Silver, E. A., Rangin, C , von Breymann, M. T., et al., 1991Proceedings of the Ocean Drilling Program, Scientific Results, Vol. 124

20. PETROLOGY OF IGNEOUS ROCKS FROM THE CELEBES SEA BASEMENT1

G. Serri,2 P. Spadea,3 L. Beccaluva,4 L. Civetta,5 M. Coltorti,4

J. Dostal,6 F. Sajona,7 C. Vaccaro,4 and O. Zeda8

ABSTRACT

Major and trace elements, mineral chemistry, and Sr-Nd isotope ratios are reported for representative igneousrocks of Ocean Drilling Program Sites 767 and 770. The basaltic basement underlying middle Eocene óradiolarian-bearing red clays was reached at 786.7 mbsf and about 421 mbsf at Sites 767 and 770, respectively. At Site 770 thebasement was drilled for about 106 m. Eight basaltic units were identified on the basis of mineralogical,petrographical, and geochemical data. They mainly consist of pillow lavas and pillow breccias (Units A, B, D, andH), intercalated with massive amygdaloidal lavas (Units Cl and C2) or relatively thin massive flows (Unit E). Twodolerite sills were also recognized (Units F and G).

All the rocks studied show the effect of low-temperature seafloor alteration, causing almost total replacement ofolivine and glass. Calcite, clays, and Fe-hydroxides are the most abundant secondary phases. Chemical mobilizationdue to the alteration processes has been evaluated by comparing elements that are widely considered mobile duringhalmyrolysis (such as low-field strength elements) with those insensitive to seafloor alteration (such as Nb). Ingeneral, MgO is removed and P2O5 occasionally enriched during the alteration of pillow lavas. TI, Cs, Li, Rb, andK, which are the most sensitive indicators of rock/seawater interaction, are generally enriched. The most crystallinesamples appear the least affected by chemical changes.

Plagioclase and olivine are continuously present as phenocrysts, and clinopyroxene is confined in thegroundmass. Textural and mineralogical features as well as crystallization sequences of Site 770 rocks are, in all,analogous to typical mid-ocean-ridge basalts (MORBs). Relatively high content of compatible trace elements, suchas Ni and Cr, indicate that these rocks represent nearly primitive or weakly fractionated MORBs.

All the studied rocks are geochemically within the spectrum of normal MORB compositional variation. TheirSr/Nd isotopic ratios plot on the mantle array (87Sr/87Sr 0.70324-0.70348 with 143Nd/144Nd 0.51298-0.51291) outsidethe field of Atlantic and Pacific MORBs. However, Sr and Nd isotopes are typical of both Indian Ocean MORBs andof some back-arc basalts, such as those of Lau Basin. The mantle source of Celebes basement basalts does not showa detectable influence of a subduction-related component.

The geochemical and isotopic data so far obtained on the Celebes basement rocks do not allow a cleardiscrimination between mid-ocean ridge and back-arc settings.

INTRODUCTION

One of the major objectives of Leg 124 was to study thenature of the basement of the Sulu and Celebes seas, twosmall basins located in a structurally complex region wherethe Eurasian, Philippine Sea, and Australian plates converge.This goal was successfully reached in both basins (Fig. 1).

In particular, the basement of the Celebes Sea was drilledin two sites, 767 and 770, for lengths of 0.4 m and 106 m,respectively. Radiolarian-bearing red clays found directlyabove the basement gave a middle Eocene age (about 43 Ma)for both sites. The basement consists of a sequence of basalticpillow- and pillow-breccia lavas and sheet flows injected bydolerite sills. Preliminary on-board analyses revealed a

1 Silver, E. A., Rangin, C , et al., von Breymann, M. T., et al., 1991. Proc.ODP, Sci. Results, 124: College Station, TX (Ocean Drilling Program).

Dipartimento di Scienze della Terra, Universita di Pisa, Via S. Maria, 53,1-56100 Pisa, Italy.

3 Istituto di Scienze della Terra, Universita di Udine, Via Cotonificio, 114,1-33100 Udine, Italy.

4 Istituto di Mineralogia, Universita di Ferrara, Corso Ercole I D'Este, 32,1-44100 Ferrara, Italy.

5 Dipartimento di Geofisica e Vulcanologia, Universita di Napoli, Largo S.Marcellino 10, 1-80138 Napoli, Italy.

6 Department of Geology, St. Mary's University, Halifax, Nova Scotia B3H3C3, Canada.

7 Petrolab. Mines and Geosciences Bureau, North Avenue, Quezon City,Philippines.

Istituto di Mineralogia, Universita di Parma, Viale delle Scienze, 78,1-43100 Parma, Italy.

tholelitic composition, showing a petrogenetic affinity withmid-ocean ridge basalts (MORB).

Despite the number and quality of data obtained during Leg124, the origin of the Celebes Sea is still under discussion. Thesedimentary record indicates that the basin developed farfrom continental or major volcanic arc sources, but does notdistinguish between the following hypotheses: (1) trappedfragment of the Indian Ocean or the Philippine Sea oceaniccrusts (Lee and McCabe, 1986; Silver, et al. 1989); (2)fragmentation and rifting of the Eurasian continental marginwith a mechanism analogous to the formation of the SouthChina Sea (Rangin, 1989); (3) inter-arc, subduction-relatedmarginal basin similar to the Parece-Vela Basin.

The nature of the basaltic crust often makes it possible torecognize the different geodynamic settings in which marginalbasins may originate. Though generally similar to tholeiiteserupted at mid-ocean ridges, the basalts from subduction-related marginal basins may show some distinctive mineral-ogical, textural, and geochemical features due to the influenceof subducted lithosphere on magma generation processes.This is widely considered as related to the variable addition tothe mantle wedge of volatile and large-ion lithophile elements(LILE) derived from a subducted lithospheric slab (Saunderset al., 1984). The geochemical signature produced in thisprocess may be detected in the basalts erupted in intra-arc andback-arc settings, particularly at an early stage of marginalbasin opening. Recently it has been shown that some back-arcbasin basalts, such as those from Lau (Volpe et al., 1988) andnorth Fiji (Price et al., 1990) basins, range from normal to

271

G. SERRI ET AL.

" 10° N

- 5°N

- 0 c

120°E 125° E

Figure 1. Location of the sites drilled during ODP Leg 124. Bathymetric contours are in meters.Oceanic basement is hatched.

enriched MORBs, and even ocean island basalt-like compo-nents are occasionally present.

Mineral chemistry and texture give further indicationsabout crystal-liquid equilibria during fractionation processesin different settings: these provide alternative criteria to infermagmatic processes in altered rocks in which the content ofvolatile and most large-ion lithophile (LIL) elements waschanged by secondary mobilization processes.

This work reports petrological, mineralogical, geochemi-cal, and isotopic data obtained in various shore-based labora-tories with the aim of defining the nature of the Celebes Seabasement, and putting constraints derived from the chemicalgeodynamics on the tectonic setting of its origin.

ANALYTICAL METHODMinerals (Tables 2, 3, and 4) were analyzed at the Univer-

sity of Milan using an electron-probe microanalyzer withSEMQ-ARL unit at an accelerating voltage of 15 kV, speci-

men current of 5 nA, beam diameter of 2 to 3 µm and countingtime of 20 s. The Bence and Albee factors (1968) were used fordata reduction.

