 15 14

Sr 845 890 685 691 679

V 214 214 240 240 249

Y 35 30 26 26 26

Zr 106 112 82 84 87

Pb 4.37 4.29 4.10

Th 2.24 3.11 1.99

U 0.63 0.64 0.49

La 15.57 17.33 13.30 12.97 13.54

Ce 32.32 34.94 28.33 27.95 29.22

Pr 4.30 4.65 3.87 3.77 4.04

Nd 18.91 18.95 17.67 17.24 17.96

Sm 4.20 4.10 4.02 4.13 4.31

Eu 1.34 1.30 1.34 1.30 1.19

Tb 0.63 0.60 0.62 0.58 0.60

Dy 3.99 3.38 3.76 3.70 3.65

Ho 0.68 0.73

Er 2.16 2.01 1.90 2.03 2.12

Tm 0.31 0.30

Yb 2.15 2.02 2.00 2.02 2.11

Lu 0.34 0.30 0.36 0.31 0.30



Trace elements

Trace element abundances are listed in Table 1 togetherwith major element data and are shown graphically inFigs 3---5. The trace element contents of Mayon lavasare typical of most calc-alkaline lavas in that they have

enriched incompatible trace element patterns with negativehigh field strength element (HFSE) anomalies (Fig. 4).Specifically, Mayon lavas have high concentrations(20---130 times primitive mantle; Sun & McDonough,1989) of LILE (such as Rb, Ba, K) but have low contents

1897 1938 1968 1984 1993 MM6


SiO2 53.12 55.25 55.05 54.86 54.66 56.17

TiO2 0.75 0.71 0.71 0.72 0.75 0.64

Al2O3 18.40 18.25 18.29 18.07 18.78 17.91

Fe2O3Total 8.65 8.13 8.18 8.32 8.37 7.83

MnO 0.17 0.18 0.18 0.18 0.15 0.17

MgO 4.35 4.02 3.97 4.04 4.18 3.79

CaO 9.00 8.35 8.37 8.41 8.32 7.93

Na2O 3.31 3.64 3.60 3.34 3.42 3.36

K2O 1.00 1.16 1.12 1.13 1.13 1.11

P2O5 0.28 0.29 0.32 0.32 0.30 0.23

Total 99.03 99.97 99.80 99.41 100.06 99.13

LOI 0.34 0.10 0.18 0.35 0.01


Ba 311 318 343 346 353 364

Co 14 33 12 12 12 13

Cr 12 4 1 2 5

Nb 6 9 2 6 7 9

Ni 9 9 4 8 3 7

Rb 16 22 18 15 18 29

Sr 645 862 669 677 682 845

V 230 164 199 196 197 147

Y 25 33 26 28 28 32

Zr 85 138 106 94 92 123

Pb 5.88 4.80 5.16

Th 3.03 2.18 2.83

U 0.79 0.60 0.76

La 15.33 18.95 15.30 15.59 15.45 15.64

Ce 32.28 39.06 32.82 33.03 32.88 32.58

Pr 4.37 5.25 4.36 4.44 4.56 4.39

Nd 19.31 22.48 18.90 19.67 19.11 19.02

Sm 4.62 5.06 3.87 4.27 4.18 4.31

Eu 1.38 1.41 1.24 1.35 1.31 1.21

Tb 0.66 0.71 0.56 0.64 0.61 0.62

Dy 4.32 3.41 4.04 3.51 3.79

Ho 0.76 0.88 0.72 0.77

Er 2.15 2.54 2.05 2.24 2.11 2.26

Tm 0.38 0.33 0.34

Yb 2.05 2.54 1.96 2.21 2.15 2.34

Lu 0.37 0.31 0.35 0.33 0.35



(4---8 times primitive mantle) of moderately incompatibletrace elements (Y, Er, Yb, Lu). The HFSEs Nb and Ti aredepleted (3---14 times primitive mantle) relative to adja-cent elements. In many respects, Mayon lavas overlapwith Leyte lavas and to a lesser extent, with East

Mindanao Arc lavas, which extend to lower total rangesof incompatible trace elements (see Sajona et al., 1994,1997). Contents of compatible elements Ni, Cr, andV are fairly low in Mayon lavas, except in the twoprimitive basalts. The aforementioned East Mindanao

BO23 BO32 BO-7a An-1c MM11 Q8 SD38


SiO2 54.90 53.17 53.24 54.12 54.61 52.58 53.24

TiO2 0.74 0.77 0.77 0.81 0.77 0.74 0.73

Al2O3 18.09 18.55 18.58 17.90 18.09 19.03 18.52

Fe2O3Total 8.47 8.82 8.76 8.82 8.16 8.63 8.85

MnO 0.18 0.18 0.16 0.18 0.17 0.16 0.17

MgO 3.98 4.61 4.46 3.92 3.97 4.47 4.57

CaO 8.34 9.15 9.16 8.33 8.37 8.56 8.95

Na2O 3.49 3.40 3.29 3.81 3.56 3.67 3.22

K2O 1.17 0.95 0.94 1.09 1.07 1.03 0.91

P2O5 0.25 0.25 0.29 0.30 0.26 0.23 0.24

Total 99.59 99.85 99.65 99.29 99.01 99.10 99.40

LOI 0.22 0.14 0.00 0.35 0.71 0.15 0.18


Ba 419 300 323 315 297 334

Co 13 13 14 12 36 34

Cr 3 5 7 5

Nb 6 6 6 7 7 9

Ni 6 10 10 5 13 11

Rb 22 14 16 17 19 21

Sr 677 682 664 679 871 870

V 208 222 244 231 191 204

Y 24 26 27 26 33 37

Zr 101 79 90 88 116 118

Pb 6.12 5.50 4.64 4.99 3.80

Th 4.70 2.61 2.10 3.54 2.63

U 0.97 0.65 0.63 0.76 0.64

La 23.43 17.23 14.22 16.27 13.08 17.89 16.73

Ce 46.80 37.60 30.20 33.64 27.86 36.37 34.12

Pr 5.85 5.01 4.11 4.58 3.86 4.67 4.41

Nd 24.29 22.52 18.78 19.28 16.56 19.87 19.08

Sm 5.20 5.23 4.05 4.19 3.98 4.46 4.25

Eu 1.46 1.51 1.33 1.51 1.13 1.24 1.27

Tb 0.71 0.77 0.59 0.67 0.59 0.64 0.59

Dy 4.29 4.56 3.95 3.81 3.68 3.82 3.83

Ho 0.87 0.94 0.74 0.77

Er 2.55 2.64 2.17 2.05 2.17 2.25 2.09

Tm 0.37 0.38 0.33 0.32

Yb 2.56 2.64 2.04 2.12 2.19 2.18 2.12

Lu 0.36 0.39 0.32 0.33 0.32 0.32 0.33

Table 1: continued



adakites have the lowest contents of moderately incom-patible as well as compatible elements among eastPhilippine arc lavas.In detail, the behavior of individual trace elements in

Mayon lavas is fairly complex. The concentration of

incompatible trace elements shows more scatter than thatof major elements with increasing SiO2 content (Fig. 3).This behavior is also demonstrated by their REE concen-trations (Fig. 5). Mayon lavas show a light REE (LREE)-enriched pattern (La/SmN 42�0) similar to other

