The Nd-, Sr- and Pb-Isotopic Character of Lavas from Taal, Laguna De Bay

TECTONOPHYSICS

ELSEVIER Tectonophysics 235 (1994) 205-221

The Nd-, Sr- and Pb-isotopic character of lavas from Taal, Laguna de Bay and Arayat volcanoes, southwestern Luzon, Philippines: implications for arc magma petrogenesis

Samuel B. Mukasa a, Martin F.J. Flower b, Asta Miklius ’

a Department of Geological Sciences, University of Michigan, 1006 C.C. Little Building, Ann Arbor, MI 48109-1063, USA ’ Department of Geological Sciences, University of Illinois, P.O. Box 4348, Chicago, IL 60680, USA ’ U.S. Geological Survey, Hawaiian Volcano Observatory, P.O. Box 51, Hawaii National Park, HI 96718, USA (Accepted 18 November 1993)

Abstract

Following the amalgamation of a collage of pre-Neogene terranes largely by strike-slip and convergence mechanisms to form the Philippine islands, volcanic chains, related to oppositely dipping subduction zones, developed along the eastern and western margins of the archipelago. There is ample field evidence that this volcanic activity, predominantly talc-alkaline in chemical character, had commenced by the Oligocene. Volcanoes resulting from subduction along the Manila-Negros trench in the west (e.g., Taal, Laguna de Bay and Arayat) form a high-angle linear array, trending away from the MORB field on Pb-isotopic covariation diagrams; have the highest Sr- and lowest Nd-isotopic compositions, of the two chains (but nevertheless plotting above bulk earth on the 87Sr/86Sr versus 143Nd/‘“Nd covariation diagram); and exhibit Sm/Nd and Rb/Sr values that are lower and higher, respectively, than the estimated values for bulk earth. While the Sm/Nd and Rb/Sr characteristics are common to both chains, volcanoes associated with the Philippine-East Luzon trench have Pb-isotopic compositions that fall in the Indian Ocean MORB field and that require time-integrated evolution in a high Th/U environment. They also have higher Nd- and lower Sr-isotopic ratios. The source materials of Philippine volcanoes, therefore, have undergone varied recent enrichments in LILE, as indicated by the decoupling of isotopic and elemental ratios. These enrichments, particularly for the western volcanoes, cannot be entirely due to small degrees of partial melting in the mantle wedge, considering that they were accompanied by elevations in radiogenic Pb. Elevated Pb ratios are best explained by the introduction of subducted, continentally derived sediments. The sedimentary component in the western volcanoes is probably the South China Sea sediments derived largely from Eurasia. That this component is not available in the Philippine-East Luzon trench is reflected by the fact that the eastern volcanoes have higher Nd- and lower Sr-isotopic ratios as well as less radiogenic common Pb.

1. Introduction nent Neogene feature located in southwestern Luzon, Philippines (Fig. 1). Of the volcanic com- The Taal volcanic center, situated within a plexes lining the western margin of Luzon, Taal caldera/volcano-tectonic depression, is a promi- and a few others in its vicinity (e.g., Banahaw,

Laguna de Bay and San Cristobal) have the most its neighbors occupy an area that has been af- controversial and perhaps least understood tec- fected by deep-seated and regionally important tonic settings. According to Cardwell et al. (1980) cross faults. Implicit in this observation is that the and Hamburger et al. (19831, this group of volca- composition and processes controlling the gener- noes is the product of eastward-dipping subduc- ation of magmas in the Taal region probably tion along the Manila trench, like other volcanoes evolved with the changing stress regimes. Indeed, along western Luzon. Divis (19801, Cruz and Um- recent isotopic studies of volcanic materials from bal (1983) and Wolfe and Self (19831, however, the area have shown that there are fundamental have shown that, although subduction might have differences in source character between Neogene been important in their development, Taal and southwestern Luzon volcanoes and contempora-

120 E 125 E

15N

ION

Fig. 1. Generalized tectonic map of the northern Philippine islands showing the locations of the Taal, Laguna de Bay and Arayat volcanic complexes, as well as the oppositely dipping subduction zones; the North Palawan Continental Terrane (NCPT); the various volcanic lineaments (EH. = East Bataan Lineament; RX. = West Bataan Lineament; EML = East Mindoro Lineament; and B%&. = West Mindoro Lineament), the Philippine transform fault, and the Macoiod Corridor. The insert shows the entire archipelago, and a box in the middle of it outlines the area of the main map. S.B. Mukasa et al. / Tectonophysics 235 (1994) 205-221 207 neous volcanoes in the same chain to the north comparative assessment of rocks from the two (Mukasa et al., 1987; Defant et al., 1988; Knittel areas. Samples from two additional volcanoes in et al., 1988). southwestern Luzon are evaluated for reference: Emergence of the cross faults and the spatially the Laguna de Bay volcanic complex, which has related lavas of the Taal group-with higher K,O the same general tectonic setting as Taal Volcano and incompatible element concentrations as well but is possibly older, and Arayat Volcano, a K-rich as more radiogenic isotopic ratios-appears to member of the western Luzon chain of volcanoes, have coincided with collision in the late Miocene but distant from any possible cross-faulting ef- between the Philippine archipelago and the North fects. Palawan Continental Terrane (NPCT). Mukasa et al. (1987), Defant et al. (1988) and Knittel et al. (1988) have therefore concluded, on the basis of 2. Regional tectonic setting isotopic and major and trace element compositions as fingerprints, that introduction of crustal The pre-Neogene geology of the Philippine materials from the NPCT to the subduction zone archipelago is dominated by a collage of arc, (either slivers of continental crust or terrigenous continental and ophiolitic blocks that, on the sediments on the leading edge of the terrane), basis of structural, paleontological, stratigraphic contributed to the source character of post-colli- and paleomagnetic data, are believed to have sion lavas erupted between southwestern Luzon formed separately under different stress regimes and Panay (Fig. 1). far from their present geographic locations (Hsii, In the light of new isotopic and trace element 1972; Hamilton, 1979; De Boer et al., 1980; Balce results on post-spreading lavas scattered through- et al., 1981; Hashimoto, 1981; Fuller et al., 1983; out the South China Sea Basin (SCSB) (Tu et al., Karig, 1983; McCabe et al., 1985; Karig et al., 19911, subcontinental lithosphere, associated with 1986). Following amalgamation of these terranes the NPCT and introduced into the segment by strike-slip and convergence mechanisms, vol- boundary between the Manila and Negros sub- canic chains, related to oppositely dipping subduction systems by collision, has to be considered duction zones, developed along the eastern and as an alternative source. If post-spreading lavas western margins of the archipelago. Volcanic ac- of the SCSB reflect a lithospheric source, then tivity, associated with the eastward subduction of this source is heterogeneous, enriched in low South China Sea crust along the Manila-Negros field strength elements (LFSE) and light rare trench and the westward subduction of the earth elements (LREE) relative to N-MORB, de- Philippine Sea plate along the Philippine-East pleted in “Sr/‘%r and 143Nd/‘44Nd relative to Luzon trench (Bicol sector), has constructed two bulk earth, and similar in its Pb-isotopic signa- subparallel belts of predominantly talc-alkaline tures to global magma reservoirs, with the Dupal stratovolcanoes since the Oligocene (Hsii, 1972; isotopic anomaly as defined by Hart (1984; 1988). Hamilton, 1979; De Boer et al., 1980; Balce et al., Defined in terms of Pb-isotopic compositions, the 1981; Hashimoto, 1981; Fuller et al., 1983; Karig, Dupal anomaly is the deviation of any data set 1983; McCabe et al., 1985; Karig et al., 1986). from the Northern Hemisphere Reference Lines Taal, Laguna de Bay and Arayat, which are com- (NHRL), which pass through the data fields for pared here, occur in the central sector of the basalts from the East Pacific Rise, Mid Atlantic western volcanic belt in the southwest of Luzon Ridge and several ocean islands. The anomaly is island, in association with eastward subduction expressed as the vertical deviation in 207Pb/204Pb along the Manila-Negros trench (Fig. 1). or 208Pb/204Pb from the line, with the minimum Neogene volcanic chains along the western deviation limit having been set at a A208/204 of Philippines have been referred to as the West +60 by Hart (1988). Many (but not all) of the Philippine arc (Mukasa et al., 1987), the Bataan SCSB geochemical features are shared by the and Mindoro arcs (De Boer et al., 1980; Defant lavas from Taal Volcano, lending impetus for a et al., 1988) and the Luzon arc (Defant et al., 208 S.B. Mukasa et al. / Tectonophysics235 (1994) 205-221