X-ray fluorescence (XRF) analyses (Tables 5 and 6) weremade on board using ODP techniques on whole-rock samples(Rangin, Silver, von Breymann, et al., 1990). At the Univer-sity of Udine, XRF analyses were made for major elements,Cr, and Se on lithium borate glass disks (flux to sample ratio10:1 to overcome matrix effects) and for trace elements (V,Ni, Rb, Sr, Y, Zr, and Nb) on powder pellets (Comptonscattering technique adopted for matrix absorption correc-tions), using a Philips wavelength-dispersive automated spec-trometer. Rare-earth elements (REE) and Y were determinedin the Centre de Recherches Pétrographiques etGéochimiques (C.R.P.G.) of Nancy, France, by inductivelycoupled plasma (ICP) emission spectrometry (columns A ofTable 6). Se, V, Li, Be, Cs, Rb, Ba, Sr, Zr, Hf, Th, U, TI, Pb,Nb, Y, and REE were analyzed at St. Johns Memorial Uni-

272

CELEBES SEA IGNEOUS ROCKS

versity by ICP-mass spectrometry (ICP-MS), using the stan-dard addition method to correct for matrix effects (columns Bof Table 6). The analytical procedure is described in detail inJenner et al. (1990). Cross checking of trace-element analysesdone on the same sample at Udine, St. Johns, Nancy, and onboard allows the following conclusions to be drawn:

1. Major-element analyses are in good agreement exceptfor SiO2 and Na2O, which are systematically higher by 0.6%-1.6% and 0.4%-0.6%, respectively (absolute values) in Udinewith respect to on-board determination.

2. Excellent agreement between the different laboratories hasbeen found for elements such as Rb, Sr, and Nb. Note that, forthese elements, St.Johns and shipboard determinations are vir-tually identical. Zr and Y determinations by XRF are systemat-ically higher than ICP-MS analyses by about 10% and 15%,respectively. Y determination by ICP (Nancy) is systematicallyhigher by about 20% with respect to ICP-MS determination(except Section 124-770C-10R-2, with ICP 40% higher).

3. There is sufficient agreement between REE analyses byICP and ICP-MS; within a span of 10%-15% between data.

4. In general, agreement between XRF determination (par-ticularly that obtained on board) and ICP-MS data for V isgood, except one Sample (124-770B-20R-3, 10-15 cm).

5. For Sample 124-770C-7R-6, 44-47 cm, the powderanalyzed on the ship and in Udine are different. The Udineanalysis is reported because ICP and ICP-MS determinationwas done on the same powder.

6. For Sample 124-770C-12R-2, 38-41 cm, which is amedium-grained olivine dolerite, ICP-MS analyses might notbe representative of the bulk rock composition due to theexceedingly small amount of sample (100 mg) analyzed forsuch a non-homogeneous lithotype. The wide discrepancybetween XRF and ICP-MS analyses for Sr, Zr, Y, Rb, and Vcan be cited to support this argument.

Sr and Nd isotopic compositions were obtained at theUniversity of Naples after cold leaching with 2.5 N (A ofTable 6) or 6.5 N (the latter only for Sr, B of Table 6)hydrochloric acid for about 30s. Measurements were donewith a VG 354 mass spectrometer. For the reference sampleNBS987, the 87Sr/86Sr ratio = 0.71026 ± 2 and for La JollaSTD the 143Nd/144Nd ratio = 0.51186 ± 3 were obtained. Thereported errors are at the 95% confidence level.

GEOLOGICAL OUTLINEThe Celebes Sea is 270,000 km2 in extent and has an

abyssal plain of 5 to 5.5 km deep. It is bounded to the north bythe Sulu Archipelago, which belongs to the Sulu volcanic arcextending from Borneo to the Zamboanga Peninsula of Min-danao. Deep bathymetric depressions mark the trenches ofthe active North Sulawesi subduction zone in the south andthe Cotobato subduction zone in the west.

At Site 767, located in the central Celebes Sea at 4° 47.5' N,123° 30.2' E at a water depth of 4905 m, the basement wasreached at 786.7 meters below seafloor (mbsf). middle Eocenered clays rest directly over the lavas. A total of 42 cm ofbasaltic rocks were recovered in the core catcher.

Three holes were drilled at Site 770, which is located on abathymetric high at a water depth of 4505 m about 43 kmnorth-northeast of Site 767. Hole 770B was drilled to 474mbsf, including 50 m of basement. Hole 770C washed tobasement, then cored to 529 mbsf, or 106 m of basement.

The basal sediments at Sites 770 and 767 are middle Eoceneradiolaria- bearing red clays with a biostratigraphic age of 42Ma (Silver et al., 1989) or 45 Ma (Rangin, 1989).

Site 770 has about half the thickness of the sedimentarycover of Site 767, yet the basal ages are identical. Thisdifference reflects the higher elevation of Site 770, protectingit from turbidite influence and allowing it to be just above thecalcite compensation depth (CCD) in the lowest part. Base-ment was recovered from 421 to 529 mbsf. Holes 770B and770C were drilled so close together as to make a separatedescription of basement lithology unnecessary. The contactbetween sediment and basement was found at 421 m in Hole770B, and about 2 m deeper in Hole 770C.

LITHOLOGYCombining shipboard results from Holes 770B and 770C

with shore-based mineralogical, petrographical, and geochem-ical analyses, eight basaltic units have been identified (Fig. 2).To avoid confusion with shipboard units, the newly definedunits were named from top to bottom with capital letters inalphabetic order and divided as follows:

—Units A, B, and H are pillow basalt sequences, with thelower part of Unit B consisting of pillow breccia;

—Unit C (Cl and C2) consists of brecciated, massiveamygdaloidal lavas cemented with calcite veins;

—Unit D consists of pillow lavas and pillow breccias;—Unit E is made of relatively thin massive lava flows;—Units F and G are identified as dolerite sills.Shipboard observations indicate that small dikes and dike-

lets, maximum 30 cm, penetrate Units D, E, and H.

Unit A (Shipboard Unit 1)

Sparsely Plagioclase Olivine Phyric Basalt Pillow LavasThis unit has a thickness of about 18 m, of which approx-

imately 7 m (Hole 770B) and 6 m (Hole 770C) were recovered.It consists of sparsely Plagioclase (2%-4%) and olivine (2%-3%) phyric pillow basalts, whose quench textures indicate acrystallization at a very high cooling rate (cf. Lofgren, 1979).Fan and bow-tie morphologies of the groundmass plagioclaseswere found only in samples from the upper part of thesequence. The texture of the studied samples recovered fromthe lower part of Unit A (Table 1) indicates a lower degree ofsupercooling, also reflected by the lower amount of mesosta-sis (10%-30% vs. 40%-70%). This is in accordance with theshipboard observation, which identifies within Unit A upperand lower parts made up respectively of pillow lavas of smalland large (0.4 m-1 m in diameter) dimensions.

Some samples appear to be nonvesicular; others containfew vesicles (up to 2%) often filled with clays and carbonates.

Units B and C: a Modified Interpretation of theLithology

There is not a straightforward correlation between therocks found in Holes 770B and 770C.

Shipboard unit 3 of Hole 770B, which has been inter-preted as a brecciated pillow lava sequence about 4 m thick(454.8-458.6 mbsf), has no equivalent at about the samedepth in Hole 770C. Here, between 452.2 and 460.8 mbsf,there is a basaltic breccia essentially made up of angularclasts and separated by a network of calcite veins. This islithologically comparable to Unit 4 of Hole 770B (458.6-464.5 mbsf), which is considered to be a brecciated andveined massive lava flow.

The re-evaluation of shipboard data coupled with newpetrochemical analyses allows us to divide Unit 3 of Hole770B into an upper section (454.8-457.0 mbsf), which consistsof highly Plagioclase -olivine phyric pillow breccia, and alower section (457.0-458.6 mbsf) that is made of a breccia

273

G. SERRI ET AL.

CO

- Q

E

α>>ooα>er

CD

—OX

u

oX

_cüoα

LUC

O)

oo

400

420 -

440 -

460 -

480 -

500 —

520 -

17R

19R

21 R

2 R

3 R

5 R

B

D

G

H

.̂•. .T •••̂ .•-i.•.̂• . i . . i . . Λ. . Λ • ΛΔ. .Λ. .Λ- . Λ- . Δ .