SV35 BO28 Mi39 An4 SD39 Lignon1 Pac8


SiO2 51.99 55.95 52.86 57.02 53.89 51.08 49.40

TiO2 0.78 0.73 0.74 0.65 0.79 0.86 0.92

Al2O3 18.75 18.42 18.47 18.26 18.84 16.62 15.95

Fe2O3Total 8.91 8.04 9.03 7.20 8.23 8.98 9.20

MnO 0.17 0.16 0.17 0.14 0.16 0.17 0.18

MgO 5.15 3.88 4.82 3.26 4.38 7.70 9.00

CaO 9.04 8.08 9.26 6.81 8.37 8.86 10.23

Na2O 3.29 3.59 3.09 4.00 3.81 3.01 2.80

K2O 0.84 1.07 0.81 1.46 1.15 1.26 1.46

P2O5 0.24 0.28 0.22 0.27 0.28 0.39 0.45

Total 99.16 100.20 99.48 99.07 99.90 98.93 99.57

LOI 0.18 0.17 0.06 0.25 0.41 0.01


Ba 274 390 326 376 318 509 475

Co 36 33 34 14 34 29 34

Cr 10 11 9 13 9 250 319

Nb 3 10 7 10 10 7 9

Ni 20 15 12 13 15 131 180

Rb 17 29 17 35 27 19 23

Sr 907 870 895 832 948 971 1017

V 213 191 181 148 203 221 243

Y 30 33 33 31 35 24 26

Zr 121 138 92 147 126 95 123

Pb 4.05 4.83 4.34 4.45 4.67 8.58 5.66

Th 1.47 2.15 3.15 2.50 2.24 5.37 6.14

U 0.44 0.71 0.66 0.87 0.65 1.43 1.45

La 11.43 13.74 17.95 16.01 14.95 29.77 28.60

Ce 25.12 29.34 36.20 33.88 32.21 57.71 59.51

Pr 3.59 3.95 4.81 4.45 4.49 7.12 7.49

Nd 15.49 17.77 19.94 18.87 18.88 29.14 29.98

Sm 3.53 3.93 4.25 4.06 4.22 6.10 5.91

Eu 1.15 1.18 1.32 1.21 1.34 1.80 1.74

Tb 0.55 0.59 0.62 0.58 0.64 0.80 0.74

Dy 3.24 4.16 3.54 3.74 3.73 4.49 4.20

Ho 0.67 0.72 0.76 0.88 0.82

Er 1.96 2.21 2.12 2.09 2.26 2.49 2.28

Tm 0.30 0.32 0.35 0.35 0.32

Yb 1.97 2.28 2.13 2.14 2.27 2.37 2.17

Lu 0.30 0.36 0.32 0.34 0.34 0.34 0.32

Replicate analyses show major element precision between 0.3% and 2.5% and are considered accurate to about 1% for Si,Ti, Al, Fe, Mg, and Ca, and to within 3---5% for Mn, Na, K, and P. For trace elements, errors are estimated at 2% for Sr and V;3% for Cr, Ni, Y, Zr and REE; 5% for Nb, U, Th, and Pb; 7% for Co and Sc; and 10% for Rb. LOI indicates the weight percentof volatiles lost on ignition at 1000�C.



Philippine mafic arc lavas (e.g. Defant et al., 1988, 1989;Miklius et al., 1991). Their REE concentration patternoverlaps with that of Taal lavas that belong to the westPhilippine arc system. However, unlike typical arc lavas,the REE contents of Mayon lavas do not exhibit acontinuous enrichment from basalts to andesites.

Sr, Nd, and Pb isotopic ratios

The isotopic ratios of representative Mayon lavas are pre-sented in Table 3 and plotted in Fig. 6a and b. The mostoutstanding isotopic feature ofMayon lavas is that the totalrange is relatively small ( 87Sr/86Sr ¼ 0�70370---0�70383;143Nd/144Nd¼ 0�51292---0�51300; 206Pb/204Pb¼ 18�54---18�57). The two primitive basalts, which have higherLREE contents than the rest of Mayon lavas, are isotopi-cally indistinguishable from the other Mayon lavas.Also shown in Fig. 6 are fields of data for Leyte and EastMindanao Arc lavas and from several arc volcaniclocalities along the west Philippines arc. As pointed out byCastillo (1996), there appears to be a systematic decrease in206Pb/204Pb and to a lesser extent an increase in 87Sr/86Sr

fromMayon to Leyte to east Mindanao (i.e. from north tosouth along the east Philippine arc). Mayon lavas havelower 87Sr/86Sr, higher 143Nd/144Nd, and lower207Pb/204Pb and 208Pb/204Pb for given 206Pb/204Pbthan lavas from Taal, Laguna de Bay, and Arayat volca-noes in the central and southern Luzon segment of the westPhilippine arc system (Mukasa et al., 1994). Compared withthe west arc volcanic lavas of the Batanes Islands in north-ern Philippines (McDermott et al., 1993; Castillo, 1996;Yang et al., 1996), Mayon lavas have higher 206Pb/204Pbbut lower 207Pb/204Pb and 208Pb/204Pb, although theyoverlap with some of these lavas with low 87Sr/86Sr andhigh 143Nd/144Nd.

DISCUSSION

Compositional variation of Mayon lavas

Newhall (1977, 1979) showed that most of the observedpetrographic and major element variation of Mayonlavas can be ascribed to crystal fractionation andmagma mixing. Compositions of Mayon lavas showing

Table 2: Sr, Nd and Pb isotopic composition of representative Mayon samples

87Sr/86Sr � 143Nd/144Nd � 206Pb/204Pb � 207Pb/204Pb � 208Pb/204Pb �

1814 0.703771 9 0.512963 8 18.573 1 15.546 1 38.462 1

1881---1882 0.703797 8 0.512923 15 18.558 1 15.528 1 38.433 1

1938 0.703783 14 0.512960 20 18.536 1 15.537 1 38.423 1

1938-leached 0.703782 19 0.512955 21

1993 0.703779 17 0.512955 9 18.541 2 15.553 2 38.553 4

Q8 0.703829 9 0.512957 21

Q8-duplicate 0.703821 10

SD38 0.703826 9 0.512972 7

SD38-duplicate 0.703834 11

SV35 0.703696 15 0.512992 22 18.560 1 15.538 1 38.421 1

Mi39 0.703807 21 0.512938 20 18.561 1 15.532 1 38.442 3

Mi39-leached 0.703875 14 0.512936 21

An4 0.703720 14 0.512974 19

SD39 0.703713 14 0.512970 15 18.550 1 15.528 1 38.660 2

Lignon1 0.703799 9 0.512982 7 18.572 3 15.540 2 38.413 8

Pac8 0.703779 12 0.512952 9 18.543 2 15.551 1 38.479 4

‘Leached’ means sample was subjected to multiple HCl-leaching procedure to mitigate alteration effects, similar to thatdescribed by Castillo et al. (1991). Leached 87Sr/86Sr and 143Nd/144Nd values are either similar to or only slightly higher thanunleached values and thus the samples are most probably not altered. Analytical uncertainty for 87Sr/86Sr measurements is�0.00002 and for 143Nd/144Nd is�0.000023; uncertainties for Pb isotopic ratios are approximately�0.007 for 206Pb/204Pb and207Pb/204Pb and �0.01 for 208Pb/204Pb. The uncertainties shown are instrument errors, and refer to the last significant digits.Sr isotopic ratios were fractionation-corrected to 86Sr/88Sr ¼ 0.1194 and normalized to 87Sr/86Sr ¼ 0.71025 for NBS 987. Ndisotopic ratios were measured in oxide form and fractionation-corrected to 146NdO/144NdO¼ 0.72225 (146Nd/144Nd¼ 0.7219)and are reported relative to 143Nd/144Nd ¼ 0.511860 for the La Jolla Nd Standard (measured La Jolla Nd values are 0.511860at DTM and 0.511862 at SIO). Pb isotopic ratios were corrected for mass fractionation based on average measured values forNBS 981 using the values of Todt et al. (1996).



repeated cycles of basaltic andesite and andesite resultfrom fractional crystallization in a shallow magma cham-ber, punctuated by periodic influxes of basaltic magmafrom depth that mix with the differentiating magma.Most of the melt passes through the chamber, but a smallportion bypasses it, forming parasitic cones (e.g. Lignon1and Pac8). New batches of parental melt mix withfractionating magma and probably trigger eruptions.The trace element concentrations of Mayon lavas,

however, are variable and do not show systematic beha-vior with increasing SiO2 content (Figs 3 and 5). Ourattempts to model the trace element concentration ofindividual lavas through simple crystal fractionation orfractional crystallization plus mixing with primitivebasalts were unsuccessful. It is possible that the traceelement variation could have resulted from assimilationand fractional crystallization (AFC) processes (e.g.DePaolo, 1981; Reagan et al., 1987). We have no samplesof the basement of the Bicol Arc, but only minimalassimilation of the basement is allowed by the limitedrange of major element and isotopic compositions aswell as by the lack of correlation between isotopic ratiosand major and trace element concentrations. Thus, sometrace element variations of Mayon lavas are probablyintrinsic to a common, compositionally heterogeneoussource of the lavas. The distinct, though small variationin Sr, Nd, and Pb isotopic ratios is consistent with thisinterpretation.

The unmodified composition ofthe Mayon sub-arc mantle

The isotopic composition of Mayon lavas is closer to theisotopic composition of normal mid-ocean ridge basalt

(N-MORB) than most Philippine arc lavas, except forsome of those in the northernmost segment of the westPhilippine arc (Mukasa et al., 1987, 1994; McDermottet al., 1993; Castillo, 1996; Yang et al., 1996). In detail, theisotopic ratios of Mayon lavas are more akin to PSBC,which has an Indian MORB-like isotopic signature(Fig. 6; Hickey-Vargas, 1991, 1998; Spadea et al., 1996).This implies that the main source of Mayon lavas has ahistory of long-term (approximately billion years) deple-tion relative to bulk Earth values of incompatible traceelements (i.e. low Rb/Sr, Nd/Sm, U/Pb, and Th/Pbratios). The Mayon sub-arc mantle, however, is appar-ently unlike those beneath many other subduction zones,which are ultra-depleted in incompatible elements (seeWoodhead et al., 1993; Elliot et al., 1997). This is shownby the generally low and constant Zr/Nb ratios (averageof 16; Fig. 7a) in Mayon lavas, which are much lowerthan the Zr/Nb ratio of the average, incompatibleelement-depleted, N-MORB (Zr/Nb ¼ 60; Sun &MacDonough, 1989). The incompatible HFSE ratiossuch as Zr/Nb are regarded as good indicators of the

CaO wt%

Fe O wt%

MgO wt%1.8

2.0

2.2

2.4

48 50 52 54 56 58

Yb ppm

15

20

25

30

Rb ppm

10

20

30

La ppm6

8

10

7

8

9

2

4

6

8

10

48 50 52 54 56 58

Total

2 3

15

25

10

35

2.6

2.8

SiO2 wt%

Fig. 3. Representative major and trace elements plotted against weightpercent SiO2 of Mayon lavas. It should be noted that the major elementsshow possible crystal fractionation trends whereas the trace elements donot show consistent linear arrays with increasing SiO2.