1989); all of these names remain in use at pre- EBL. Defant et al. (1988) have pointed out, how- sent. These volcanoes form two closely spaced, ever, that Laguna de Bay and Taal’s oldest erup- subparallel chains that Defant et al. (1988) have tive products might be older than the MC and termed the eastern and western Bataan linea- should, therefore, not display cross-fault effects, ments (EBL and WBL), in northern and central if these are distinct and are reflected by the lava Luzon, and eastern and western Mindoro linea- chemistry. Depth to the top of the subducted slab ments (EML and WML) in southwestern Luzon, beneath EBL and MC volcanoes is another im- Mindoro and Marinduque. The Bataan and Min- portant distinction, possibly bearing on composi- doro lineaments are separated by a zone of tion of the magmas. Seismic studies by Cardwell northeast-trending cross-faults which define a et al. (198C0, Hamburger et al. (19831, Hayes and rift-like tectonic feature, referred to by Defant et Lewis (1984) and Rangin et al. (1988) have re- al. (1988) and Fiirster et al. (1990) as the Macolod vealed that the top of the subducted slab is 180- Corridor (MC). Taal and Laguna de Bay are 200 km beneath Arayat, but much deeper be- located in this feature, while Arayat is part of the neath the MC volcanoes.

T _ 4 Tagayby kilometers

Fig. 2. Map of Taal Volcano, showing the locations of the caldera rim samples at Tagaytay, Napa~n Island and Mt. Macolod. The insert is an enlargement of Volcano Island and shows the lava flows sampled for this study. S.B. Mukasa et al. / Tectonophysics 235 (1994) 205-221 209

Immediately west of the Manila-Negros trench tions in Fig. 2). The differences between calc-al- system is the South China Sea Basin (SCSB), a kaline and tholeiitic suites on Volcano Island are marginal basin which began to develop in the best explained by their evolution in separate sys- Paleocene, along what was originally an Andean- tems without material exchange between the two type convergence zone, and evolved through rock series. Chemical and petrographic evidence crustal stretching and sea-floor spreading (Taylor suggest that fractional crystallization and mixing and Hayes, 1983; Ru and Piggott, 1986; Tappon- between less- and more-evolved magmas were nier et al., 1986). Remnant crustal blocks sepa- important processes in development of both se- rated from Eurasia by the marginal basin opening ries, with mixing predominating in the calc-al- are presently scattered throughout the basin, but kaline suite. are concentrated along the eastern margin where Compared to the WBL volcanoes, Taal lavas they collided with Borneo and the Philippine (and lavas of the other EBL and MC volcanoes) archipelago, effectively terminating spreading. are more alkalic and have higher contents of The NPCT, which abuts islands in the western incompatible elements (Miklius et al., 1991). De- central Philippines (Fig. l), and appears to have fant et al. (1988) attribute these features to very contaminated post-collision magma sources in the small degrees of partial melting as compared to area, is one of these crustal blocks. WBL lavas, which have lower contents of these After termination of spreading in the SCSB, elements. Contradiction between certain trace el- widespread volcanism penetrated the basin floor, ement ratio covariations (e.g., Ba/Y versus Zr/Y) stranded microcontinents and the mainland, from and incompatible element contents, with respect Thailand in the south to Korea in the north. Tu to degrees of partial melting, have led Miklius et et al. (1991) have argued that these post-spread- al. (1991) to conclude that source heterogeneity ing lavas probably sampled the subcontinental was the principal factor determining parent melt lithosphere-dispersed with basin opening-that differences. had been enriched along the, now defunct, An- In their comprehensive study of post-spreading dean-type margin of Southeast Asia. It is worth lavas in the SCSB region, Tu et al. (1991) ob- considering and investigating the possibility that served LFSE and REE enrichments relative to such a lithospheric source was introduced into N-MORB, as well as 87Sr/86Sr and 143Nd/‘44Nd the MC region with collision between the NPCT ratios that are variable but are depleted relative and the Philippine archipelago. to bulk earth values. In addition, they discovered not only heterogeneities in isotopic and trace element ratios, but a decoupling of the two pa- 3. Previous geochemical studies of Taal Volcano rameters. The observed decoupling is best ex- and South China Sea Basin hasalts plained in terms of recent but protracted enrichment histories, of the type believed to occur in Geochemical studies by Miklius et al. (1991) of supra-subduction zone mantle wedge environ- Taal lavas have revealed that talc-alkaline and ments. Pb-isotopic compositions for most of the tholeiitic magmas evolved coevally. On Volcano lavas fall in the fields for the Dupal isotopic Island, the medial post-caldera edifice of cones anomaly (A208/204 = 45-73), implying that ei- and tuff rings of the complex (Fig. 2), the two ther the anomaly, documented chiefly in the lava suites are easily distinguished from one an- southern hemisphere (Hart, 1984; 1988), extends other, and from pre-caldera eruptive products into this region, or that the SCSB source is a rimming Taal Lake, on the basis of major ele- relatively isolated domain produced by local en- ments and incompatible trace element ratios. For richment processes. Dupal-type Pb-isotopic com- the equivalent MgO content, post-caldera lavas positions have also been documented on samples on Volcano Island have higher incompatible ele- from the Philippine Sea by Hickey-Vargas (1991), ment contents than pre- and syn-caldera collapse and from further north, in the Sea of Japan, by lavas at Mt. Macolod and Napayun Island (loca- Tatsumoto and Nakamura (1991). 210 S.3. Mukasa et al. / Tectonophysics 235 (1994) 205-221

4. Sampling of Mt. Tabaro on Volcano Island, a post-caldera edifice, constitute the youngest group of samples. The Taal samples used in this study, for which No relative time constraints have been attached locations are shown in Fig. 2, are a sub-set of the to the Laguna de Bay and Arayat samples, all of much larger suite studied and described in detail them basaltic, since they are so few in number. by Miklius et al. (1991); descriptions here are therefore brief. A representative compositional spectrum and broad time sequences were the 5. Analytical procedures principal criteria used to choose a sample sub-set. Thus basalt, basaltic andesite, andesite and dacite, Whole-rock powders were leached with warm ranging in relative age from pre-caldera to his- 2 N HCl and 3 N HNO, sequentially, and then toric post-caldera collapse, have been analyzed. washed with doubly distilled H,O. Samples of The presence of multiple vents has produced approximately 100 mg for Pb and 50 mg for complex stratigraphic relations, requiring tight Rb-Sr and Sm-Nd were dissolved in a 10: 1 isotopic age constraints, as yet unavailable. In the solution of HF: HNO,, with 14 N HNO, added present study time resolution is limited to pre-, first to minimize the formation of fluorides. The syn- and post-caldera collapse, which is neverthe- dissolutions were carried out in screw-top PFA less instructive. The oldest samples come from Teflon beakers on a hot plate set at 150°C. Mt. Macolod, a deeply eroded satellite plug, and Pb was isolated using AG 1X8 anion resin on from pre-caldera bombs at the northwestern rim 0.5 ml and 0.2 ml HBr-HCl ion exchange of the volcanic center. Lavas of the 1969 eruption columns. A procedural blank run with the sam-