.Δ.'Λ'.Δ. .Δ. Δ.. . .\ . . . .•.c

-•.Δ. .Δ.'.Δ. .Δ. Δ.•^ ^ Λ Δ . . Δ Λ Λ <

.Δ • Λ Δ Δ. '

FOR AMINIF ER-N ANNO F0 S S1 L MAR LUnit HC

DARK GREY CLAY STONEUnit HC

PILLOW LAY ASUnit A: sparsely PI-OI phytic basalt pillow lavas

PILLOW LAYAS AND PILLOW BRECCIASUnit B: highly PI-OI phyric basalts, pillowed andbrecciated

MAS S1 YE LAYASUnitCl and C2: moderately to highly PI-OI phyr-ic amyg dal oi dai basalt lava fi ovs, breccia* e d an dveined

PILLOW LAYAS AND PILLOW BRECCIASUnit D: moderately to highly PI-OI phyric basalt-ic pillowlava3 and pillow breccias

MASSIYE LAYAS WITH DIKESUnit E: moderately PI-OI phyric basalt massivelavas spanely intruded by basaltic dikes

DOLERITESILLUnit F: sparsely PI phyric to aphyric doleritesill

OLIYINE DOLERITE SILLUnit G: s para ely t o m o d erat ely Ol-Pl p hyric d ol-eritesill

PILLOW LAYAS WITH DIKESUnit H: m o d erat ely PI-OI p hyric basaltic pill owlavas cut by basaltic dikes

Figure 2. Merged stratigraphic log of Site 770 Holes B and C - Celebes Sea basement.

with angular clasts of a massive lava cemented by veinsdominantly of calcite and clays.

Conclusions

—Unit B (pillow lavas and pillow breccias) of the modifiedlithostratigraphic sequence of Site 770 basement comprisesUnit 2 of Holes 770B and 770C as well as the upper part ofUnit 3 of Hole770B;

—Unit Cl includes Unit 3 of Hole 770C, Unit 4 of Hole770B, and the lower section (457.0-458.6 mbsf) of Unit 3 ofHole 770B;

—Unit C2 comprises Units 5 and 6 of Hole B and Unit 4 ofHole 770C.

UnitB

Highly Plagioclase and Olivine Phyric Basaltic Pillow Lavasand Pillow Breccias

This unit shows a thickness in Hole 770B nearly twice asgreat as that presented in Hole 770C (17.7 m vs. 9.5 m) ofwhich 7 m (Hole 770B) and 4 m (Hole 770C) were recovered.

The rocks are pillow basalts that become brecciated at thebase of the newly defined unit. These pillow lavas are easilydistinguished from those of Unit A by their high content ofphenocrysts of Plagioclase (109£ 20%) and olivine (4%-\0%).Groundmass textures and amount of mesostasis are compara-

274

CELEBES SEA IGNEOUS ROCKS

ble to the upper Unit A that petrographically represents themost quenched facies of Site 770 (Table 1).

Vesicles/amygdales of Unit B are, on average, slightlymore abundant (up to 4%) and of bigger size (up to 5 mm) thanin Unit A.

Unit C (Cl + C2)

Moderately to Highly Plagioclase and Olivine PhyricAmygdaloidal Basalt Massive Lavas, Brecciated and Veined

The thickness of this unit is comparable in the two holes(770B, about 17 m of which 7.5 m recovered; 770C, about 20m, of which 14 m recovered).

Lithologically, it consists of amygdaloidal massive lavaflows showing strong evidence of in-situ brecciation andcementation through a network of veins, predominantly cal-cite, and subordinate clays, silica, and Fe-hydroxides. Chilledmargins are absent throughout the sequence, but texturalvariation (Table 1) suggests the existence of at least threecooling units.

The studied samples are moderately Plagioclase (3%-10%)and olivine (l%-10%) phyric basalts with intersertal, inter-granular, interfasciculate and subophitic groundmass: tex-tures that are normally found in such types of massive lavaflows. The lowermost studied sample of Unit C2 is a highlyPlagioclase (15%) phyric basalt, characterized by rare plagio-clase megacrysts greater than 1 cm in size.

The distribution of the amygdales varies throughout thesequence from 2% to 15%. Altered mesostasis is alwayspresent. In contrast with the pillow lava sequences of Units Aand B, the mesostasis is volumetrically subordinate andinvariably occurs as intersertal patches.

Unit D (Shipboard Unit 5 of Hole 770C)

Moderately to Highly Plagioclase and Olivine Phyric BasalticPillow Lavas and Pillow Breccias

This unit has a thickness of 18.7 m. (472.1-490.8 mbsf) ofwhich about 5.5 m were recovered.

Lithologically, it consists of pillow lavas, commonly 10-20cm in diameter, with pillow breccia horizons. The studiedsamples are characterized by variable abundance of Plagio-clase (8%-15%) and olivine (4%) phenocrysts set in a ground-mass with quench textures dominated by spherulitic crystalsof Plagioclase.

The vesicles (l%-4%), typically filled with green andbrown clays and carbonate, are heterogeneously distributedthrough the sequence.

Shipboard observation suggests that Sample 124-770B-7R-6, 43-44 cm can be interpreted as a cryptocrystallinedikelet about 5 cm thick intruded in the brecciated part of theunit.

Unit E (Shipboard Unit 6 of Hole 770C)

Moderately Plagioclase and Olivine Phyric Basalt MassiveLavas cut by Dikes

This unit is 10.5 m thick (490.8-501.3 mbsf); about 3 mwere recovered by drilling.

The lithology is not well constrained. Shipboard observationsuggests that this sequence consists of three separate coolingunits cut by flat-lying minor intrusions. The studied sample wasinterpreted as a small dike. This is in accordance with thephaneritic medium grain size, subophitic texture, and virtualabsence of vesicles/amygdales and mesostasis (Table 1).

Unit F (Shipboard Unit 7)

Sparsely Plagioclase Phyric to Aphyric Dolerite Sill

This is a relatively homogeneous massive unit 11.2 m. thick(501.3-512.5 mbsf) of which 5.5 m were recovered.

The topmost 25 cm of the unit is characterized by about10% of spherical vesicles/amygdales that become scattered orvirtually absent going downward. The unit is classified as adolerite sill with uniform phaneritic appearance. The rocks aresubaphyric with uneven distribution of Plagioclase phenoc-rysts up to 2%. The texture is subophitic, with a significantamount (about 15%) of interstitial mesostasis (Table 1).

Unit G (Shipboard Unit 8)

Sparsely to Moderately Olivine and Plagioclase Phyric DoleriteSill

This unit has a thickness of 10.6 m (512.5-523.1 mbsf) ofwhich 3.5 m were recovered.

Lithologically, it is a massive dolerite sill characterized bya variation of both grain size and abundance of phenocrystswith depth. The lower part is coarser-grained and moreporphyritic than the upper part. The two studied samples weretaken near the base of the sill. Despite their proximity (15 cmapart), they show a significant variation in grain size fromcoarse-medium- grained (Sample 124-770C-12R-2, 31-33 cm)to fine-medium-grained (Sample 124-770C-12R-2, 38-41 cm)rocks. In the latter, the abundance of olivine phenocrysts(Table 1) shows a heterogeneous distribution even at the scaleof thin section. Vesicles are virtually absent.

Unit H (Shipboard Unit 9)

Moderately Plagioclase and Olivine Phyric Basalt Pillow Lavaswith Dikes

This is the lowermost unit of Site 770. Its thickness is 6.4 m(523.1-529.5 mbsf). Shipboard observation suggests that mostof the recovered rocks (about 1.5 m) represent pillow lavas.Groundmass texture, amount of mesostasis (about 20%) andpresence of abundant vesicles/amygdales (about 6%) in thesingle studied sample (Table 1) do not contradict this interpre-tation.

PETROGRAPHY

Primary FeatureThe wide textural variation observed in the Site 770 rocks

is a common feature of submarine basaltic rocks and largelydepends on the mode of crystallization of the magmas at, orby, the seafloor. Well-developed quench textures with a highbut variable (20%-70%) amount of cryptocrystalline/glassymesostasis were found in the pillow lava sequences (Units A,B, and D, Table 1). In addition to the skeletal/spheruliticplagioclases and acicular/dendritic clinopyroxenes, thegroundmass of the studied pillows is characterized, particu-larly in Unit A and D, by skeletal/branching olivine. Incontrast, the massive lava flow sections (Units C and E) anddolerite sills (Units F and G) generally show groundmasses ofsubophitic texture with a low, but variable (0%-10%) amountof mesostasis, typically as intersertal patches.