0

1

2

3

4

45 55 65

K O 2

SiO2

MayonLeyte E. MindanaoTaal

BasaltBasalticandesite Andesite Dacite Rhyolite

IAT

CA

HKCA

SH

Fig. 2. Plot of K2O vs SiO2 for Mayon lavas. The diagram is modifiedslightly from Peccerillo & Taylor (1976). Other Philippine arc lavasmentioned in the text (after Miklius et al., 1991; Sajona et al., 1997) areshown for comparison. SH, shoshonite; HKCA, high-K calc-alkaline;CA, calc-alkaline; IAT, island arc tholeiite.



unmodified sub-arc mantle composition because theygenerally are thought to be not modified by dehydrationfluids (e.g. Woodhead et al., 1993; Pearce & Peate, 1995;Pearce et al., 1999). The Mayon Zr/Nb values are also

close to the average value for the PSBC subductingbeneath Mayon Volcano (Zr/Nb ¼ 13; Hickey-Vargas,1998). Thus, assuming that the mantle source of PSBChas not been affected by a prior history of subduction

Phil. Sea sed.

Mayon lavas

PSBC

S. China Sea sed.

Taal lavas

PSBC

Rb Ba Th U Nb K La Ce Pb Sr Nd Sm Zr Eu Ti Tb Dy Y Er Yb Lu

Sam

ple

/ Pri

miti

ve m

antle

(a)

(b)

1

10

100

1

10

100

Fig. 4. Primitive mantle-normalized diagrams for (a) Mayon lavas and (b) Taal lavas (Miklius et al., 1991) compared with those for the bulk sedimentfrom the Philippine Sea Plate (Solidum & Castillo, 2001; Solidum, 2002), Philippine Sea basaltic crust (PSBC; Hickey-Vargas, 1998), and the fewavailable data for sediments from the South China Sea Plate (McDermott et al., 1993). For clarity, only the total range of each lava series is shown.Taal lavas do not have Nb data. Primitive mantle-normalizing values are from Sun & McDonough (1979).

10

100

Taal lavas

Pac8

Lignon1

Other Mayon lavas

La Ce Pr Nd Sm Eu Tb Dy Er Yb LuHo Tm

200

Sam

ple

/ C

hond

rite

Fig. 5. Chondrite-normalized REE concentration patterns for Mayon lavas. Chondrite-normalizing values are from Sun & McDonough (1989).



and/or that HFSE have not been added to the Mayonsub-arc mantle by subduction in a significant way,then both isotopic and HFSE ratios indicate that theunmodified Mayon sub-arc mantle probably has anIndian MORB-like composition.

It is also tempting to attribute the isotopic and Zr/Nb ratios of Mayon lavas to direct melting of thesubducted PSBC. Indeed, Sajona et al. (1993, 2000)have proposed that melts coming from the basaltic por-tion of subducted slabs play a significant role in generat-ing many Philippine arc magmas. Along the easternMindanao segment of the east Philippine arc, Sajonaet al. (1993, 1994, 1997) pointed out that some silicic(456 wt % SiO2) lavas are pure PSBC-derived melts

87 86Sr/ Sr

143N

d/ N

d14

4

0.5122

0.5124

0.5126

0.5128

0.5130

0.7028 0.7034 0.7040 0.7046 0.7052 0.7058 0.7064

15.45

15.50

15.55

15.60

15.65

15.70

38.00

38.20

38.40

38.60

38.80

18.20 18.30 18.40 18.50 18.60 18.70

Pb/

Pb20

720

4Pb

/Pb

208

204

Pb/ Pb206 204

MayonLeyteEast

TaalLagunaArayat

(a)

(b)

northern Luzon segment,

West Phil. Arc

to S. China Sea sed.

to Philippine Sea sed.

Indian0.05

0.20.3

0.5

MORB

0.02

0.05

Philippine Sea bulk sediment

PSBC

PSBC0.4

NHRL

NHRL

northern Luzon segment, West Phil. Arc

northern Luzon segment, West Phil. Arc

Mindanao

Fig. 6. Plots of (a) 87Sr/86Sr vs 143Nd/144Nd and (b) 206Pb/204Pb vs207Pb/204Pb and 208Pb/204Pb for Mayon lavas, relative to lavas fromother arc volcanoes in the Philippines (shaded field is for the northernLuzon section of the west Philippine arc). The Indian MORB87Sr/86Sr and 143Nd/144Nd values shown are for average N-MORBfrom the Indian Ocean, which have higher 87Sr/86Sr and lower 143Nd/144Nd ratios than average N-MORB from the Pacific Ocean (Whiteet al., 1987). The NHRL (Northern Hemisphere Reference Line) passesthrough the data fields for MORB from the Pacific and North AtlanticOceans and several ocean islands from the northern hemisphere (Hart,1984). Field for Indian MORB plots above NHRL particularly in the206Pb/204Pb vs 208Pb/204Pb diagram (e.g. Hart, 1984; Castillo, 1996;Hickey-Vargas, 1998). Mayon data plot within the Indian MORB field.Lines connecting bulk sediment from the Philippine Sea Plate andrepresentative analysis of PSBC represent mixing paths between thesetwo end-members. Tick marks and numbers along the mixing pathsrepresent the fraction of sediment in the mixture. Sources of data areMukasa et al. (1987, 1994), McDermott et al. (1993), Castillo (1996) andSolidum & Castillo (2001).

Mayon lavas

Mariana lavas

PS bulk sed.

PSBC

0 0.2 0.4 0.6 0.8

Th/Nb

Zr/Nb

La/Sm

U/Nb

La/Nb

(a)

(b)

(c)

(d)

MO

RB

OIB

MORB - OIB

20

60

80

100

40

1

2

3

0.1

0.2

0.3

0

2

4

6

0

(N)

Fig. 7. Plots of Th/Nb vs (a) Zr/Nb, (b) La/Sm(N), (c) U/Nb, and(d) La/Nb for Mayon lavas. Mayon lavas, similar to Mariana arc lavas,show a range of Th/Nb ratios and these correlate with some incompa-tible trace element ratios.



because of their high Sr/Y (4100) and La/Yb (47) ratiosand low Y (510) and Yb (51) values. These chemicalcharacteristics are also possessed by lavas generated bymelting of underplated basalt or lower mafic crust (e.g.Atherton & Petford, 1993; Xu et al., 2002), and some ofthese lavas have already been proposed to occur alongthe west Philippine arc (e.g. Yumul et al., 1999). Mayonlavas and particularly the primitive basalts, however, donot possess these and other criteria for melts coming froma basaltic source (see Rapp et al., 1999). It should be notedthat in Mayon lavas, the primitive basalts have thehighest La/Yb (�15) and among the highest Sr/Y(�40) ratios, so we infer that processes other than meltingof the subducted basaltic crust can produce high Sr/Yand La/Yb ratios in arc lavas (see also Atherton &Petford, 1993; Kay & Kay, 1993; Castillo et al., 1999;Yumul et al., 1999; Xu et al., 2002).