Table 1 Elemental concentrations and isotopic compositions for lavas from Taal, Laguna de Bay and Arayat volcanoes, Luzon Island, Philippines Sample Concentrations (ppm) a Isotopic Ratios ’ Rb Sr Sm Nd Th/ “Rb/ 87Sr/86Sr 14’Sm/ 143Nd/ ‘=Pb/’ “‘Pb/ ‘03Pb/ Ta s%r 144Nd ‘“Nd 204Pb ‘“Pb ‘04Pb

Tad Volcano B4 26 385.6 2.83 11.77 23.1 0.195 0.70457 f 2 0.145 0.51281 + 2 18.581 15.614 38.729 BALl 86 211.9 4.01 18.76 29.9 1.175 0.70459 * 4 0.129 0.51281 + 2 18.577 15.584 38.626 BM8 81 237.5 4.30 19.91 28.6 0.987 0.70458 + 4 0.130 0.51283 + 2 18.635 15.656 38.846 CAL1 28 355.6 2.99 12.41 20.5 0.228 0.70452 f 4 0.145 18.554 15.586 38.647 CY3 32 328.2 3.20 13.60 24.5 0.282 0.70458 * 2 0.142 18.585 15.599 38.687 MB5 26 345.5 2.97 12.11 18.5 0.218 0.70454 + 2 0.147 0.51282 &-2 18.639 15.668 38.867 MC1 23 344.5 3.19 12.06 14.8 0.193 0.70443 * 4 0.159 0.51284 f 4 18.570 15.610 38.675 NPl 47 358.5 4.03 17.58 24.5 0.380 0.70454 + 2 0.138 0.51279 + 2 18.593 15.620 38.743 T5.5 29 328.4 3.20 12.96 - - 0.70449 rt: 2 0.149 0.51285 rt 2 18.567 15.591 38.652 TT4 SO 378.6 4.02 17.14 26.7 0.382 0.70472 f 2 0.141 0.51281 f 2 18.623 15.643 38.803

Laguna De Bay Volcano LBMI-1 - - 6.44 29.32 - - 0.70426 f 4 0.132 0.51276 & 2 - LTN-2 - _ 6.59 29.88 - - 0.70409 * 2 0.133 0.51285 f 2 18.621 15.594 38.605 TLHI-9 - 413.0 3.90 15.96 - - 0.70426 + 2 0.147 0.51285 + 2 18.473 15.552 38.456

Arayaf Volcano EG80-89 - 446.9 4.01 16.37 - - 0.70410 + 4 0.148 0.51287 f 2 18.473 15.584 38.604 EG80-90 - 480.3 5.59 27.42 - - 0.70432 f 2 0.123 0.51281 + 2 18.505 15.605 38.695 a Rb concentrations determined by XRF; Th and Ta by INCA; and Sr, Sm and Nd by isotope dilution. b Errors on the *‘Sr/%r and 143Nd/~~Nd values represent 20~ based on wjthin-Nn statistics. Errors for the Pb ratios are smalier than 0.1%. Normalization values are given in the analytical techniques section. S.B. Mukasa et al. / Tectonophysics 235 (1994) 205-221 211 ples had 0.7 ng of Pb, a small amount (< 1%) of on rhenium triple filaments with weak phosphoric the total Pb in each sample. The Pb samples were acid. Pb, Nd, Sm and Rb were measured in loaded on rhenium single filaments using the single-collector mode and Sr multi-dynamically phosphoric acid-silica gel method. using a VG Isomass 354 triple collector, thermal Sample aliquots dissolved for Rb-Sr and Sm- ionization mass spectrometer. Nd isotopic measurements were dried down and Within-run precisions for the measured Pb ra- redissolved three times in 6 N HCl, and finally tios were better than 0.05% at the 95% confi- equilibrated in 2.5 N HCl to be loaded onto dence level. These ratios have been corrected for columns. Rb and Sr were separated from other fractionation during ionization using replicate elements and from each other using 2.5 N HCl analyses of the standard SRM-981 using a factor on cation exchange columns; REE concentrates of 0.09 k 0.02% per atomic mass unit. Sr and Nd were removed from the same columns using 6 N isotopic fractionations were corrected by normal- HCl. The REE fractions were dried and then izing to 86Sr/88Sr and 146Nd/‘44Nd values of picked up in 0.25 N HCl for introduction onto 0.11940 and 0.72190, respectively. The 87Sr/86Sr columns tightly packed with Bis (2-ethylhexyl) value for NBS standard SRM-987, based on repli- hydrogen phosphate-coated Teflon powder. Tight cate analyses, is 0.71024 &-1 and the 143Nd/144Nd packing prevented solution channelling that might value for the La Jolla standard is 0.51185 k 1. cause incomplete separation of the REE; a substantial hydrostatic head helped to speed up the flow rate. Nd was recovered with 0.25 N HCl, 6. Results and Sm with 0.6 N HCl. Procedural blanks for Nd and Sr were 0.04 ng and 0.1 ng, respectively. Isotopic compositions, and elemental concen- Rb samples were loaded on tantalum single trations and ratios, are presented in Table 1 and filaments with 1 N HCl, Sr on rhenium single plotted on various correlation diagrams in Figs. filaments with a TaCl, solution, and Sm and Nd 3-8. Detailed eruption chronologies for Taal, La-