Although subaphyric to highly phyric types are found, thephenocryst mineralogy of Site 770 rocks is relatively homo-geneous. Plagioclase (1%- 20%) is constantly present, andsystematically dominant over olivine (0%- 10%) in all thestudied samples (Plagioclase - olivine phyric basalts), except

275

G. SERRI ET AL.

Table 1. Synthetic petrographic description and modal proportion (% volume) ofmajor phase from Celebes basement rocks.

HoleCore-Sec.Int. (cm) Depth

UNIT A—Sparsely Pl-Ol770B16R-495-96770C2R-358-61770B17R-3134-137770C3R-24-7770C3R-3128-129770B18R-290-93

422.6

426.7

430.1

434.4

437.1

438.0

UNIT B—Strongly Pl-Ol770B18R-3144-145770C4R-126-27770B19R-198-99770B19R-242-45

440.0

442.9

446.1

447.6

Phenocrysts Ground mass*

Phyric Pillow Lavasol3, pl3

ol2, pl3

ol2, pl3

ol2, pl3

ol2, pl4

ol2, pl2

cpx2, meso60

cpx5, meso40 vesl

cpxl, meso60 ves2

cpx5, meso50 ves2

cpx30, meso20

cpx20, meso30

Textural features

skeletal/spherulutic pidendritic cpx

skeletal/spherulutic pidendritic cpx

skeletal/spherulutic pidendritic cpx

skeletal/spherulutic pidendritic cpx

tabular/skeletal piacicular/dendritic cpx

skeletal/acicular pidendritic cpx

Phyric Pillow Lavas and Pillow BrecciasollO, pll5

ol4, pllO

ol5, p!20

ol8, p!20

cpx2, meso50ves2

cpx2, meso40 vesl

meso60 ves3

cpx5, meso40 ves4

skeletal/spherulitic pidendritic cpx

skeletal/spherulitic pidendritic cpx

skeletal/spherulitic pidendritic cpx

skeletal/spherulitic pidendritic cpx

UNIT Cl and C2—Moderately to Strongly Pl-Ol Phyric, Brecciated, and Veined

770C5R-239-41770B20R-2101-102770B20R-312-14770B20R-4135-140770C5R-770-71770C6R-344-45770B21R-199-100770B21R-434-35770B21R-477-78770B21R-618-19

(amygdaloidal) Massive Basaltic Lava Flows453.9

457.2

457.6

460.3

460.6

465.1

465.5

468.7

469.2

471.3

ol5, pi10

ol5, pllO

ol2, pl5

ollO, pl5

oll,pl4

oll,pl4

ol5, pl3

ol5, pl6

ol5, pl5

ol5, p!15

cpx30, meso3 ves8

cpx30, meso5 ves2

cpx30, mesol2 ves3

cpx30, meso5 vesl5

cpx35, mesol2 ves8

cpx30, meso4 ves8

cpx35, meso8 vesl5

cpx30, meso5 vesl5

cpx30, mesoδ ves3

cpx35, meso5 vesl5

UNIT D—Moderately to Strongly Pl-Ol Phyric Pillow Lavas and770C7R-53-6770C7R-643-44

477.2

478.9

ol4, pll5

ol4, pl8

meso70 vesl

cpx5, meso60 ves4

UNIT E—Moderately Pl-Ol Phyric Massive Lavas cut by Dykes

intersertal

interfasciculateintergranular/intersertal

subophitic with intersertalpatches

subophitic with intersertalpatches

subophitic with intersertalpatches

interfasciculateintergranular/intersertal

subophitic with intersertalpatches

subophitic with intersertalpatches

intergranular/intersertal

subophitic with patches ofaltered mesostasis

Pillow Brecciasskeletal/spherulitic pi

skeletal/spherulitic pi

276

CELEBES SEA IGNEOUS ROCKS

Table 1 (continued).

HoleCore-Sec.Int. (cm)

770C9R-242-48

Depth

492.7

Phenocrysts

ol5, pl4

Groundmass*

cpx35

UNIT F—Sparsely PI Phyric to Aphyric Dolerite Sill770C10R-242-43770C11R-196-98

502.4

511.2

pH

pl2

cpx40, mesolO

cpx40, mesolO

UNIT G—Sparsely to Moderately Ol-Pl Phyric Dolerite Sill770C12R-221-33770C12R-238-41

521.4

521.6

oll5, pl5

ollO, pl5

cpx30, mesol7

cpx35, meso5

UNIT H—Moderately Pl-Ol Phyric Lavas770C12R-366-67

523.3 o!3, pll2 cpx30, meso20 ves6

Textural features

subophitic

subophitic with intersertalpatches

subophitic with intersertalpatches

subophitic

subophitic

interfasciculate/intersertal

Abbreviations: ol, olivine; pi, Plagioclase; cpx, clinopyroxene; meso, mesostasis; ves, vesicles. Heremesotasis refers to the sum of cryptro-crystalline material and plan.

* Among groundmass minerals only the clinopyroxene modol proportion is reported; vesiclesincludes also partial filling with secondary material and amygloles.

the sill of Unit G, which can be defined as an olivine(10‰l 5%) - Plagioclase (about 5%) phyric doleritic basalt.Cr-spinel is occasionally found as individual crystals and/orincluded in olivine and, more rarely, in Plagioclase phenoc-rysts. This is particularly evident in the pillows of Units B andD. Clinopyroxene and Fe-Ti oxides are always confined to thegroundmass. The prevailing opaque is Ti-magnetite, but il-menite has been identified as an additional oxide in most of themassive lavas and dolerite sills.

Textural observation of the quenched rocks indicate thatseafloor crystallization invariably took place in the followingsequence: olivine + Plagioclase, clinopyroxene, Fe-Ti oxides.Spatial relation among phenocrysts indicates that, in thepre-emptive stage (magma chamber?) the Cr-spinel appearson the liquidus earlier than olivine and Plagioclase. In a fewoccurrences, particularly in Unit B, olivine is found includedin Plagioclase phenocrysts, but not the other way around. Thissuggests that, for the less evolved magmas, olivine precedesPlagioclase on the liquidus. Sparse Plagioclase and olivineclusters as well as plagioclase-olivine glomerocrysts are oftenpresent, particularly in the most phyric rocks.

Textural and mineralogical features as well as crystallizationsequences of Site 770 rocks are, in all, analogous to typical,relatively primitive mid-ocean- ridge basalts (MORBs).

Alteration

All the rocks of Site 770 suffered the effects of low-temperature seafloor alteration that produced a variation incolor from dark gray to brownish gray to brownish. This oftengives to the rocks a mottled appearance, occasionally even inthin section, particularly in the pillow lava sequences.

Carbonates, essentially calcite, "brown" and "green"clays, and Fe-hydroxides are the most abundant secondaryphases. They invariably occur as pseudomorphs after olivineand glassy/cryptocrystalline mesostasis as well as authigenicphases filling veins and vugs. Although no glass has beenpreserved in the studied samples, the palagonitization (cf.Honnorez, 1978) of the mesostasis has not proceeded to a

mature stage. Prevailing alteration minerals are clays. Well-developed crystals of phillipsite have been identified in few ifany of the samples, with the probable exception of the pillowsof Units B and D, where a fibrous/patchy phase with lowbirefringence and extremely low refringence was found asso-ciated with clays and Fe-hydroxides as a devitrification prod-uct of the glassy mesostasis.

Olivine relicts were found in only one sample of Unit A(Table 4). Titanomagnetite is largely oxidized, probably to Ti-maghemite and locally also to secondary Fe-oxides, particularlyin the most altered "brown" zones of the pillow lavas (i.e., UnitD). Clinopyroxene and Plagioclase are virtually unaltered in allthe studied samples. In one strongly brecciated sample fromUnit Cl (124-770B-20R-4, 72-79 cm), well-crystallized second-ary K-feldspar partially replaces Plagioclase.