The subduction component beneath thesouthern Luzon east Philippine arc

Although the Mayon sub-arc mantle has an IndianMORB-like composition, Mayon lavas are enriched inLILE and depleted in Nb (Fig. 4). The enrichment inLILE indicates that a subduction component, derivedfrom subducted sediment and basaltic crust, is beingadded to the mantle wedge source of the arc lavas(e.g. Gill, 1981; Hawkesworth et al., 1991, and referencestherein). Interestingly, the absolute concentrations andnormalized patterns of the highly incompatible trace ele-ments of the Mayon lavas are similar to those of the bulksediments (Fig. 4). In fact, the key Zr/Nb ratios of Mayonlavas are close to the average of the bulk pelagic sedi-ments overlying the subducting PSBC (bulk sedimentZr/Nb ¼ 18; Fig. 7a). The major questions then are:(1) how is a pelagic sediment compositional signaturephysically being transferred as a subduction component?(2) Is pelagic sediment the only source of subductioncomponent in the Mayon sub-arc mantle?To constrain the possible sources of subduction com-

ponent in the Mayon sub-arc mantle, we compare theTh/Nb ratios of our Mayon data with those for lavas ofthe Mariana Arc. The ratio of Th/Nb is a good tracer ofthe source of arc lavas because both Nb and, to a certainextent, Th are relatively immobile in aqueous fluids(Brenan et al., 1995b; Elliot et al., 1997); a recent experi-mental study showed that Th can be efficiently trans-ferred from subducted sediments to the mantle sourceof arc lavas through melting ( Johnson & Plank, 1999).Mayon lavas have a wide range of Th/Nb ratios(Fig. 7a---d). Similar to Mariana arc lavas, there is littleor no correlation between Th/Nb ratios and SiO2 con-tents of Mayon lavas. In Mariana Arc lavas, Th/Nbratios show systematic correlations with other indices ofgeochemical enrichment (e.g. La/Sm, U/Nb, La/Nb,

Zr/Nb). The same general relationships are shown byMayon lavas (Fig. 7b---d). Elliot et al. (1997) proposed thatthe wide range of Th/Nb ratios in Mariana arc lavasresults from a bimodal source of the subduction com-ponent. The main subduction component involved inthe production of lavas that have the highest Th/Nbratios (i.e. lavas from Agrigan Island) is melted pelagicsediment. The subduction component responsible forlavas that have the lowest Th/Nb ratios (i.e. lavas fromGuguan Island) was postulated by Elliot et al. (1997) to beaqueous fluid dehydrated from the subducted basalticcrust. In the Izu---Bonin arc volcanic front, a relativelymore geochemically enriched aqueous fluid componentfrom the subducting basaltic crust and sediment is beingadded to the mantle wedge whereas in the backarc, anaqueous fluid dehydrated further from the residual slabis being added to a relatively more enriched mantle(Hochstaedster et al., 2001). The wide range of Th/Nbratios of Mayon lavas, therefore, by analogy suggests thatboth melt and aqueous fluid phases may be involved intransferring the subduction component into the Mayonsub-arc mantle.

Melting of sediment

The high Th/Nb end of the Mayon trace element arraysoverlaps with or plots close to the bulk sediment value(Fig. 7a---d). Because Nb is immobile in hydrothermalfluid (e.g. Brenan et al., 1995b; You et al., 1996) andthere is no compelling evidence in Mayon lavas that Thwas mobilized by hydrothermal fluid (e.g. absence ofpositive Th concentration anomaly; Fig. 4), it is unlikelythat the bulk sediment Th/Nb signature was impartedinto Mayon lavas through an aqueous fluid phase. On theother hand, although experimental results show that par-tial melts of sediments have different trace elementcompositions and ratios, Th and Nb are both effectivelymobilized from sediment above the solidus ( Johnson &Plank, 1999). Hence, the Mayon data support the ideathat melt is more effective than aqueous fluid in transfer-ring the bulk sediment signature of the subducted pelagicsediment into the source of Mayon Arc lavas.To further constrain the source of the subduction com-

ponent in the Mayon sub-arc mantle, we plot Th/Nbagainst Ba/La and La/Sm against Ba/La ratios of theMayon lavas (Fig. 8a and b). Barium is only slightly moreincompatible than La. Hence, during the normal range ofpartial melting of the mantle (�10---20%), the Ba/Laratio of the resultant magma is similar to that of its sourcematerial and only becomes slightly higher at extremelysmall degree of melting (�1%). On the other hand,experimental results show that Ba is extremely mobilein aqueous fluids during hydrothermal dehydration ofsubducted sediments (You et al., 1996) and altered basalts(e.g. Brenan et al., 1995b; Kogiso et al., 1997a). Thus ahigh Ba/La ratio is an excellent index of an aqueous fluid



contribution from the subducted sediment and basalticcrust. West Pacific chalk has extremely high Ba/La ratio(Lin, 1991), but no such sediments have been found inthe subducting oceanic crust beneath Mayon Volcano(Solidum, 2002). Figure 8a shows an overall negativecorrelation between Ba/La and Th/Nb ratios of Mayonlavas. This is opposite to what is expected if Th is mobilein aqueous fluid and requires that Th be more incompa-tible in aqueous fluid than Ba during slab dehydration(e.g. Elliot et al., 1997; Johnson & Plank, 1999). Finally,La/Sm is opposite Ba/La ratio to a certain extentbecause La is more incompatible than Sm during partialmelting and La is only slightly more mobile than Sm inhydrothermal fluid (e.g. You et al., 1996). TheLa/Sm ratio therefore should become more elevatedduring partial melting than during dehydration of mate-rials from a subducted slab. This means that a subductioncomponent carried by an aqueous fluid should have thehighest Ba/La and lowest La/Sm ratios whereas thatcarried by a melt phase should have lowest Ba/La andhighest La/Sm ratios. Figure 8b shows an overall negat-ive correlation between Ba/La and La/SmN ratios ofMayon lavas, with the low Ba/La---high La/SmN end ofthe array overlapping with the bulk sediment. This indic-ates that the Ba/La and La/SmN ratios of the subductedsediments probably are being transferred to the Mayonsub-arc mantle through a melt phase.

Dehydration of basaltic crust

The low Th/Nb end of the Mayon trace element arraystrends toward PSBC (Fig. 7a---d), the isotopic compositionof which is presumed to be that of the Mayon sub-arcmantle. Because PSBC is evidently not being melted, thelow Th/Nb signature of Mayon lavas could be comingdirectly from the sub-arc mantle, or the PSBC signatureis being transferred to the mantle through aqueous fluid,or both. The low Th/Nb end of the array, however, has ahigh Ba/La ratio, a signature of slab-derived aqueousfluid (Fig. 8a). Thus it is most likely that the lowTh/Nb ratio of some of the Mayon lavas is comingfrom fluids dehydrated from PSBC.Finally, we plot Ce/Pb against 207Pb/204Pb ratio

(Fig. 8c; Miller et al., 1994). It has been shown experi-mentally that Pb is highly mobile whereas Ce is onlyslightly mobile in hydrothermal fluids (e.g. Brenan et al.,1995a; You et al., 1996). The lead isotopic signature is notaffected by dehydration or melting. Mayon lavas have thesame Pb isotopic signature as the PSBC, indicating thattheir Pb isotopic composition is not coming from thesediment, but from either the mantle wedge or subductedPSBC (Figs 6b and 8c). Indeed, Pb isotopic data indicateonly a small bulk sediment contribution of �1%,although Sr and Nd isotopic ratios indicate a slightlyhigher sediment contribution of �5% to the Indian

Fig. 8. Plots of (a) Th/Nb vs Ba/La, (b) La/Sm(N) vs Ba/La, and(c) Ce/Pb vs 207Pb/204Pb for Mayon and Taal lavas. In all plots, line1 with arrow represents the range of composition of fluid dehydratedfrom the subducted basaltic crust. In (b) and (c), line 2 represents themixing line between Philippine Sea bulk sediment and fluid dehydratedfrom the basaltic crust that passes through Mayon lavas, and line 3represents the line defined by Taal lavas that goes to the fluid dehy-drated from the basaltic crust. In (b), Mayon lavas have a large con-tribution from Philippine Sea bulk sediment. There are no La dataavailable for the South China Sea terrigenous sediment (McDermottet al., 1993) to compare with Taal lavas. In (c), Philippine Sea bulksediment appears to have little influence on the composition of Mayonlavas whereas Taal lavas appear to have a large proportion of terrige-nous sediment. (b) and (c) indicate that Mayon and Taal lavas need‘variable amounts’ of fluid dehydrated from the subducted basaltic crustto explain their trace element and isotopic compositions. Some of theTaal data are from Miklius et al. (1991).



MORB-like sub-arc mantle (Fig. 6a and b). Mayon lavas,however, have much lower Ce/Pb ratios than the PSBCand mantle wedge, so the bulk of their Pb content musthave been added through hydrothermal fluids comingfrom PSBC.A possible reason why the Pb dehydrated from the Pb-

enriched West Philippine Sea sediment was not trans-ferred into the mantle source of Mayon lavas is its highmobility in hydrothermal fluids. The bulk of the Pb con-tent of the subducted sediment may have been lost earlyat shallower depths in the subduction zone (You et al.,1996). The fact that the Pb content of Mayon lavas ismuch lower than that of sediments (Fig. 4) is consistentwith this interpretation. This also implies that the result ofmass balance calculation based on Pb isotopic composi-tions and concentrations in sediments and mantle wedge(Fig. 6b) should be considered a minimum, as Pb is not aconservative element during sediment subduction.In summary, the combined isotopic and trace element

data suggest that the mantle wedge beneath MayonVolcano is enriched by a subduction component. Thesubduction component comes from partial melting ofthe subducted pelagic sediment and from dehydrationof the subducted PSBC (see also Plank & Langmuir,1998). Surprising in this light, a preliminary 10Be resultfor a Mayon lava failed to detect the sediment input(Morris & Tera, 1989).