15.6

18.0 19.0 20sPb/204Pb Fig. 3. 207Pb/204Pb versus 206Pb/‘04Pb covariation diagram with data for Taal, Laguna de Bay and Arayat volcanoes as well as data from Tu et al. (1991, and unpublished) for lavas from Mindoro ophiolites and post-spreading eruptions in the South China Sea Basin. Fields for the Bicol (eastern Philippines), Halmahera (Indonesia) and Taiwanese arcs, and Atlantic, Pacific and Indian Ocean MORB (see references in the text) are included for comparison. Note that the three Philippine volcanoes studied yield a linear array at a high angle to the MORB trend and are more radiogenic than the volcanoes in the Bicol arc of the eastern Philippines. 212 S.B. Mukasa et al. / Tectonophysics 235 (1994) 205-221 guna de Bay and Arayat are still not understood array relative to the MORB trend, lying totally well enough for complete assessments of the outside the generalized MORB field which in- time-integrated isotopic record of the magma cludes fields for Indian Ocean MORB and OIB. supply systems. It should be noted, however, that, Line-fitting through the points yields a correla- at Taal, isotopic ratios do not vary as a function tion coefficient of 0.939, but no meaningful age of rock composition, distance from the trench, information. The two Laguna de Bay samples proximity to the cross-faults, or relative age in the analyzed for Pb-isotopic compositions fall at the pre-, syn- and post-caldera collapse groupings. low 207Pb/204Pb end of the Taal array, and one of the two samples lies in the MORB field. Arayat 6.1. Fb-isotopic uariations data, in contrast, lie slightly to the left of the Taal-Laguna de Bay array, but are close to the The Pb-isotopic data for Taal, Laguna de Bay MORB field. On the 208Pb/204Pb versus “‘Pb/ and Arayat volcanoes listed in Table 1 are plotted 204Pb diagram (Fig. 41, the Taal-Laguna de Bay on 207Pb/204Pb versus 206Pb/204Pb and “‘Pb/ array is reproduced, albeit without as good a ‘04Pb versus *06Pb/204Pb covariation diagrams in correlation coefficient as on the diagram in Fig. Figs. 3 and 4, respectively. Also shown for com- 3. As in Fig. 3, data for Mt. Arayat fall to the low parison are the fields for MORB (Sun et al., 2osPb/204Pb side of the array. However, data for 1975; Sun, 1980; Vidal and Clauer, 1983; Cohen all three volcanoes fall within the field for Indian and O’Nions, 1982; DuprC and Allegre, 1983; Ocean MORB (Hamelin and Allegre, 1985; Hamelin et al., 1984, 1986; Hart, 1984; Hamelin Hamelin et al., 1986; Mahoney et al., 1989), con- and Allegre, 19851, the Halmahera arc (Morris et siderably above Atlantic and Pacific Ocean al., 1983), the Taiwan arc (Sun, 1980) and data MORB, and overlap with the fields for the points for post-spreading lavas in the South China Halmahera and Taiwan arcs. Sea Basin and ophiolites on Mindoro island iTu On a A2”Pb/ 204Pb versus A*“Pb/ 204Pb dia- et al., 1991). The most striking feature of the Taal gram (Fig. 51, which complements the other Pb data in Fig. 3 is the remarkably high-angle, linear covariation plots, the data fall along a broadly

o Laguna de Bay

H S. China Basin

MORB (Atlamic and Pacific)

- w 37.0 f I /I I I I I I 17.0 18.0 19.0 20.0 21.0 206Pb/204Pb Fig. 4. 20sPb/204Pb versus 206Pb/Z04Pb covariationdiagram for the three Philippine volcanoes studied with various fields for comparison, as described in the caption for Fig. 3. Note that in this covariation space as well, the western Philippine volcanoes are more radiogenic than the volcanoes of the Bicot arc in the eastern Philippines. S.B. Mukasa et al. / Tectonophysics 235 (19941205-221 213

14 l Taal Pacific o Laguna de Bay Sediments 12 A Arayat n S. China Basin 10 q Mindoro ophiolites I 8

-80 -60 -40 -20 0 20 40 60 80 100 A 208Pb/204Pb Fig. 5. A7/4 versus A8/4 diagram comparing the *‘*Pb and ‘07Pb enrichments of Taal, Laguna de Bay and Arayat volcanoes with enrichments for the Bicol arc in the eastern Philippines and the post-spreading lavas of the South China Sea Basin. The end-members enriched mantle of type I (EM II and type II (EM II), depleted MORB mantle (DMM), high m (HIMU), the local Mindoro ophiohtes and Pacific sediments are also included on the diagram. For the three western Philippine volcanoes, ‘OsPb and ‘07Pb enrichments are positively correlated. In contrast, the Bicol arc volcanoes and some of the post-spreading lavas in the South China Sea display preferential enrichment in aasPb compared to =‘.Pb. Source materials for the latter appear to have had a time-integrated record of high Th/U. While it is possible that the source of magmas for western Philippine volcanoes also had this elemental signature, the Pb budget is dominated by components enriched in both 208Pb and ‘07Pb and believed to be largely subducted sediments.

linear trend in the top right-hand quadrant, indi- For comparison, the diagram includes fields for cating 208Pb and 207Pb enrichments in the lavas MORB, OIB and several island arcs, using data relative to northern hemisphere basaltic rocks from various sources. Data for post-spreading with similar 206Pb/204Pb values. In contrast to lavas of the SCSB and ophiolitic slivers in Min- other volcanic centers, in both the Bataan- doro and Palawan from Tu et al. (1991) are also Mindoro and Bicol arcs, studied by Mukasa et al. included. Taal data are remarkably homoge- 11987), Taal, Laguna de Bay and Arayat have neous, defining a tight cluster toward the lower A207Pb/ 204Pb and A208Pb/ 204Pb values that end of the field for the Izu-Honshu arc in Japan. place them above the best fit regression line for All three Laguna de Bay samples have initial southern hemisphere basaltic rocks as calculated 87Sr/86Sr values lower than those of the Taal by Hart (1984). 207Pb enrichment for a given samples. They aIso show a wider range in initial 208Pb is more pronounced for these samples than ‘43Nd/*44Nd values, despite the very small sam- in those studied earlier. In this respect, the three ple population. Similarly, the two Arayat samples volcanoes are similar to some of the post-spread- have initial Sr ratios lower than those of the Taal ing lavas in the SCSB; the latter, however, in- samples and are notably different from each other clude samples which fall below the regression in their initial Nd ratios. Quite clearly, these line. three western Philippine volcanoes have Nd-Sr isotopic signatures that are discernible from those 6.2. ‘43Nd /144Nd- 87Sr/?!?r couariation of the large majority of post-spreading SCSB Variations in Nd and Sr initial isotopic ratios lavas, the source materials of ophiolitic slivers in for the three volcanoes are illustrated in Fig. 6. Mindoro and Palawan and the Bicol arc (Mukasa 214 S.B. Mukasa et al. / Tectonophysics235 (1994) 205-221 et al., 1986; Mukasa et al., in review). Nd-Sr 0.716 isotopic data for the Bicol arc fall in the field for the Mariana, New Britain, Aleutian and South Sandwich arcs and are remarkably homogeneous Modem Ocean over a wide area. 0.712 6.3. 87Sr /86Sr versus =06Pb /204Pb c~u~~~t~o~

*?Sr/*%r versus 206Pb/204Pb values have been plotted together in Fig. 7, which includes the 3j fields for MORB, modern ocean sediments, and Sf30.708 2 6 0.706

E Sr 0 30 60 0.5132 IO 0.5131 South Sandwiches 8 0.702 IArc, East Philippines 0.5130 6

0.5129 4 i 8.0 19.0 2U.O 0.5128 20sPb/204Pb 2 s 0.5127 Fig. 7. s7Sr/8aSr versus 206Pb/204Pb covariation diagram for 0 $ Taal, Laguna de Bay and Arayat volcanoes compared to the -% 0.5126 Bicol arc of the eastern Philippines, post-spreading lavas of -2 'Nd the South China Sea Basin (SCSB), MORB, Lesser Antilles $ 0.5125 -4 arc and modern ocean sediments. Initial srSr/s%r values for the three western Philippine volcanoes are higher than the 0.5124 -6 values for SCSB lavas of comparable 206Pb/2”Pb ratios. 0.5123 However, when compared to their generally high 207Pb/204Pb -8 and 2”sPb/204Pb ratios, the western Philippine volcanoes have 0.5122 relatively non-radiogenic Sr. -10 0.5121 -12 0.5120 -14 the Lesser AntiIles arc, as well as data points for 0.5119 0.703 0.705 0.707 0.709 0.711 the Mindoro-Palawan ophiolites and post- 87W 93 spreading lavas of the SCSB. The diagram illus- Fig. 6. ‘43Nd/1~Nd versus 87Sr/86Sr covariat~on diagram trates the point that Taal, Laguna de Bay and showing the data for the Philippine volcanoes studied and for Arayat *‘Sr/‘%r values for all but two samples the post-spreading lavas of the South China Sea Basin. These are higher than those of post-spreading lavas in are compared to the fields for ocean island basaits (OIB), the SCSB, despite the samples having compara- mid-ocean ridge basaits (MORB) and various arcs (references given in the text). Taal, Laguna de Bay and Arayat volcanoes ble 206Pb/204Pb values. This feature points to a all have more radiogenic Sr and less radiogenic Nd than the slight but significant difference in the character Bicol arc volcanoes in the eastern Philippines. Their *43Nd/ of magma sources beneath the two regions. Com- r”Nd values, however, are less radiogenic than bulk earth pared to data for the Lesser Antilles arc (David- and yet their r4’Sm/‘“Nd ratios are lower. This characteris- son, 19861, the three Philippine volcanoes have tic indicates that the source materials were enriched in light rare earth elements, possibly derived from the subducted remarkably high Sr initial ratios for their low components, shortly before melt extraction. 206Pb/204Pb.