Petrographic observation does not reveal any significantvertical variation in the distribution of secondary minerals,except perhaps calcite. This mineral is rare in the two upperpillow lava sequences (Units A and B), whereas it suddenlybecomes abundant as a space-filling phase in the massivebrecciated lavas of Unit C, and persists in variable amountsdown to the bottom of the drilled sequence.

Where it is possible to evaluate the deposition sequence ofauthigenic phases, such as in vugs and vesicles, it is evidentthat calcite is, in most cases, the last mineral to form. Perhapsits rare occurrence in Units A and B depends on the quickerself-sealing of the voids by clays, due to a higher reactivity toalteration of the abundant glassy mesostasis of the pillowlavas.

The extent of alteration appears to be strongly dependenton the crystallinity of the rocks, which probably drasticallyreduces the rate of reaction between basalt and seawater, asstudied experimentally by Seyfried and Bischoff (1979). In thestudied thin sections, the highly oxidized brownish alterationzones are more diffuse in the pillows than in the massive lavasand dolerite sills, independent of their lithostratigraphic posi-tion. Brown and green clays are present throughout the Site770 sequence. The latter are dominant as pseudomorphingphases after olivine and mesostasis, as well as replacing glassy

277

G. SERRI ET AL.

inclusions in the plagiociase phenocrysts, in the massive lavasand dolerite sills.

The observed secondary minerals (calcite, Fe-hydroxides,brown and green clays, dominantly smectites and celadonite)and the absence of albite (occasional presence of K-feldspar),chlorites and sphene indicate that the alteration processestook place at a temperature certainly below 140°-150°C (Sey-fried and Bischoff, 1979; Alt et al., 1986).

Although numerous experimental data on element mobili-zation due to basalt/seawater interaction at low to moderatetemperature have been gathered (Seyfried and Bischoff, 1979;Mottl and Holland, 1978; Menzies and Seyfried, 1979; Sey-fried and Mottl, 1982), and many studies on mid-ocean ridgebasalts altered at low temperature have been performed (Hartet al, 1974; Böhlke et al, 1980; Staudigel et al., 1981; Staudigeland Hart, 1983; Bienvenu et al., 1990) the direction andmagnitude of chemical exchange in the altered rocks studieddoes not follow any simple rule. For instance, importantfactors such as temperature, water/rock ratio, pH, and oxida-tion state may easily change in the same structural site duringthe evolution of halmyrolysis and hydrothermal alterationsystems related to the formation and cooling of the oceaniccrust, and can therefore produce an complex pattern ofelement mobilization.

The careful petrographic analysis of secondary phases,together with the recognition that the alteration magnitude inthe studied rocks is primarily controlled by the degree ofcrystallinity, gives important indications for the interpretationof geochemical data on whole rocks.

The major- and trace-element mobility can be inferred bycomparing basalts belonging to the same flow unit but showingdifferent degrees of petrographically determined alteration.

Sample 770B-19R, 98-99 cm, is highly oxidized (brownzone), and more altered petrographically than 770B-19R2,42-45 cm, sampled only 1.5 m deeper. K2O and P2O5 aresignificantly enriched in the more altered sample of Unit B(0.14%-0.43% and 0.10%-0.15% respectively). The P2O5 en-richment may be due to the incorporation of seawater phos-phate into secondary ferric oxide, replacing olivine in thehighly oxidized zones (cf. Böhlke et al., 1981). In this case,there is also a small but significant (about 10%) increase ofTiO2. Nb varies from 2 to 3 ppm, but this is obviously withinthe analytical uncertainty. V, Cr, Zr, Y, and, surprisingly,also Ni and Sr, appear unaffected. There is also a significantdecrease (1% absolute value) of CaO among major elements.

Another way to evaluate element mobility is to examinebulk composition in relation to depth (Figs. 3 and 4). Becauseof analytical errors SiO2 and Na2O were not plotted (seeanalytical method). As it is widely recognized that at least Zr,Nb, and TiO2 are virtually immobile during strong low-temperature alteration even of glassy rocks (Bienvenu et al.,1990; Hart et al., 1974) the ratio between these elements isused to give an initial characterization of magma types. Ti/Zr(75-110) and Zr/Nb (>20) indicate that all of the studied rocksare consistent with normal MORB (cf. Sun et al., 1979;Saunders et al., 1988; Sun and McDonough, 1989; Le Roex etal., 1989). This result, which is confirmed by the wholetrace-element pattern, permits an easier evaluation of elementdistribution with height. In general, the lithological units alsocorrespond to chemical units.

Unit A (sparsely plagioclase-olivine phyric pillow lavas) iswell defined by the homogeneous and distinctly higher TiO2,P2O5, V, Zr, Nb, Sr contents and Zr/Y ratio than in the otherlithological types. In contrast, the erratic distribution of K2Oand FeO3 tot suggests that these elements were mobilized.

Unit B (highly plagioclase-olivine phyric pillow lavas)evidently represents a different magma type. The distributionof TiO2 (1.18%-1.32%) and P2O5 (0.10%) is relatively homo-

geneous, except for one strongly altered sample with higherP2O5 (0.15%) content (seawater phosphate into secondaryferric oxide?). K2O is highly variable. Low TiO2 and P2O5,high Cr (404-383 ppm), and Ni (150-144 ppm) indicate arelatively primitive nature of the original magma; mg numbers(58.6-59.5) are relatively low for such high Cr and Ni mag-mas. It is hard to say whether or not this is due to spinel andolivine accumulation. At any rate the comparison, as regardsTiO2 and Ni vs. MgO, of the studied rocks with the spectrumof variation of N-MORB (Fig. 5) indicate that the basalts ofUnit B are too low in MgO for the given TiO2 and Ni contents.A possible explanation, also supported by experimental data(Seyfried and Bischoff, 1979) and petrography (Table 1), isthat MgO is leached from the pillow basalts (in this caseolivine and glass) during alteration under oxidizing conditions.Analogous reasoning is likely to apply also to the upper pillowlava sequence (Unit A), where the rocks show an anomalouslyhigh Ni/MgO ratio compared to the maximum variation ofnormal MORBs (Fig. 5). This is also valid when Ni-MgOvariation of bach-arc basin basalts from the Mariana Troughand the Lau Basin are taken into consideration (Fig. 5).

Unit C (moderately to highly plagioclase-olivine phyricmassive lava flows) is chemically distinguishable into twosubunits, Cl and C2. The latter shows higher TiO2, P2O5 andZr contents. Cl is also transitional (for Zr, Y, V, Cr, Sr)between Unit B and C2. TiO2 of Unit Cl and C2 shows arelatively small scatter, whereas K2O and Fe2O3 are highlyvariable. The MgO/TiO2 ratio of Unit C is apparently notanomalous among MORBS (Fig. 5), whereas Ni/MgO ratio isevidently quite high and generally outside of MORB spectrum(Fig. 5). Whether this is due to olivine accumulation or slightMgO removal is hard to say. In the case of the massive rocks,it is not necessary to invoke extensive MgO removal toexplain Mg and Ni distribution. This is in accordance withpetrographic observations, which show a smaller amount ofsecondary minerals, and with the conclusion of Böhlke et al.(1981) on similar massive rocks from DSDP Site 396B in theNorth Atlantic. Ubiquitous metastable clinopyroxene andPlagioclase as well as minor microcrystalline/glassy mesosta-sis indicate that Al and Ca are probably not too far from themagmatic value, particularly in massive rocks. The high CaOand A12O3 content of Units B and C is generally associatedwith the presence of abundant Plagioclase phenocrysts.