Implication for HFSE contents of arc lavas

Another interesting implication of our results concernsthe suggested role of a titanate phase in creating negativeHFSE anomalies in arc lavas (e.g. Green, 1981; Morriset al., 1990). Lavas generated from the Mayon sub-arcmantle mixed with bulk sediment-derived melt do notneed residual titanate phase in the mantle wedge, as thesediment itself has negative HFSE anomalies. A similarconclusion was arrived at by Elliot et al. (1997) regardingsome of the Mariana Arc lavas. When the proportion ofsediment contribution decreases and the amount ofPSBC-hydrated aqueous fluid increases, the negativeHFSE anomaly diminishes. Such positive correlationbetween sediment contribution and HFSE depletionclearly indicates that, at least in the present case, a titan-ate phase is not necessary to create negative HFSEanomalies.

The subduction component beneath TaalVolcano, revisited

Taal Volcano erupts a large volume of lavas in the south-ern Luzon segment of the west Philippine arc (Fig. 1) andis also one of the better-studied arc volcanic systems in thePhilippines (e.g. Miklius et al., 1991; Mukasa et al., 1994;Knittel & Yang, 1998). The modern Taal Volcano is a

remnant of an eruptive complex occupying a massivevolcano-tectonic depression, now occupied by TaalLake, resulting from multiple phases of collapse (Mikliuset al., 1991). The lava flows and pyroclastic materials fromseveral of the eruption centers produced during differentphases of volcanic activity in the complex as well as fromthe older NW margin of the volcanic depression rangefrom basalt to dacite. Despite the wider range in composi-tion (Fig. 2), the Taal lava series are similar to Mayonlavas in that they fall along trends defined by crystalfractionation and magma mixing (Miklius et al., 1991).Could Taal lavas be affected by crustal contamination?

SiO2 and MgO are not correlated with isotopic ratios,particularly the Pb isotopes, which show the most sys-tematic variation (Fig. 6b). Moreover, petrological investi-gations along the Luzon Arc show that lavas containingthe largest subduction components (i.e. those with thehighest 87Sr/86Sr ratios and the most enriched in LILE)are those that are most primitive (i.e. have the highestMgO, Cr, and Ni contents, and hence are the leastcontaminated). Examples include several rock suites inthe Taiwan segment of the Luzon Arc (McDermott et al.,1993) and the Mt. Pinatubo rock suite (Castillo &Punongbayan, 1996). Karig (1983) found no evidencethat the Luzon Arc is underlain by continental crust,and the nearby Palawan continental block (Cardwellet al., 1980; Defant et al., 1988) has 206Pb/204Pb and207Pb/204Pb ratios (Tu et al., 1992) that are too low tobe the appropriate high 207Pb/204Pb contaminant for theTaal lavas. Finally, the high 3He/4He ratios reported forTaal geothermal gases (7�5 Ra; Poreda & Craig, 1989) donot support significant crustal contributions of helium tothe Taal lava sources. Thus, the highly active magmasupply to Taal does not appear to melt and assimilatecrust through which it rises.Instead, we think the compositional heterogeneity of

Taal lavas is due to a common, although heterogeneousmantle source (Miklius et al., 1991; Mukasa et al., 1994;Knittel & Yang, 1998). Strontium and Nd isotopic ratiosindicate that the Taal sub-arc mantle has had long-termdepletion in incompatible trace elements, but the lavasare enriched in incompatible trace elements (i.e. highRb/Sr and Nd/Sm ratios; Mukasa et al., 1994). Thisdecoupling between isotopic and elemental ratios wasinterpreted as due to recent addition to the mantlewedge of subduction component, consisting mainly ofmaterials from subducted sediments (see also Knittelet al., 1988).One of the best arguments for the sedimentary origin of

the subduction component in the Taal sub-arc mantle isPb isotopic data. Taal lavas have variable 207Pb/204Pband 208Pb/204Pb for given 206Pb/204Pb ratios, forminglinear arrays that trend from the low 207Pb/204Pb and208Pb/204Pb values of lithospheric mantle toward thehigh 207Pb/204Pb and 208Pb/204Pb signatures of oceanic



sediment (Fig. 6b). Taal lavas also have lower143Nd/144Nd but higher 87Sr/86Sr than Mayon lavas,closer to the sediment Nd and Sr isotopic signature(Fig. 6a). A significant sediment contribution to the gen-esis of Taal lavas is consistent with the thick terrigenoussediment cover of the South China Sea basaltic crust andabsence of sedimentary wedge along the Manila Trench(Silver & Rangin, 1991).There is no information regarding sediment subduct-

ing directly beneath Taal Volcano, but there are limiteddata available for sediments currently subductingbeneath the northern segment of the west Philippine arc(Fig. 1; McDermott et al., 1993). These South China Seasediments are mainly derived from continental Eurasiaand are similar to Sulu Sea sediments sampled duringOcean Drilling Program Leg 124 (e.g. Brass et al., 1991;Solidum & Castillo, 2001; Solidum, 2002). The composi-tion of the South China Sea terrigenous sedimentsis distinct from that of the pelagic sediments in thePhilippine Sea being subducted beneath MayonVolcano. This is clearly shown by the relatively higher

Rb and Th but lower Sr contents of the South China Seathan Philippine Sea sediments, which are respectivelyreflected by Taal and Mayon lava compositions (Figs 4and 9). Thus, it appears that the main compositionaldifference between Mayon and Taal lavas can be tracedto the different types of sediments being subducted oneither side of southern Luzon.We proposed in the previous section that the sediment

compositional signature is being transferred to the sourceof Mayon lavas mainly through a melt phase. On theother hand, Knittel & Yang (1998) proposed that anaqueous fluid phase is responsible for the transfer ofsediment component into the Taal sub-arc mantle. Oneclue to the composition of the unmodified Taal sub-arcmantle is that in the sediment discrimination diagrams(Fig. 9a---d), Taal lava compositional trends point towardPSBC. Previous studies have also shown that the Taalsub-arc mantle, like the Mayon sub-arc mantle, is not asdepleted in highly incompatible trace elements as manyother sub-arc mantles (e.g. Miklius et al., 1991; Knittel &Yang, 1998). Thus, we assume that both Taal and

200

400

600

800

250

500

750

1000

0

10

20

30

Phil. Sea sed.Taal lavasMayon lavas

Ba

Sr

Nd

Th

Ba

Sm

Rb0 5 10 15 20

250

500

750

250

500

750

1000

2

4

6

8

0 50 100 150

Sr

S. China Sea sed.

PSBC

(a) (b)

00

Fig. 9. Plots of (a) Th vs Ba, Sr, and Nd and (b) Rb vs Ba, Sr, and Sm for Mayon and Taal lavas. Both lava suites define distinct linear arrays thatoriginate from PSBC and/or mantle wedge. The Mayon array either overlaps with or trends toward the value for Philippine Sea bulk sedimentwhereas the Taal array either overlaps with or trends toward the majority of the values for South China Sea sediments. Additional data are fromMiklius et al. (1991), McDermott et al. (1993) and Hickey-Vargas (1998).



Mayon sub-arc mantles have similar, Indian MORB-likecompositions.There are no available Nb data for Taal arc lavas to

directly compare their Th/Nb ratios with those of Mayonand Mariana arc lavas. Nevertheless, Taal lavas formlinear arrays in both Ba/La vs La/SmN and Ce/Pb vs207Pb/204Pb plots (Fig. 8b and c), indicating that, similarto the Mayon sub-arc mantle, the subduction componentin the Taal sub-arc mantle is also coming from subductedsediment and basaltic crust. Sediment contribution ismore significant in Taal sub-arc mantle than in Mayonsub-arc mantle. This is clearly shown in the subverticalarray of Taal lavas in the Ce/Pb vs 207Pb/204Pb plot(Fig. 8c). The high 207Pb/204Pb end of the array pointsto the high 207Pb/204Pb signature of terrigenous sedimentwhereas the low 207Pb/204Pb end points to the aqueousfluid signature of the subducted basaltic crust (see Milleret al., 1994; Brenan et al., 1995a). In the Ba/La vs La/SmN plot (Fig. 8b), Taal lavas have a fairly limited andslightly higher Ba/La ratio than both the PSBC andWestPhilippine Sea bulk terrigenous sediment. This is consist-ent with the derivation of the low La/SmN---high Ba/Laend of the array from fluids derived from the subductedbasaltic crust. The high La/SmN---high Ba/La end of thearray, on the other hand, indicates that the Ba/La sig-nature of subducted sediments is being transferred toTaal lavas either through a small-degree melt or an aque-ous fluid phase, or both. Further evaluation of whetherthe terrigenous sediment signature is being transferred toTaal lavas through either melt or aqueous fluid phase,however, will have to wait until analyses of terrigenoussediments subducting beneath Taal Volcano havebecome available.In summary, as at Mayon, the mantle wedge beneath