216 S.B. Mukasa et al. / Tectonophysics235 (1994) 205-221

high Th/La and low HFSE/LREE in arcs are subsequent subduction beneath new arcs (Varne, quite prominent (cf. Navon and Stolper, 1987; 1985; Varne and Foden, 1986; Tu et al., 1991). Kelemen et al., 19901, as are steep trends in Recognizing eruptive products with enriched sub- Pb-Pb space (cf. White and Dupre, 1986; Mukasa continental lithospheric materials of this nature is et al., 1987). extremely challenging (e.g., Hawkesworth and El- Wedge and arc crust mixtures have also been lam, 1989). At any rate, geochemical comparison considered viable magma sources in a number of of lavas erupted in a marginal basin with those in studies (e.g., Tilton and Barreiro, 1980; Arculus arc stratovolcanoes beneath which the marginal and Johnson, 1981; Leeman, 1983; Davidson, basin floor subducts is desirable. Considering the 1986). This model is particularly appropriate for rapidity with which subducted sediments are re- continental arc lavas that have isotopic signatures cycled to the mantle wedge, it should be possible, resembling those of local ancient sialic crust, yet using “Be and B systematics (cf. Morris et al., have mafic varieties not easily explained as pure 1990), to distinguish enriched subcontinental crustal derivatives. It also seems appropriate for lithosphere from recently subducted sedimentary suites of rocks with isotopic patterns (e.g., in components. al80 versus 87Sr/86Sr space) more suggestive of upper level crustal contamination than source 7.2. Isotopic character of the source materials contamination. Other studies have recognized that H,O-rich A striking feature of Taal Volcano isotopic and/or low-melting subducted slab components compositions is the relative homogeneity of “Sr/ are released into the supra-subduction mantle ‘%r and 143Nd/144Nd (Fig. 6) with high and wedge, forming pervasively metasomatized rock variable 207Pb/2”Pb and 208Pb/2”4Pb values mixtures with high potential to produce island (Figs. 3 and 4, respectively). On the 143Nd/144Nd arc-type magmas on melting (Kay, 1980; Sun, versus 87Sr/%r diagram (Fig. 61, the data form a 1980; Gill, 1981; McCulloch and Perfit, 1981; very small cluster of points near the depleted end Mukasa et al., 1987). This model is consistent in of the much larger Lesser Antilles field and within many arcs with the observed alkalis/LREE, alka- the field for the Izu-Honshu arc of Japan. With line earths/LREE and Th/La values. However, points falling well above bulk earth 143Nd/‘44Nd devolatilization mechanisms remain poorly con- values, and having “Sr/%r values either similar strained and the composition of subducted sedi- to or less than bulk earth, the source materials ments, which include ferruginous clays and appear to have evolved in a long-term depleted pelagic, hemipelagic, metalliferous and volcani- environment. The samples are, however, less de- elastic sediments, are poorly documented. That pleted compared to lavas from the Mariana, New not withstanding, ‘*Be studies of arc lavas (Tera Britain, Aleutian or South Sandwich arcs. In con- et al., 1986; Morris and Tera, 1989; Morris et al., trast to the homogeneity suggested by Nd and Sr 1990) have provided strong evidence to support ratios, the high-angle Pb-isotopic arrays indicate the notion that a subducted sediment component a more complex heterogeneous source, involving contributes to the source materials of arc lavas. mixing between crustal and MORB-like compo- With a half-life of only 1.5 X lo6 Ma, “Be decays nents, irrespective of the bulk composition or age rapidly in the mantle wedge. Its detection is, of the rocks. therefore, a good indicator of the rapidity with A comparison of data for Arayat, Laguna de which sedimentary components recycle through Bay and Taal shows that isotopic differences be- subduction zones. tween them are small, suggesting that source ma- Finally, another model worth considering, par- terials for the three volcanoes were probably sim- ticularly for arcs abutting marginal basins, in- ilar, despite respective differences in the angle of volves the disaggregation by rifting of Andean- dip for the subducted slab. type continental margins, resulting in the disper- Recent enrichment in LILE of the source is sal of enriched subcontinental lithosphere and its indicated by the decoupling of isotopic and ele- S. B. Mukasa et al. / Tectonophysics 235 (1994) 205-221 217

mental ratios, whereby low Sm/Nd, and high concomitant decreases in compatible elements Rb/Sr values relative to bulk earth (Fig. 8) coin- such as Ni, Cr and Sr. MORB-normalized incom- cide with a time-integrated Nd-Sr isotopic record patible element distributions (spidergrams) re- of depletion. If the inferred enrichment event flect pronounced ‘troughs’ for Ta, Nb and Ti, were old (> 200 Ma), there would be more sub- characteristic of arc magmas and possibly their stantial retardation and elevation in the 143Nd/ sources (Ryerson and Watson, 1987; Kelemen et 144Nd and the 87Sr/86Sr values, respectively. It is al., 1990). Despite any isotopic similarities be- unlikely that LILE enrichment reflects small de- tween the post-spreading SCSB and Bataan- grees of partial melting as proposed by Defant et Mindoro eruptives, differences in their spider- al. (19881, inasmuch as this mechanism fails to grams appear to preclude a common source for explain the elevated 207Pb/204Pb values observed. these magmas. Modification of elemental ratios is best explained by the recent addition of an enriched component, 7.4. Comparison of isotopic and elemental varia- the most likely candidate being subducted marine tions between Taal Volcano, the Bicol arc and sediment. This view was rejected by Defant et al. post-spreading lavas in the SCSB (1988) on the basis of trace element data, which may not be uniquely interpretable. The relatively A comparison of isotopic and trace element slight enrichment in the Nd- and Sr-isotopic com- data for samples from Taal, the Bicol arc in the positions (cf. the Banda and Lesser Antilles arcs eastern Philippines and post-spreading lavas in in Fig. 61, may indicate less sedimentary contami- the SCSB provides some insights into the types of nation of the southwestern Luzon source, assum- components that might be available in the magma ing, reasonably, that sediment characteristics are sources for each group. Trace element data for not overly different in the two areas. Assessing the Bicol arc are only now being acquired by our when the sedimentary component might have group and so an interim comparison here must be been introduced into the source materials is vi- based on the isotopic data in Mukasa et al. (1986; tally important in distinguishing whether the sedi- 1987) and Mukasa et al. (in review). ment involved was subducted since the Eocene Of the three areas, the Bicol arc appears to inception of the Manila-Negros Trench or was have the simplest isotopic signatures, explicable recycled into the subcontinental lithosphere by magma derivation from source materials with (along the southeast Asia margin in what is now a a long, time-integrated record of LILE depletion defunct continental arc) prior to the opening of relative to bulk earth. In Figs. 3 and 4, samples the SCSB. Presumably, the SCSB source was for the Bicol arc fall in the MORB fields (ruled enriched prior to its opening via disaggregation of fields), and in Fig. 5, exhibit signs of having been the continental arc and, according to this model, derived from materials with a long history of high west Philippine volcanoes would have acquired Th/U values, although not as extreme as MORB the same signature as the result of subduction or and OIB from the Indian Ocean (DuprC and delamination during and after collision of the Allbgre, 1983; Hart, 1984; 1988). SCSB lithosphere beneath the western margin of While their 206Pb/204Pb values are not as low the Philippine archipelago. The question of tim- as those of the Bicol arc (Figs. 3,4 and 7), a large ing is considered below after comparison of ele- number of the post-spreading lava samples from mental and isotopic variations in Philippine and the SCSB fall in the MORB field, and exhibit post-spreading SCSB eruptives. Dupal anomaly characteristics as well (Tu et al., 1991). A few samples, however, have elevated 7.3. Observations based on trace elements 207Pb/204Pb values, placing them outside the MORB field in Fig. 3 and suggesting sediment Trace element data for Taal lavas in Miklius et addition to their source materials sometime prior al. (1991) reflect an increase in the incompatible to magma extraction. With such isotopic hetero- elements Ba, K, Rb, Th and Zr with SiO,, and geneities, the SCSB group has had a much more 218 S.B. Mukasa et al. / Tectonophysics 235 (1994) 205-221