Only one or two samples from Units D, E, F, G, and Hwere studied (Table 1 and Figs. 3 and 4). In general the Ti, Mg,Ni variation is within the MORB spectrum. This suggests thatMgO removal was not strong below the massive lava sequenceof Unit C. Higher incidence of "green" over "brown" clay,as well as predominance of crystalline rocks, suggest that thismight be related to a weaker oxidation, possibly due to a lowerwater/rock interaction (cf. Seyfried and Bischoff, 1979). It isto be noted that the dolerite sill of Unit F resulted from theintrusion of a relatively evolved magma (Tables 5 and 6; Figs.3 and 4) approaching the chemical composition of Unit A.These doleritic basalts are the least-altered rocks of theCelebes Basin. Accordingly, concentration of K2O (0.24%),Rb (3.25 ppm) and Cs (0.05 ppm), and even Li (9.5 ppm), arewithin the range of evolved fresh N-MORB glasses (cf.Hertogen et al., 1980; Seyfried et al., 1984). Nevertheless, TIis enriched by about a factor of 10 relative to fresh N-MORBglasses (cf. McGoldrick et al., 1979; Hertogen et al., 1980). Ingeneral, TI, Cs, and Li as well as Rb are the most sensitiveindicator of rock/sea water interaction at low temperature.The fact that only TI appears to be significantly enrichedindicates that the effect of low temperature alteration onwhole-rock chemistry of the studied doleritic basalt (124-770C-10R-2, 43-47 cm) was so weak as to be consideredvirtually absent.

278

mg

I

TiO2/P2O5 TiO2% P2O5%

CELEBES SEA IGNEOUS ROCKS

KOO% Unit

4 4 0 •

460 "

480 -

500 "

520 "

540 •

•

••

Φ

o

o

a aπ α

D

••

Δ

O

o

-

••••t *

** * o

oa a

o oo

•

Δ

o

o

•ββ•

V

-

•

•

•

•

t**oOo

0 D

o oa

mm

Δ

o

o

if

•

••••*o#o

D Gα

••

Δ

o

o

«

•

-

•••••** *o

CO

oDO

D

••

Δ

o

o

V

C 2

D

E

F

G

H

50 60 70 80 10 14 18 1.0 14 1.8 2.2 0 0.1 0.2 0.3 0.1 0.3 0.5

MgO% Fe 2 O 3 % AI2O3% CaO% Unit

4dV

440 "

460 "

480 -

500 "

520 "

540 •

•

••J

*

o

oa a

o aa

mm

Δ

~δ~

0

-β- •V

-

-

4•••^ »*o

α α

° ao

••

Δ

o

o

v

•

-

_

• •••*_•*o

o oo

D O

α αα

Δ

ö~

o

• •• • V

-

-

•

••

•

jr_ _ _ _ _

* *

Ooo

oα α

G Oa

mm

Δ

o

o

• • wm_V

c 2

D

E

F

G

H

5 6 7 8 9 10 7 8 9 10 11 15 16 17 18 19 20 11 12 13 14 15

Figure 3. Stratigraphic plots for some major elements of Celebes Sea basalts, mg = 100 Mg/(Mg+Fe2+). Fe2O3/FeO = 0.15.

279

G. SERRI ET AL.

Nb(ppm) Zr(ppm) Y(ppm) V(ppm) Ni (ppm) Crlppm) Unit

&

•Q.

Φ

440"

460-

480-

500-

520-

540 -

•••

8oD

aa

m

o

•

-

•

•

•

eo

α

Do

••

o

•θ• »

-

-

•

•

•

ooo

aa

m•

o

» «-V

-

•

•

•

CO

oa

a

o

V

•

•

$

o oo

a

aa

mm

o

β• *

2 4 6 8 60 100 140 180 20 30 40 50 250 350 100 200 300 300 400 500

Sr(ppm) Rb(ppm) Ti/Zr Ti/V Zr/Y Zr/Nb U n i t

Q.

Q

440 -

460-

480 -

500 -

520-

540 -

•

4

o

aαD

aa

o

«- «-

••

•

Φ

o $

DD

D

••

O

•

•

•

•

oooD

DD

-*-

••

•

\

ooo

αa

a

mm

o

-

•

•

•

oo

oD

• •

c 2

D

E

F

G

H

100 150 200 2 4 6 8 10 80 120 160 30 40 3 4 20 30 40

Figure 4. Stratigraphic plots for some trace elements and Ti/Zr, Ti/V, Zr/Y, and Zr/Nb ratios of Celebes Sea basalts.

280

C\J

O

ao -

^ 5 -

zo -

1.5 -

1.0 "

ACelebesSite 770

g \ \

• v\\ \ \ \\V \ V—Cayman

Φ \ T \ \ ] * ~ Tamayo

Sea

Φ UNIT

Φ »

O »

D >

• >Δ »O »

+- »v »

\ \ —A A Discordance

^ – — Kolbeinsey

AB

C1

C2

D

EF

G

H

10 12

MgO%

14

βoo -

500 "

E400 -

αQ.

300 -

200 -

100 -I

B

Back-arc Basalts(Mariana Trough,(Lau Basin)

A A Discordance

Tamayo

KolbeinseyCayman

10 12 14

MgO%16

Figure 5. A. TiO2 vs. MgO and B. Ni vs. MgO diagrams of Celebes Sea basalts. Fields for rocks from Mid-Cayman Rise, East Pacific Rise southward of Tamayo transform,Australian-Antarctic discordance and Kolbeinsey Ridge after Klein and Langmuir, 1987. Field of Mariana Trough and Lau back-arc basins after Hawkins and Melchior, 1985.

δ

G. SERRI ET AL.

Element mobilization can also be evaluated by comparingelements that may be enriched by halmyrolis such as Ce, Ba, Th,U, and K with Nb, which is considered geochemically immobileduring low-temperature seafloor alteration (cf. Bienvenu et al.,1990; Hart et al., 1974). For these elements, MORB-OIB sys-tematics have recently been studied by Saunders et al. (1988) andare presented in Figure 6. Th, Ce, and even Ba show coherentbehavior with Nb. This indicates that they were not significantlymodified (enriched) by halmyrolis. In contrast, the wide vari-abilty of the K/Nb ratio at constant Ce/Nb value indicates thatthe K is strongly enriched during the low-temperature alterationsuffered by the studied samples. By contrast, the Th/Nb ratio isconstant, apart from the sample with analytical uncertainties(see analytical methods). This supports the hypothesis of arelative immobility of Th during the alteration of the studiedrocks. U, whose concentration is very close to the detectionlimit of ICP-MS, does not appear, within the analytical uncer-tainties, to be strongly enriched by alteration, except for thehighly oxidized Site 767 basalts.

MINERAL CHEMISTRY

Site 770Petrographically, the rocks of Site 770 are relatively homo-

geneous (Table 1). They are basalts, with Plagioclase and olivineas ubiquitous phenocrysts. Plagioclase and groundmass clinopy-roxene are commonly fresh, whereas olivine is altered to sec-ondary minerals in all of the samples except in Unit A, wheresporadic relics of olivine microphenocrysts (Fo 83.3) are found(Table 2). The olivine shapes indicate equilibrium growth fromthe host magma. The calculated FeO/MgO ratio (D = 0.27) of theliquid is 1.3, much lower than the whole-rock FeO*/MgO ratio(1.7 calculated assuming Fe2O3/FeO = 0.15). As inferred in thealteration chapter, this is mainly due to MgO removal duringhalmyrolysis under highly oxidizing conditions. FeO*/MgO ra-tios varying between 1.6 and 1.8 allow us to extend this inter-pretation to the whole of Unit A.

Based on clinopyroxene analyses, Unit A is clearly distin-guishable from all the other basaltic units (Table 3; Figure 7),except the pillow lavas of Unit B. Ca/Mg+Fe ratios of thesepillow sequences are significantly higher that the rest of Site770 clinopyroxenes, which are, however, crystallized with alower cooling rate either in massive lavas or in dolerite sills.These pyroxenes show a typical subalkaline trend in thepyroxene quadrilateral. Whether or not these differences aredue to the cooling rate it is difficult to say. Although someexperimental observations (Gamble and Taylor, 1980) showthat the partitioning of Ca, Mg, and Fe is essentially rate-independent, other results (Lofgren, 1979) indicate that, inrelation to increasing cooling rate, a slight increase of Ca/Mg+Fe and signifcant increase of the Mg/Fe ratio occursduring quenching of some magma types.