Taal Volcano is being enriched by a subduction com-ponent. At both volcanoes, the subduction componentscome from subducted sediment and basaltic crust. Itappears that in both cases the basaltic crust componentis being transferred in an aqueous fluid phase. The sedi-ment component in Mayon is being transferred through amelt phase whereas at Taal it is being transferred eitheras a small-degree melt or through an aqueous fluid, orboth. The main difference between the two settings is thedifference in the composition of the pelagic and terrigen-ous sediments being subducted beneath the east andwest Philippine arcs (Karig et al., 1975; Rangin et al.,1990; Solidum, 2002). More sediment is also involved inthe generation of the subduction component beneathTaal Volcano than beneath Mayon Volcano. TheSouth China Sea Basin being subducted eastward alongthe Manila Trench has moderate to thick cover of terri-genous sediments (Cardwell et al., 1980; Taylor & Hayes,1983; Rangin et al., 1990; Silver & Rangin, 1991;Solidum, 2002) and delivers a relatively large volume ofterrigenous sedimentary component into the mantle

beneath Taal Volcano. In contrast, the Philippine SeaPlate being subducted westward along the PhilippineTrench has a thin cover of pelagic sediments (Karig et al.,1975; Cardwell et al., 1980; Silver & Rangin, 1991;Solidum, 2002). This thin sediment cover of the subduc-ting slab limits the amount of pelagic sediment in themantle beneath the east margin of southern Luzon.

IMPLICATIONS FOR THE REGIONAL

VARIATION OF PHILIPPINE ARC

LAVAS

Lavas erupted in various segments of the east and westPhilippine arcs are diverse in their chemical and isotopiccompositions (e.g. Mukasa et al., 1987, 1994; McDermottet al., 1993; Castillo, 1996), and the origin of the variationis controversial. A few investigators call for variations inthe amount of melt from subducted basaltic crust toexplain some of the compositional variabilities (Sajonaet al., 2000), but others believe that the entire Philippinearc setting is underlain by a common, geochemicallydepleted mantle source and that the arc compositionalvariation results from variable type and proportion ofsediments being subducted beneath the different segmentsof Philippine arcs (e.g. Defant et al., 1989; McDermottet al., 1993). For example, the Sr isotopic ratios along thecentral Luzon segment of the west Philippine arc displaylatitudinal variation starting from a low, almost MORB-like value (0�7030) in the north to a high of 0�7047 in TaalVolcano in the south (Figs 1 and 6). Defant et al. (1989)suggested that such variation is due to increasing amountsof continental sediment subducting from north to southalong the Manila Trench. Subvertical trends in 207Pb/204Pb and 208Pb/204Pb vs 206Pb/204Pb plots are also acharacteristic feature of lavas from different segments ofthe Philippine arc systems, and McDermott et al. (1993)have proposed that this feature is also the result of addi-tion of sediments with variable isotopic compositions towhat might otherwise be an isotopically homogeneousmantle wedge.Still other investigators have proposed that the composi-

tional variation in the lavas is a combined effect of thevariable composition of the unmodified mantle wedgeand later added components from subducted slab(e.g. Chen et al., 1990; Castillo, 1996). The Philippinearchipelago is an amalgamation of different tectonicterranes from different locations (e.g. Hamilton, 1979;Hall, 1996), and hence it is possible that these terranescarried with them portions of the lithospheric mantle anddeep mafic lower crust from previous locations. Along theeast Philippine arc, for example, there is also distinctlatitudinal variation in Sr and Pb isotopic compositionsalthough there is neither continental sediment nordemonstrated compositional variation in the sediment



being subducted along the Philippine Trench (Castillo,1996).Our results show that sediments play a major role in

generating the overall arc compositional signature ofMayon and Taal arc lavas. Thus, the results favor vari-ation in the type and amount of subducting sedimentsas the major factor behind the regional variation ofPhilippine arc lavas. However, we do not yet knowwhether the sediment signature is transferred throughmelt in all Philippine arcs. Neither do we know whetherdehydration or melting of the basaltic crust plays a majorrole in generating compositional variation (e.g. White &Dupre, 1986; McDermott et al., 1993; Miller et al., 1994;Elliot et al., 1997). Long-lived, but small-scale, hetero-geneities in the mantle wedge (e.g. Morris & Hart,1983; Wallace & Carmichael, 1999; Castillo et al., 2002)might also add to regional compositional variability.Finally, the subparallel, subvertical trends in 207Pb/204Pb and 208Pb/204Pb vs 206Pb/204Pb plots are incon-sistent with a homogeneous mantle source for these lavas.These trends must be due to fairly recent addition ofPb from subduction components to a mantle wedgeoriginally variable in 206Pb/204Pb (Castillo, 1996).

SUMMARY AND CONCLUSIONS

Lavas from Mayon Volcano in southeastern Luzon,Philippines, are compositionally variable. A part of thevariation can be ascribed to shallow-level fractionalcrystallization and magma mixing, but the other part isdue to compositional variability of the source. Combinedtrace element and isotopic data show that the composi-tion of the unmodified mantle wedge beneath MayonVolcano has an Indian MORB-like composition. Tothis mantle wedge is added a small amount of subductioncomponent consisting of a few percent partial melt ofthe subducted pelagic sediment and aqueous fluiddehydrated from the subducted PSBC. Interestingly, thedepletion of HFSE in Mayon lavas could be inheritedfrom the sediment HFSE depletion, and hence Mayonlavas do not require the formation of a residual titanatephase in their source.Previous studies have also shown that the composition

of lavas erupted from Taal Volcano on the western side ofsouthern Luzon resulted from sediment addition to itsmantle wedge source. Compared with Mayon lavas, Taallavas have a much wider range of composition. The widecompositional range is most probably due to the largeramount of subduction component being added to themantle beneath Taal. Combined trace element and iso-topic data suggest that the subduction component in theTaal sub-arc mantle is similar to that in the Mayon sub-arc mantle, coming from both the subducted basalticcrust and sediment. The contribution from basalticcrust is transferred in the form of an aqueous fluid

phase; sediment contribution is transferred in either amelt or a fluid phase, or in both.Our results suggest that sediment input exerts a major

control on the composition of arc lavas on both sides ofsouthern Luzon. This brings into question the origin ofthe observed regional variation of arc lavas in the entirePhilippine arc systems. Our results favor the idea thatthe type and amount of sediment can account for mostobserved variation. However, factors such as the composi-tion of the unmodified mantle wedge and the processesby which the sediment signature is being transferred tothe lava source (i.e. whether through an aqueous fluid ormelt phase) may also add to generating compositionalvariations of arc lavas.

ACKNOWLEDGEMENTS

We are grateful to R. Solidum for his help in the analyt-ical work, comments, and suggestions; to M. Flower forproviding the Taal samples; and to C. MacIsaac, theDTM staff, and the SIO Analytical Facility for the useof analytical laboratory and instruments. We also muchappreciate the thorough reviews of W. Hildreth, J. Ryan,T. Sisson, and two anonymous reviewers and editorialhandling of D. Geist, which significantly improved themanuscript. This work is supported by NSF grantEAR00-01212 and Carnegie Institution of WashingtonPostdoctoral Fellowship to P.C.

REFERENCESAtherton, M. P. & Petford, N. (1993). Generation of sodium-rich

magmas from newly underplated basaltic crust. Nature 362,

144---146.

Brass, G., Sims, D., Calvert, S. E. & Breymann, M. T. (1991). Data

report: major- and minor-element analysis of sediments from Sites

767, 768 and 769. In: Silver, E., Rangin, C., et al. (eds) Proceedings of

the Ocean Drilling Project, Scientific Results, 124. College Station, TX:

Ocean Drilling Program, pp. 531---539.

Brenan, J. M., Shaw, H. F. & Ryerson, F. J. (1995a). Experimental

evidence for the origin of lead enrichment in convergent-margin

magmas. Nature 378, 54---56.

Brenan, J. M., Shaw, H. F., Ryerson, F. J. & Phinney, D. L. (1995b).

Mineral---aqueous fluid partitioning of trace elements at 900�C and

2�0GPa: constraints on the trace element chemistry of mantle and

deep crustal fluids. Geochimica et Cosmochimica Acta 59, 3331---3350.

Cardwell, R. K., Isacks, B. L. & Karig, D. E. (1980). The spatial

distribution of earthquakes, focal mechanism solutions, and

subducted lithosphere in the Philippine and northeastern Indonesian

islands. In Hayes, D. E. (ed.) The Tectonic and Geologic Evolution of

Southeast Asian Seas and Islands. Geophysical Monograph, American

Geophysical Union 23, 1---35.

Castillo, P. R. (1996). The origin and geodynamic implication of the

Dupal isotopic anomaly in volcanic rocks from the Philippine island

arcs. Geology 24, 271---274.