complex history of LILE depletion and enrichment than has the source material for lavas in the Bicol arc. In having comparatively low Sri ratios, falling between 0.70355 and 0.70376 (Figs. 6 and 7), the Bicol arc and post-spreading lavas of the SCSB are similar. However, as already stated, their 206Pb/204Pb values are quite different. On the other hand, Taal Volcano (and for that matter Laguna de Bay and Arayat volcanoes) have consistently the lowest 143Nd/144Nd, highest 87Sr/86Sr, and most radiogenic Pb ratios in the three sample groups. For the Taal and SCSB Withinplate Lavas samples, this is despite the fact that Sm/Nd and Rb/Sr values are similar for all rocks of basaltic I I I I I composition. If similarity of these elemental ra- 0 10 20 30 40 50 t D tios was the result of western Philippine volca- ThlTa noes inheriting enriched Eurasian subcontinental Fig. 9. A comparison of the Sr/Nd and Th/Ta ratios of Taal lithosphere, delivered to the archipelago by Volcano, South China Sea Basin post-spreading lavas, within- plate lavas and average continental crust. The SCSB lavas spreading and subduction associated with open- overlap with the field for within-plate lavas, while the Taal ing of the SCSB, then Sr-, Nd- and Pb-isotopic samples fall on a mixing line between MORB and sediment compositions of lavas from the two areas would dominated end-members. The two Taal samples off the mix- have been the same. This is not what is observed. ing line are dacitic in composition and might owe their low Therefore, either the elemental ratios for the two Sr/Nd to plagioclase fractional crystallization. areas were equalized in different and unrelated enrichment events or, following a common enrichment event, west Philippine lava sources were 207Pb/204Pb values also to involvement of re- endowed with yet another enriched component, cently subducted SCSB sediments at the source. apparently not made available to the post-spread- Varied but small degrees of partial melting, ing lavas of the SCSB. long-held source heterogeneities and the intro- The data require Taal and its neighboring vol- duction of a sedimentary component into the canoes to have acquired an LILE enriched com- supra-subduction mantle wedge are the processes ponent recently. This component is not available that determine the trace element inventories of to the source regions of the post-spreading lavas derivative melts prior to any fractionation effects. in the SCSB or even to the Bicol arc. Recently Fig. 9, a plot of Sr/Nd versus Th/Ta for SCSB subducted SCSB sediments, derived largely from post-spreading lavas (solid squares) and Taal Vol- young Eurasian crust, comprise the most logical cano (circles), highlights our view that different component to introduce the observed LILE en- sources or magma evolution mechanisms have richment. prevailed in the two areas. For comparison, no- Fig. 5 illustrates well the fact that samples tice the location of average crust (Taylor and from Taal, Laguna de Bay and Arayat volcanoes McLennan, 1985), within-plate lavas and fresh do not have the classic Dupal-type isotopic MORB glass (Hofmann, 1988), as well as the anomaly. Elevated 208Pb/204Pb values are accom- mixing line between altered MORB and sedi- panied by elevated *“Pb/*04Pb values in all sam- ment-dominated end-members, first modelled by ples, indicating unexceptional Th/U values in the Rogers et al. (1985). If the two dacite samples source. Besides that, 87Sr/86Sr values for all sam- from Taal (the two points below the mixing trend) ples are less than the minimum limit of 0.705 for are excluded to minimize the effects of plagio- the Dupal isotopic anomaly. We attribute the clase fractional crystallization on the observed positive correlation between 208Pb/204Pb and Sr/Nd values, then Taal samples lie reasonably S.3. Mukasa et al. / Tectonophysics 235 (1994) 20.5221 219 close to the mixing line between ahered MORB Th/Ta and, for the most part, lower 143Nd/144Nd and sediment-dominated end-members. In con- values. In addition, all but one of the samples trast, SCSB post-spreading lavas have much lower from the three volcanoes analyzed fall above the Th/Ta values and are essentially indistinguish- MORB field, whereas the majority of the post- able from within-plate lavas and fresh MORB spreading lavas from the SCSB fall within it on glass. We view this difference as another good both Pb-covariation diagrams. indication that magma sources of the three west- (4) The sedimentary component that infiu- ern Philippine volcanoes studied have been per- enced the trace element and isotopic character of meated with an LILE-enriched component that lavas from the three western Philippine volcanoes appears to be absent in SCSB post-spreading is not available to Bicol arc volcanoes lining the lavas. Subducted, ~OntineRtally derived sediments eastern margin of the archipelago. This is re- explain satisfactorily the observed high Th/Ta flected in the latter having comparatively higher values in Taal Volcano and other volcanoes in 143Nd/‘44Nd and lower s7Sr/86Sr values, as well the western Philippines. as less radiogenic common Pb with Dupal-type isotope characteristics. The most logical enriched ~rn~nent available to the Manila-Negros sub- 8. Conclusions duction zone, which is completely missing on the eastern side of the archipelago, is South China From the above the following conclusions can Sea basin sediments, derived largely from Eura- be drawn: sia. (1) The source materials of lavas from Taal and other western Philippine volcanoes have undergone recent enrichment in LILE, as indicated Acknowledgments by the decoupling of isotopic and elemental ratios. Their derivative magmas have low Sm$Nd We are indebted to personnel at the Philip- relative to bulk earth, and yet their 143Nd/i 4Nd pine Institute of Volcanology and Seismology, values exhibit a time-integrated record of deple- particularly Director Ray Punongbayan, for assis- tion. Rb/Sr values of the magmas are also con- tance with logistics in the field. Research funding siderably higher than the bulk earth value, for this study came chiefly from National Science whereas corresponding s7Sr,/%r values are ei- Foundation grants EAR-8721416 (M.F.J.F) and ther less than, or are close to, the estimated bulk EAR-8996316 (S.B.M.). Thoughtful reviews have earth ratio. been provided by Pat Castilio, John EncamaciBn (2) LILE enrichment in Iavas from TaaI, La- and Robert Stern, to whom we are most grateful. guna de Bay and Arayat volcanoes cannot be entirely due to small degrees of partial melting. The trace element enrichment is accompanied by References substantial eIevations in radiogenic Pb, a feature best explained by the introduction of subducted, Arcuius,R.J., 1981. Island arc magmatism in relation to the continentally derived sediments. evolution of the crust and mantle. Tectonophysics, 75: (3) Lavas from the three western Philippine 113-133. Arculus, R.J. and Johnson, R.W., 1981. Island arc magma volcanoes studied and from the post-spreading sources: A geochemical assessment of the roles of slab-de- group in the South China Sea have some small, rived com~nents and crustal contamination. Geochem. J., but nevertheless important, differences in both 15: 109-133. elemental and isotopic ratios, requiring different Balce, G.R., Crispin, O.A., Samaniego, S.M. and Miranda, source materials. Although Sm/Nd and Rb/Sr F.E., 1981. Metallogenesis in the Philippines. Ceol. Surv. Jpn. Rep., 261: 125-148. values for basaltic rocks from the two areas are Cardwell, RX., Isacks, B.L. and Karig, D.E., 1980. The spa- virtually the same, western Philippine volcanoes tial distrjbution of earthquakes, focal mechanism solu- have consistently higher s7Sr/s6Sr, Sr/Nd, and tions, and subducted lithosphere in the Philippine and 220 S.3. Mukasa et al. / Tectonophysics 235 (1994) 20.5-221