The strong variation of A12O3 and TiO2 in quenched py-roxenes from pillow units (Unit A, A12O3, 4.3%-6.4%; TiO2,2.4%-3.4%. Unit B, A12O3, 2.9%-7.0%; TiO2, 1.3%-3.0%)suggests that the cooling rate was very high, and that thereforesome increase of Ca/Mg+Fe and particularly Fe/Mg mightindeed have taken place.

In conclusion, clinopyroxene composition in terms ofquadrilateral elements as well as TiO2 and A12O3 appears, inthe analyzed rocks, to be more dependent on the kinetics ofcrystallization rather than the chemistry of parental magmas.

In general the plagioclases show a wide compositionalrange both in pillow basalts and dolerites (Table 4; Fig. 8). Allanalyzed units show maximum anorthite content of the pla-gioclases greater than 84%, reaching the highest value of An88.5% in Unit B. Plagioclases from dolerite units (F and G)evidently reach the lowest anorthite content (around 30% An),

while the pillow basalts show groundmass Plagioclase withminimum anorthite content around 50%. In the Plagioclaserims of Units C2 (massive basalts), anorthite contents be-tween 35 and 40 are attained. Note that the phenocrysts ofUnit A (Sample 124-770B-16R-4, 95-96 cm) are slightly moreenriched in orthoclase molecules than all the rest of thestudied samples from Site 770 at the same anorthite content.This suggests a slightly higher K2O content in the parentalmelt in this unit.

MgO and iron calculated as Fe2O3 are found as minorelements (less than 1%) in all the analyzed crystals. FeO (andK2O) are positively correlated with anorthite content as com-monly observed in typical MORB (Ayuso et al., 1976). Inabsolute value, Fe vs. Na cation variation is very similar toPlagioclase from DSDP Leg 64 basalts (cf. Perfit and Fornari,1983). The variation of MgO is also typical of MORB plagio-clase, with the classic Mg drop which is observed in this caseat about 60% An content. Whether this is due to the onset ofclinopyroxene crystallization (Ayuso et al., 1976) or crystal-lochemical effects (Perfit and Fornari, 1983) is not clear.However, the ubiquitous absence of clinopyroxene phenoc-rysts and the MgO drop being confined to groundmass plagio-clases and to the rim of phenocrysts does not allow us toexclude the clinopyroxene effect.

Spinels reveal a rather uniform compositional range (Table2). In Figure 9 they plot well inside the fields for MORB.

Site 767The single sample studied from Site 767 is an altered, fine-

grained olivine- basalt. The rock, which shows a pseudo-spinifextexture, is characterized by chain-like olivine up to 0.5 cm (15%),and has fasciculated, fan-shaped spherulitic Plagioclase (30%)intergrown with skeletal dendritic clinopyroxene (20%). Ground-mass olivine also shows branching texture. Vesicles (2%) arefilled with clays, calcite, and Fe-hydroxides. The same mineralsreplaced the glassy /criptocrystalline mesostasis (30%). Olivine isdominantly replaced by Fe-hydroxides. The low-temperaturealteration evidently took place under highly oxidizing conditionsas indicated by the variable brownish color that gives the rocksa mottled appearance.

GEOCHEMISTRYHigh-quality analyses have been performed on eight se-

lected samples, seven of which come from different strati-graphic levels within the two holes of Site 770. At Site 767 thebasement was drilled only for a few centimeters; therefore,only one sample has been analyzed.

Major-element analyses of 29 samples from Holes 770B,770C, and one from Hole 767 have been reported in Table 5.Trace-element analyses determined in various laboratories onthe same samples are reported in Table 6.

A general overview of the effects of low-temperatureseafloor alteration on the whole geochemistry is given in thepetrography chapter, alteration section.

Abundance and ratios between incompatible elements,which are widely considered to be immobile or weakly af-fected by seawater rock alteration processes (cf. Hart et al.,1974), show that all Celebes basement rocks are within therange of variation of MORBs (Tables 5 and 6). Relatively highcontent of compatible trace elements, such as Ni and Cr,indicate that these rocks represent nearly primitive or weaklyfractionated basaltic magmas despite generally low MgOcontents (5.2%-7.7%; Table 5), largely due to removal duringhalmyrolysis.

Chondrite-normalized REE patterns (Fig. 10A) allow us toclassify the studied samples as normal MORB. It is evidentthat the basalt of the upper part of Unit A, as also shown bythe mineral chemistry, is a distinctly different magma type.

282

1.0 i

0.5

\.Qü-

0.1

0.05

C. CRUST

N-MORB

— P-MANTLE

— E-MORB

0.01 0.05 0.1

U/Nb0.5

50

nZ\CD

o

10 N-MORB

E MORB7 6

CelebesSite 770

Sea

• UNITΦ »

O »

α »

• »A »

O »

-•- »

V »

A

B

C1

C2

D

E

F

G

H

0.05 0.1

Th/Nb0.5

1 0 •,

.Qz\Φ

o

N-MORB

E-MORB-

P-MANTLE C. CRUST v

OIB

10

Ba/Nb50

1 0 η

5

MORB-

100

C. CRUST

500

K/Nb2000

Figure 6. Pb/Nb vs. U/Nb, Ce/Nb vs. Th/Nb, Ce/Nb vs. Ba/Nb, and Ce/Nb vs. K/Nb of Celebes Sea basalts. Field of N-MORB, E-MORB, and ocean island basalts (OIB)are reported for comparison (after Saunders et al., 1988). Averages for bulk continental crust (Weaver and Tarney, 1984) and primordial mantle estimates (Sun, 1980) arealso indicated. The cross identifies the Hole 767C basalt.

G. SERRI ET AL.

Table 2. Representative microprobe analyses (wt% oxide) and atomic proportions of olivine (ol) and Cr-spinel(sp) from Celebes basement rocks. Fe3+/Fe2+ partitioning is according to charge balance.

Hole 770B

Core-SectionInterval (cm)

SiO2

TiO2

A12O3

Fe2O3

FeOMnOMgOCaONa2ONiOCr2O3

Total

SiTiAlF e + 3

F e + 2

MnMgCaNaNiCr

Total

Unit A16R-495-96

olAc

39.610.020.05

-15.830.23

44.150.330.050.240.01

100.52

0.99540.00040.0015

_0.33270.00491.65350.00890.00240.00490.0002

3.0048

Am

39.340.030.07

-15.930.25

44.120.320.000.250.06

100.37

0.99120.00060.0021

_0.33570.00531.65680.00860.00000.00510.0012

3.0066

Ar

39.500.040.07

-15.820.29

44.270.340.020.230.05

100.63

0.99210.00080.0021

_0.33230.00621.65700.00910.00100.00460.0010

3.0062

Unit B18R-3

144-145spAc

0.4635.945.68

12.480.21

16.460.05

_0.20

28.72

100.20

_

0.07909.67840.97662.38460.04065.60450.0122

_0.03685.1873

24.0000

Ar

0.4636.755.32

12.280.19

16.770.02

_0.19

28.46

100.44

_

0.07859.82950.90852.33050.03655.67130.0049

_0.03475.1055

23.9999

UnitC221R-434-35

spAc

0.5531.405.49

14.720.23

14.650.02

0.1933.17

100.42

_

0.09708.68370.96932.88840.04575.12260.0050

_0.03596.1525

24.0001

Hole 770C

Ar

0.6929.956.04

19.310.28

11.670.05

_0.13

32.83

100.95

_

0.12428.45071.08813.86590.05684.16340.0128

_0.02506.2129

23.9998

Uni tD7R-53-6spAc

0.6929.767.62

15.490.22

14.060.13

-0.13

32.62

100.72

_

0.12288.29751.35643.06440.04414.95700.0330

_0.02476.1007

24.0006

Ar

0.6528.47

7.2915.070.23

14.150.14

_0.15

34.54

100.69

_

0.11627.97431.30362.99510.04635.01170.0356

_0.02876.4895

24.0010

Abbreviations: c, crystal core; m, crystal middle; r, crystal rim.