Castillo, P. R. & Punongbayan, R. S. (1996). Petrology and Sr, Nd and

Pb isotopic geochemistry of Mt. Pinatubo volcanic rocks. In:

Newhall, C. G. & Punongbayan, R. S. (eds) Fire and Mud: Eruptions



and Lahars of Mount Pinatubo, Philippines. Quezon City and

Seattle: PHIVOLCS and University of Washington Press,

pp. 799---805.

Castillo, P. R., Carlson, R. W. & Batiza, R. (1991). Origin of Nauru

Basin igneous complex: Sr, Nd and Pb isotope and REE constraints.

Earth and Planetary Science Letters 103, 200---213.

Castillo, P. R., Janney, P. E. & Solidum, R. U. (1999). Petrology and

geochemistry of Camiguin Island, southern Philippines: insights to

the source of adakite and other lavas in a complex arc tectonic

setting. Contributions to Mineralogy and Petrology 134, 33---51.

Castillo, P. R., Solidum, R. U. & Punongbayan, R. S. (2002). Origin of

high field strength element enrichment in the Sulu Arc, southern

Philippines, revisited. Geology 30, 707---710.

Chen, C.-H., Shieh, U.-N., Lee, T., Chen, C.-H. & Mertzman, S. A.

(1990). Nd---Sr---O isotopic evidence for source contamination and

unusual component under Luzon arc. Geochimica et Cosmochimica Acta

54, 2473---2483.

Class, C., Miller, D., Goldstein, S. & Langmuir, C. (2000).

Distinguishing melt and fluid subduction and components in Umnak

volcanics, Aleutian Arc. Geochemistry, Geophysics, Geosystems 1, paper

number 1999GC000010.

Defant, M. J. & Kepezhinskas, P. (2001). Evidence suggests slab

melting in arc magmas. EOS Transactions, American Geophysical Union

82, 67---70.

Defant, M. J., De Boer, J. Z. & Oles, D. (1988). The western Central

Luzon volcanic arc, the Philippines: two arcs divided by rifting?

Tectonophysics 145, 305---317.Defant, M. J., Jacques, D., Maury, R. C., De Boer, J. & Joron, J. (1989).

Geochemistry and tectonic setting of the Luzon arc, Philippines.

Geological Society of America Bulletin 101, 635---672.

DePaolo, D. J. (1981). Trace element and isotopic effects of combined

wall rock assimilation and fractional crystallization. Earth and

Planetary Science Letters 53, 189---202.

Elliot, T., Plank, T., Zindler, A., White, W. & Bourdon, B. (1997).

Element transport from slab to volcanic front at the Mariana arc.

Journal of Geophysical Research 102, 14991---15019.

Forster, H., Oles, D., Knittel, U., Defant, M. J. & Torres, R. C. (1990).

The Macolod Corridor: a rift crossing the Philippine island arc.

Tectonophysics 183, 265---271.

Gill, J. B. (1981). Orogenic Andesites and Plate Tectonics. New York:

Springer.

Green, T. H. (1981). Experimental evidence for the role of accessory

phases in magma genesis. Journal of Volcanology and Geothermal Research

10, 405---422.

Hall, R. (1996). Reconstruction of Cenozoic SE Asia. In: Hall, R. &

Blundell, D. J. (eds) Tectonic Evolution of Southeast Asia. Geological Society,

London, Special Publications 106, 153---184.

Hamilton, W. (1979). Tectonics of the Indonesian Region. US Geological Survey

Professional Paper 1078.

Hart, S. R. (1984). The Dupal anomaly: a large-scale isotopic anomaly

in the southern hemisphere. Nature 309, 753---756.

Hawkesworth, C. J., Hergt, J. M., Ellam, R. M. & McDermott, F.

(1991). Element fluxes associated with subduction related magma-

tism. Philosophical Transactions of the Royal Society of London, Series A 335,

393---405.

Hickey-Vargas, R. (1991). Isotope characteristics of submarine lavas

from the Philippine Sea: implications for the origin of arc and basin

magmas of the Philippine tectonic plate. Earth and Planetary Science

Letters 107, 290---304.

Hickey-Vargas, R. (1998). Origin of the Indian Ocean-type isotopic

signature in basalts from Philippine Sea plate spreading centers; an

assessment of local versus large-scale processes. Journal of Geophysical

Research 103, 20963---20979.

Hochstaedter, A., Gill, J., Peters, R., Broughton, P., Holden, P. &

Taylor, B. (2001). Across-arc geochemical trends in the Izu---Bonin

arc; contributions from the subducting slab. Geochemistry, Geophysics,

Geosystems 2, paper number 2000GC000105.

Janney, P. E. & Castillo, P. R. (1996). Basalts from the Central Pacific

Basin: evidence for the origin of Cretaceous igneous complexes in

the Jurassic Western Pacific. Journal of Geophysical Research 101,

2875---2894.

Johnson, M. C. & Plank, T. (1999). Dehydration and melting

experiments constrain the fate of subducted sediments. Geochemistry,

Geophysics, Geosystems 1, paper number 1999GC000014.

Kamber, B. S. & Collerson, K. D. (2000). Role of ‘hidden’

deeply subducted slabs in mantle depletion. Chemical Geology 166,

241---254.

Karig, D. (1983). Accreted terranes in the northern part of the

Philippine Archipelago. Tectonics 2, 211---236.Karig, D. E., Ingle, J., et al. (eds) (1975). Initial Reports of the Deep Sea

Drilling Project, 31. Washington, DC: US Government Printing

Office.

Kay, R. W. (1980). Volcanic arc magmas: implications of a

melting---mixing model for element recycling in the crust---upper

mantle system. Journal of Geology 88, 497---522.

Kay, R. W. & Kay, S. M. (1993). Delamination and delamination

magmatism. Tectonophysics 219, 177---189.

Knittel, U. & Defant, M. J. (1988). Sr isotopic and trace element

variations in Oligocene to Recent igneous rocks from the Philippine

island arc: evidence for recent enrichment in the sub-Philippine

mantle. Earth and Planetary Science Letters 87, 87---99.

Knittel, U. & Yang, T. F. (1998). Source components and enrichment

processes in the mantle wedge beneath Luzon (Philippines). In:

Flower, M., Chung, S.-L., Lo, C.-H. & Lee, T.-Y. (eds) Mantle

Dynamics and Plate Interactions in East Asia. American Geophysical Union,

Geodynamics Series 27, 385---403.

Knittel, U., Defant, M. J. & Raczek, I. (1988). Recent enrichment in

the source region of arc magmas from Luzon Island, Philippines: Sr

and Nd isotopic evidence. Geology 16, 73---76.

Kogiso, T., Tatsumi, Y. & Nakano, S. (1997a). Trace element transport

during dehydration processes in the subducted oceanic crust:

1. Experiments and implications for the origin of oceanic island

basalts. Earth and Planetary Science Letters 148, 193---205.

Kogiso, T., Tatsumi, Y., Shimoda, G. & Barsczus, H. G. (1997b). High

m (HIMU) ocean island basalts in southern Polynesia: new evidence

for whole mantle scale recycling of subducted oceanic crust. Journal of

Geophysical Research 102, 8085---8103.

Lin, P.-N. (1991). Trace element and isotopic characteristics of

western Pacific pelagic sediments: implications for the petro-

genesis of Mariana Arc lavas. Geochimica et Cosmochimica Acta 56,

1641---1654.

Lugmair, G. W. & Galer, S. J. G. (1992). Age and isotopic relationships

among the angerites Lewis Cliff 86010 and Angra dos Reis.

Geochimica et Cosmochimica Acta 56, 1673---1694.

McDermott, F., Defant, M. J., Hawkesworth, C. J., Maury, R. C. &

Joron, J. L. (1993). Isotope and trace element evidence for three

component mixing in the genesis of the North Luzon arc lavas

(Philippines). Contributions to Mineralogy and Petrology 113, 9---23.

Miklius, A., Flower, M. F. J., Huijsmans, J. P. P., Mukasa, S. B. &

Castillo, P. R. (1991). Geochemistry of lavas from Taal Volcano,

southwestern Luzon, Philippines: evidence for multiple magma

supply systems and mantle source heterogeneity. Journal of Petrology

32, 593---627.Miller, D. M., Goldstein, S. L. & Langmuir, C. H. (1994). Cerium/lead

and lead isotopic ratios in arc magmas and the enrichment of lead in

the continents. Nature 368, 514---520.



Morris, J. & Tera, F. (1989). 10Be and 9Be in mineral separates and

whole rocks from volcanic arcs: implications for sediment subduc-

tion. Geochimica et Cosmochimica Acta 53, 3197---3206.

Morris, J. D. & Hart, S. R. (1983). Isotopic and incompatible element

constraints on the genesis of island arc volcanics from Cold Bay and

Amac Island, Aleutians and implications for mantle structure.