northeastern Indonesian islands. In: D.E. Hayes (Editor), isotopic data of Indian Ocean Ridges: new evidence of The Tectonic and Geologic Evolution of Southeast Asian large-scale mapping of mantle heterogeneities. Earth Seas and Islands. Am. Geophys. Union Geophys. Monogr., Planet. Sci. Lett., 76: 288-298. 23: l-35. Hamilton, W.B., 1979. Tectonics of the Indonesian region. US Cohen, R.S. and O’Nions, R.K., 1982. The lead, neodymium Geol. Surv. Prof. Pap. 1078: l-345. and strontium isotopic structure of ocean ridge basalts. J. Hart, S.R., 1984. A large-scale isotopic anomaly in the south- Petrol., 23: 299-324. ern hemisphere. Nature, 309: 7.52-757. Cruz, J. and Umbal, J., 1983. Volcanology and Philippine Hart, S.R., 1988. Heterogeneous mantle domains: Signatures, volcanism: some issues and problems. Phil. J. Volcanol., 1: genesis and mixing chronologies. Earth Planet. Sci. I.&t., 3-10. 90: 273-296. Davidson, J.P., 1986. Isotopic and trace element constraints Hashimoto, W., 1981. Geologic development of the Philip- on the petrogenesis of subduction-related lavas from Mar- pines. Geol. Paleontol. Southeast Asia, 22: 83-170. tinique, Lesser Antilles. J. Geophys. Res., 91: 5943-5962. ~awkeswo~h, C.J. and Ellam, R., 1989. Chemical flutes and De Boer, J.Z., Odom, L.A., Ragland, P.C., Snider, F.C. and wedge replenishment rates along recent destructive plate Tilford, N.R., 1980. The Bataan orogene: eastward sub- margins. Geology, 17: 46-49. duction, tectonic rotations, and volcanism in the western Hayes, D.E. and Lewis, SD., 1984. A geophysical study of the Pacific (Philippines). Tectonophysics, 67: 251-282. Manila Trench, Luzon, Philippines 1. Crustal structure, Defant, M.J., De Boer, J.Z. and Oles, D., 1988. The western gravity and regional tectonic evolution. J. Geophys. Res., Central Luzon volcanic arc, the Philippines: two arcs di- 89: 9171-9195. vided by rifting? Tectonophysics, 145: 305-317. Hickey-Vargas, R., 1991. Isotope characteristics of submarine Defant, M.J., Jacques, D., Maury, R.C., De Boer, J.Z. and lavas from the Philippine Sea: implications for the origin Joron, J.-L., 1989. Geochemistry and tectonic setting of of arc and basin magmas of the Philippine tectonic plate. the Luzon arc, Philippines. Geol. Sot. Am. Bull., 101: Earth Planet. Sci. L&t., 107: 290-304. 663-672. Hofmann, A.W., 1988. Chemical differentiation of the Earth: Divis, A.F., 1980. The petrology and tectonics of recent vol- the relationship between mantle, continental crust, and canism in the Central Philippine Islands. In: D.E. Hayes oceanic crust. Earth Planet. Sci. Lett., 90: 297-314. (Editor), The Tectonic and Geologic Evolution of South- Hsii, I., 1972. Magnetic properties of igneous rocks in the east Asian Seas and Islands. Am. Geophys. Union Geo- northern Philippines. PhD Diss., Washington Univ., St. phys. Monogr., 23: 127-144. Louis, MO, 165~~. Dupre, 3. and Allegre, C.J., 1983. Pb-Sr isotopic variations in Karig, D.E., 1983. Accreted terranes in the northern part of Indian Ocean basalts and mixing phenomena. Nature, 303: the Philippine Archipelago. Tectonics, 2: 211-236. 142-146. Karig, D.E., Sarewitz, D.R. and Haeck, G.D., 1986. Role of Fiirster, H., Oles, D., Knittel, U., Defant, M.J. and Torres, strike-slip faulting in the evolution of allochthonous ter- R.C., 1990. The Macolod Corridor: A rift crossing the ranes in the Philippines. Geology, 14: 852-855. Philippine island arc. Tectonophysics, 183: 26.5-271. Kay, R.W., 1980. Volcanic arc magmas: Implications for a Fuller, M.D., McCabe, R., Williams, I., Almasco, J., Encina, melting-mixing model for element recycling in the crust- R.Y., Zanoria, AS. and Wolfe, J.A., 1983. Paleomag- upper mantle system. J. Geol., 88: 497-522. netism of Luzon. In: D.E. Hayes (Editor), The Tectonic Kelemen, P.B., Johnson, K.T.M., Kinzler, R.J. and Irving, and Geologic Evolution of Southeast Asian Seas and Is- A.J., 1990. High-field-strength element depletions in arc lands, Part 2. Am. Geophys. Union Geophys. Monogr., 27: basalts due to mantle-magma interaction. Nature, 345: 79-94. 521-524. Gill, J.B., 1981. Orogenic Andesites and Plate Tectonics. Knittel, U., Defant, M.J. and Raczek, I., 1988. Recent enrich- Springer, Berlin, 39th~~. ment in the source region of arc magmas from Luzon Hamburger, M.W., Cardwell, R.K. and Isacks, B.L., 1983. island, the Philippines: Sr and Nd isotopic evidence. Geol- Seismotectonics of the Northern Philippine Island Arc. In: ogy, 16: 73-76. D.E. Hayes (Editor), The Tectonic and Geologic Evolu- Leeman, W.P., 1983. The influence of crustal structure on tion of Southeast Asian Seas and Islands, Part 2. Am. compositions of subduction-related magmas. J. VolcanoI. Geophys. Union Geophys. Monogr., 27: l-22. Geotherm. Res., 18: 561-588. Hamelin, B. and AHtgre, C.J., 1985. Large-scale regional Mahoney, J.J., Natland, J.H., White, W.M., Poreda, R., units in the depleted upper mantle revealed by an isotope Bloomer, S.H., Fisher, R.L. and Baxter, A.N., 1989. fso- study of the south-west Indian Ridge. Nature, 315: 196- topic and geochemical provinces of the Western Indian 198. Ocean spreading centers. J. Geophys. Res., 94: 4033-4052. Hamelin, B., Duprt, B. and Allegre, C.J., 1984. Lead- McCabe, R., Almasco, J. and Yumul, G., 1985. Terranes of strontium isotopic variations along the East Pacific Rise the Central Philippines. In: D. Howell (Editor), Tectonos- and the Mid-Atlantic ridge: a comparative study. Earth tratigraphic terranes of the circum-Pacific region. Coun. Planet. Sci. Lett., 67: 340-350. Energy Miner. Resour. Earth Sci. Ser., 1: 421-435. Hamelin, B., Dupre, B. and Allegre, C.J., 1986. Pb-Sr-Nd McCulloch, M.T. and Perfit, M.R., 1981. 143Nd/144Nd, 87Srf S.B. Mukasa et al. / Tectonop~ysics 235 (19941 20.5221 221

s6Sr and trace element constraints on the petrogenesis of Stern, R.J., 198X. A common mantle source for western Pa- Aleutian island arc magmas. Earth Planet. Sci. Lett., 56: cific island arc and “hot spot” magmas-Implications for 167-179. layering in the upper mantle. Carnegie Inst. Year Book, Miklius, A., Flower, M.F.J., Huijsmans, J.P.P., Mukasa, S.B. Washington, 81: 455-462. and Castillo, P., 1991. Geochemistry of lavas from Taal Sun, S.-s., 1980. Lead isotopic study of young volcanic rocks Volcano, southwestern Luzon, Philippines: Evidence for from mid-ocean ridges, oceanic islands and island arcs. multiple magma supply systems and mantle source hetero- Philos. Trans. R. Sot. London, Ser. A, 297: 409-455. geneity. J. Petrol., 32: 593-627. Sun, S.-s., Tatsumoto, M. and Schilling, J.G., 1975. Mantle Morris, J.D. and Hart, S.R., 1983. Isotopic and incompatible plume mixing along the Reykjanes ridge axis: lead isotope element constraints on the genesis of island arc volcanics evidence. Science, 190: 143. from Cold Bay and Amak Island, Aleutians, and implica- Tapponnier, P., Peltzer, G. and Armijo, R., 1986. On the tions for mantle structure. Geochim. Cosmochim. Acta, mechanics of the collision between India and Asia. In: 47: 201.5-2030. M.P. Coward and AC. Reis (Editors), Collision Tectonics. Morris, J. and Tera, F., 1989. “Be and “Be in mineral Geol. Sot. London Spec. Publ., 19: 115-157. separates and whole rocks from volcanic arcs: Implications Tatsumoto, M. and Nakamura, Y., 1991. DUPAL anomaly in for sediment subduction. Geochim. Cosmochim. Acta, 53: the Sea of Japan: Pb, Nd and Sr isotopic variations at the 3197-3206. eastern Eurasian continental margin. Geochim. Cos- Morris, J.D., Jezek, P.A., Hart, S.R. and Gill, J.B., 1983. The mochim. Acta, 55: 3697-3708 Halmahera island arc, Molucca Sea collision zone, Indone- Taylor, B. and Hayes, D.E., 1983, Origin and history of the sia: a geochemical survey. In: D.E. Hayes (Editor), The South China Sea Basin. In: D.E. Hayes (Editor), The Tectonic and Geologic Evolution of Southeast Asian Seas Tectonic and Geologic Evolution of Southeast Asian Seas and Islands, Part 2. Am. Geophys. Union Geophys. and Islands, Part 2. Am. Geophys. Union Geophys. Monogr., 27: 373-387. Monogr., 27: 23-56. Morris, J.D., Leeman, W.P. and Tera, F., 1990. The sub- Taylor, S.R., and McLennan, SM., 1985. The Continental ducted component in island arc lavas: constraints from Be Crust: its Com~sition and Evolution. Blackwell, Oxford, isotopes and B-Be systematics. Nature, 344: 31-36. UK, 312~~. Mukasa, S.B., McCabe, R., Gill, J.B. and Newhall, C., 1986. Tera, F., Brown, L., Morris, J., Sacks, I.S., Klein, J. and Philippine island arc Pb-isotopic compositions: the Dupal Middleton, R., 1986. Sediment incorporation in island-arc anomaly in a northern hemisphere island arc? EOS, Trans. magmas: inferences from “Be. Geochim. Cosmochim. Am. Geophys. Union, 67: 1273. Acta, 50: 535-550. Mukasa, S.B., McCabe, R. and Gill, J.B., 1987. Pb-isotopic Tilton, G.R. and Barreiro, B.A., 1980. Origin of lead in compositions of volcanic rocks in the West and East Andean talc-alkaline lavas, southern Peru. Science, 210: Philippine island arcs: presence of the Dupal isotopic 1245-1247. anomaly. Earth Planet. Sci. Lett., 84: 153-164. Tu, K., Flower, M.F.J., Carlson, R.W., Zhang, M. and Xie, Mukasa, S.B., Flower, M.F.J. and Arcilla, C., Differences in G., 1991. Sr, Nd, and Pb isotopic compositions of Hainan the isotopic and trace element compositions of volcanic basalts (south China): Implications for a subcontinental chains associated with the op~sitely-dipping Philippine lithosphere Dupal source. Geology, 19: 567-569. subduction zones: constraints on the character of the Varne, R., 1985. Ancient subcontinental mantle: a source of source components. Geochim. Cosmochim. Acta (in re- K-rich erogenic volcanics. Geology, 13: 405-408. view.). Varne, R. and Foden, J.D., 1986. Geochemical and isotopic Navon, 0. and Stolper, E., 1987. Geochemical consequences systematics of Eastern Sunda arc volcanics: implications of melt percolation: the upper mantle as a chromato- for mantle sources and mantle mixing processes. In: F.C. graphic column. J. Geol., 95: 285-307. Wezel (Editor), Origin of Arcs. Elsevier, Amsterdam, pp. Rangin, C., Stephan, J.F., Blanchet, R., Baladad, D., Bouysee, 159-189. P., Chen, M.P. and Chotin, P., 1988. Seabeam survey at Vidal, Ph. and Clauer, N., 1981. Pb and Sr isotopic systemat- the southern end of the Manila trench. Tectonophysics, its of some basalts and sulfides from the East Pacific Rise 146: 261-278. at 21”N (project RITA). Earth Planet. Sci. Lett., 55: 237- Rogers, N.W., Hawkesworth, C.J., Parker, R.J. and Marsh, 246. J.S., 1985. The geochemistry of potassic lavas from Vulsini, White, W.M. and Dupre, B., 1986. Sediment subduction and central Italy and implications for mantle enrichment pro- magma genesis in the Lesser Antilles: isotopic and trace cesses beneath the Roman region. Contib. Mineral. Petrol., element constraints. J. Geophys. Res., 91: 5927-5941. 90: 244-257. Wolfe, J.A. and Self, S., 1983. Structural lineaments and Ru, K. and Pigott, J.D., 1986. Episodic rifting and subsidence Neogene volcanism in southwestern Luzon. In: D.E. Hayes in the South China Sea. Am. Assoc. Pet. Geol. Bull., 70: (Editor), The Tectonic and Geologic Evolution of South- 1136-1155. east Asian Seas and Islands, Part 2. Am. Geophys. Union Ryerson, F.J. and Watson, E.B., 1987. Rutile saturation in Geophys. Monogr., 27: 157-172. magmas: Implications for Ti-Nb-Ta depletion in island- arc basalts. Earth Planet. Sci. Lett., 86: 225-239.