Morris, J. D., Leeman, W. P. & Tera, F. (1990). The subducted

component in island arc lavas: constraints from Be isotopes and

B---Be systematics. Nature 344, 31---36.

Mukasa, S. B., McCabe, R. & Gill, J. B. (1987). Pb-isotopic

compositions of volcanic rocks in the west and east Philippine island

arcs: presence of the Dupal isotopic anomaly. Earth and Planetary

Science Letters 84, 153---164.

Mukasa, S. B., Flower, F. J. & Miklius, A. (1994). The Nd-, Sr- and Pb-

isotopic character of lavas from Taal, Laguna de Bay and Arayat

volcanoes, S.W. Luzon, Philippines: implications for arc magma

petrogenesis. Tectonophysics 235, 205---221.

Newhall, C. G. (1977). Geology and petrology of Mayon Volcano,

southeastern Luzon, Philippines. M.S. thesis, University of

California (Davis), 292 pp.

Newhall, C. G. (1979). Temporal variation in the lavas of Mayon

Volcano, Philippines. Journal of Volcanological and Geophysical Research

6, 61---83.

Norrish, K. & Chappell, B. (1977). X-ray fluorescence spectrometry.

In: Zussman, J. (ed.) Physical Methods in Determinative Mineralogy.

New York: Academic Press, pp. 201---272.Norrish, K. & Hutton, J. T. (1969). An accurate X-ray spectrographic

method for the analysis of a wide range of geological samples.


Pearce, J. A. & Peate, D. W. (1995). Tectonic implications of the

composition of volcanic arc magmas. Annual Review of Earth and

Planetary Sciences 23, 251---285.

Pearce, J. A., Kempton, P. D., Nowell, G. M. & Noble, S. R. (1999).

Hf---Nd element and isotope perspective on the nature and

provenance of mantle and subduction components in Western

Pacific arc---basin systems. Journal of Petrology 40, 1579---1611.

Peccerillo, A. & Taylor, S. R. (1976). Geochemistry of Eocene calc-

alkaline volcanic rocks from the Kastamonu area, northern Turkey.

Contributions to Mineralogy and Petrology 58, 63---81.

Plank, T. & Langmuir, C. H. (1998). The chemical composition of

subducting sediment and its consequences for the crust and mantle.

Chemical Geology 145, 325---394.

Poreda, R. & Craig, H. (1989). Helium isotope ratios in circum-Pacific

volcanic arcs. Nature 338, 473---478.

Rangin, C., Silver, E., von Breymann, M. T., et al. (eds) (1990).

Proceedings of the Ocean Drilling Program, Initial Reports, 124. College

Station, TX: Ocean Drilling Program.

Rapp, R. P., Shimizu, N., Norman, M. D. & Applegate, G. S. (1999).

Reaction between slab-derived melts and peridotite in the mantle

wedge: experimental constraints at 3�8GPa. Chemical Geology 160,

335---356.

Reagan, M. K., Gill, J. B., Malavassi, E. & Garcia, M. O. (1987).

Changes in magma composition at Arenal volcano, Costa Rica,

1968---1985: real-time monitoring of open system differentiation.

Bulletin of Volcanology 49, 415---434.

Sajona, F. G., Maury, R. C., Bellon, H., Cotten, J., Defant, M. J.,

Pubellier, M. & Rangin, C. (1993). Initiation of subduction and the

generation of slab melts in western and eastern Mindanao,

Philippines. Geology 21, 1007---1010.Sajona, F. G., Bellon, H., Maury, R. C., Pubellier, M., Cotten, J. &

Rangin, C. (1994). Magmatic response to abrupt changes in

geodynamic settings: Pliocene---Quaternary calc-alkaline lavas and

Nb enriched basalts of Leyte and Mindanao (Philippines).


Sajona, F. G., Bellon, H., Maury, R. C., Pubellier, M., Quebral, R. D.,

Cotten, J., Bayon, F. E., Pagado, E. & Pamatian, P. (1997). Tertiary

and Quaternary magmatism in Mindanao and Leyte (Philippines):

geochronology, geochemistry and tectonic setting. Journal of Asian

Earth Sciences 15, 121---153.

Sajona, F. G., Maury, R. C., Pubellier, M., Leterrier, J., Bellon, H. &

Cotten, J. (2000). Magmatic source enrichment by slab-derived melts

in a young post-collisional setting, central Mindanao (Philippines).

Lithos 54, 173---206.

Silver, E. & Rangin, C. (1991). Development of the Celebes Basin in

the context of western Pacific marginal basin history. In: Silver, E.,

Rangin, C., et al. (eds) Proceedings of the Ocean Drilling Program, Scientific

Results, 124. College Station, TX: Ocean Drilling Program,

pp. 39---49.Solidum, R. U. (2002). Geochemistry of volcanic arc lavas in

central and southern Philippines: contributions from the sub-

ducted slab. Ph.D. dissertation, University of California, San Diego,

197 pp.

Solidum, R. U. & Castillo, P. R. (2001). Geochemical characteristics of

sediments potentially subducted in western and eastern Philippines

(abstract). EOS Transactions, American Geophysical Union 82, F1305.

Spadea, P., D’Antonio, M. & Thirlwall, M. F. (1996). Source

characteristics of the basement rocks from the Sulu and Celebes

Basins (Western Pacific): chemical and isotopic evidence. Contributions

to Mineralogy and Petrology 123, 159---176.Sun, S.-S. & McDonough, W. F. (1989). Chemical and isotopic

systematics of oceanic basalts: implications for mantle composition

and processes. In: Saunders, A. D. & Norry, M. J. (eds) Magmatism in

the Ocean Basins. Geological Society, London, Special Publications 42,

313---345.

Taylor, B. & Hayes, D. E. (1983). Origin and history of the South

China Sea basin. In: Hayes, D. E. (ed.) The Tectonic and Geologic

Evolution of Southeast Asian Seas and Islands, Part 2. Geophysical Monograph,

American Geophysical Union 27, 23---56.

Todt, W., Cliff, R. A. Hanser, A. & Hofmann, A. W. (1996).

Evaluation of a 202Pb---205Pb double spike for high precision lead

isotope analyses. In: Basu, A. & Hart, S. (eds) Earth Processes: Reading

the Isotopic Code. Geophysical Monograph, American Geophysical Union 95,

429---437.

Tu, K., Flower, M. F. J., Carlson, R. W., Xie, G., Chen, C.-Y. &

Zhang, M. (1992). Magmatism in the South China Basin 1. Isotopic

and trace element evidence for an endogenous Dupal mantle

component. Chemical Geology 97, 47---62.

Walker, R. J., Carlson, R. W., Shirley, S. B. & Boyd, F. R. (1989). Os,

Sr, Nd and Pb isotope systematics of southern African peridotite

xenoliths: implications for the chemical evaluation of subcontinental

mantle. Geochimica et Cosmochimica Acta 53, 1583---1595.Wallace, P. & Carmichael, I. (1999). Quaternary volcanism near the

Valley of Mexico: implications for subduction zone magmatism and

the effects of crustal thickness variations on primitive magma

compositions. Contributions to Mineralogy and Petrology 135, 291---314.White, W. M. & Dupre, B. (1986). Sediment subduction and magma

genesis in the Lesser Antilles: isotopic and trace element constraints.

Journal of Geophysical Research 91, 5927---5941.

White, W. M., Hofmann, A. W. & Puchelt, H. (1987). Isotope

geochemistry of Pacific mid-ocean ridge basalt. Journal of Geophysical

Research 92, 4881---4893.

Woodhead, J., Eggins, S. & Gamble, J. (1993). High field strength and

transition element systematics in island and back-arc basin basalts:

evidence for multi-phase melt extraction and a depleted mantle

wedge. Earth and Planetary Science Letters 114, 491---504.



Xu, J.-F., Shinjo, R., Defant, M. J., Wang, Q. & Rapp, R. (2002).

Origin of Mesozoic adakitic intrusive rocks in the Ningzhen area of

East China; partial melting of delaminated lower continental crust?

Geology 30, 1111---1114.

Yang, T. F., Lee, T. Chen, C.-H., Cheng, S.-N., Knittel, U.,

Punongbayan, R. S. & Rasdas, A. R. (1996). A double island arc

between Taiwan and Luzon; consequence of ridge subduction.


You, C.-F., Castillo, P. R., Chan, L. H., Gieskes, J. M. & Spivack, A. J.

(1996). Trace element behavior in hydrothermal experiments:

implications for fluid processes at shallow depths in subduction

zones. Earth and Planetary Science Letters 140, 41---52.

Yumul, G. P., Jr, Dimalanta, C. B., Faustino, D. V. & de Jesus, J. V.

(1999). Silicic arc volcanism and lower crust melting: an

example from the central Luzon, Philippines. Journal of Geology 154,

13---14.



Documents

Geochemical Constraints on Possible Subduction Components in