High Zr/Nb ratio, between 25 and 35, rules out the affinity ofthis magma with transitional MORB (Le Roex et al., 1989).The MORB-normalized patterns (Fig. 10B) of Samples 124-770B-21R-6, 18-19 cm and 124-767C-12R, CC, 36-39 cm,show an anomalous enrichment of K20 and Rb, due to thelow-temperature alteration.

A careful examination of the REE and incompatible ele-ment patterns shows increasing La/Sm, La/Yb, Zr/Y, Nb/Y,and Th/Yb with increasing Th, Nb, and Ce. This is compatiblewith different degrees of melting of the same mantle source.

The apparently high Ce and P abundances relative to ahypothetical baseline drawn between Zr and Nb (Fig. 10)cannot be considered indicative of a subduction-related en-richment, for different reasons.

1. The Ce/Nb (Fig. 6) and Ce/Zr ratios are well within therange of MORBs. The variation of the Ce/Zr ratio (0.106-0.143,ICP-MS data; 0.099-0.127, on the basis of XRF Zr data) iswithin the range of normal MORB from southwest Indian Ridge(0.108-0.157, Le Roex et al., 1983) and comparable to N-MORBfrom various oceans (0.102-0.123, Sun et al., 1979; N-MORBaverage of Ce/Zr = 0.115, Hofmann, 1988).

2. As shown by numerous authors (cf. Pearce, 1982;Beccaluva et al., 1984), Th is much more sensitive than Ce andP to subduction-zone enrichment. Evidently the absence of Thenrichment in the Celebes basalts suggests that the very weakP and Ce positive anomaly shown in Figure 10 does notrepresent the signature of a subduction-related component.

A synthesis of the most relevant geochemical features isgiven in a diagram (Fig. 11) taken from Beccaluva et al., 1984,where the relative variation of Th, Nb, and Zr of the mostcommon magmas from subduction- and nonsubduction-re-lated settings are reported. All the samples from the Celebes

basement fall within the array of basic lavas from nonsubduc-tion settings, perfectly overlapping the N-MORB field. This isalso evident in the Ce-, Th-, Ba-, and Pb-normalized to Nbdiagrams (Fig. 6) and in the Zr/Y vs. Zr/Nb plot of Figure 12.

The Sr isotopic composition of eight samples, many ofwhich were analyzed in duplicate, ranges between 0.70324 and0.70350 (Table 6). One sample shows a ratio as high as0.70483, which was reduced to 0.70344 by 6.5 N HC1 coldleaching. 143Nd/144Nd ratios measured in three samples of UnitA, B, and Cl gave nearly homogeneous values in the range0.51291-0.51298. In the Sr-Nd isotopic diagram (Fig. 13), theCelebes basalts plot within the mantle array, but outside thefield of Atlantic and Pacific MORBs. The obtained values aretypical of some back-arc basin basalts, such as those of theLau Basin (Volpe et al., 1988) and of MORBs from the IndianOcean (Mahoney et al., 1989).

CONCLUSIONSThe middle Eocene oceanic basement of the Celebes Sea,

drilled at Sites 767 and 770 during ODP Leg 124, consists ofeight units made of basaltic lava and doleritic sills, all of whichsuffered various degree of low-temperature seafloor alter-ation, at temperatures certainly below 140°-150°C. Secondarymobilization generally involved TI, Cs, Li, Rb, K, and, to alesser extent, Mg, Ca, Na, and, occasionally, P, whereas themost crystalline samples appear to be only slightly affected bychemical changes.

Textural and mineralogical features as well as the crystal-lization sequence (olivine, Plagioclase, clinopyroxene, Fe-Tioxides) of Site 770 basement rocks are, in all, analogous totypical mid-ocean-ridge basalts. All of the studied rocks aregeochemically within the spectrum of normal MORB compo-sitional variation. Relatively high content of compatible traceelements, such as Ni and Cr, indicate that these rocks

284

CELEBES SEA IGNEOUS ROCKS

Table 3. Representative microprobe analyses (wt.% oxide) and atomic proportions of pyroxenes from Celebes basement rocks, Fe3+/Fe2+ partitioningaccording to Papike et al., 1974.

Hole 770B

Core-SectionInterval (cm)

SiO2

TiO2

A12O3

Fe2O3

FeOMnOMgOCaONa2ONiOCr2O3

TotalSiTiAlFe+3

F e + 2

MnMgCaNaNiCr

Total

Unit A

18R-290-93

g

47.462.925.801.03

10.120.23

11.7420.050.520.010.09

99.971.7910.0830.2580.0290.3190.0070.6600.8110.0380.0000.003

4.000

Hole 770C

Core-SectionInterval (cm)

SiO2TiO2

A12O3

Fe2O3

FeOMnOMgOCaONa2ONiOCr2O3

TotalSiTiAlF e + 3

F e + 2

MnMgCaNaNiCr

Total

UnitB

g51.08

1.302.990.777.560.20

15.2919.920.370.040.04

99.561.9000.0360.1310.0220.2350.0060.8480.7940.0270.0010.001

4.001

g

47.932.404.411.529.980.26

11.3221.010.460.010.13

99.431.8250.0690.1980.0440.3180.0080.6430.8570.0340.0000.004

4.000

UnitCl5R-7

70-71

g49.11

1.994.850.758.530.16

13.3919.760.570.040.03

99.181.8460.0560.2150.0210.2680.0050.7500.7960.0420.0010.001

4.000

UnitB

18R-3144-145

g

47.732.967.011.006.910.08

12.7521.550.520.000.25

100.761.7660.0820.3060.0280.2140.0030.7030.8540.0370.0000.007

4.000

g50.41

1.333.661.526.620.21

15.3920.050.330.020.11

99.651.8720.0370.1600.0430.2060.0070.8520.7980.0240.0010.003

4.000

Unit Cl

20R-2101-102mph-c

50.891.184.920.586.420.15

15.1721.010.300.050.29

100.%1.8600.0320.2120.0160.1960.0050.8260.8230.0210.0020.008

4.001

Unit C26R-3

44-45

g49.28

1.644.491.287.210.23

14.6319.570.390.040.42

99.181.8440.0460.1980.0360.2260.0070.8160.7850.0280.0010.012

4.001

mph-r

50.781.294.141.186.990.15

15.0620.330.400.010.28

100.611.8680.0360.1800.0330.2150.0050.8260.8020.0290.0000.008

4.000

g49.36

1.984.651.437.260.18

14.2620.290.420.040.14

100.011.8350.0550.2040.0400.2260.0060.7900.8080.0300.0010.004

4.000

Unit C2

21R-199-100

g

50.801.363.160.689.230.23

15.2118.590.340.020.13

99.751.8930.0380.1390.0190.2880.0070.8450.7420.0250.0010.004

4.000

Uni tF11R-196-98

mph-c49.09

1.755.161.626.830.19

13.8321.130.330.010.24

100.181.8240.0490.2260.0450.2120.0060.7660.8410.0240.0000.007

4.000

21R-434-35

g

49.531.863.800.88

10.230.27

13.6118.990.390.020.11

99.691.8630.0530.1690.0250.3220.0090.7630.7650.0280.0010.003

4.000

mph-r50.71

1.613.591.157.500.22

14.7620.340.430.030.29

100.631.8720.0450.1560.0320.2320.0070.8120.8050.0310.0010.009

4.000

21R-477-78

g

49.601.875.171.216.880.18

14.2820.780.400.000.41

100.781.8280.0520.2250.0340.2120.0060.7840.8200.0290.0000.012

4.000

mph-c48.932.263.381.67

12.480.39

12.7617.790.460.000.00

100.121.8530.0640.1510.0480.3950.0130.7200.7220.0340.0000.000

4.000

g

50.011.184.101.406.640.18

14.8120.130.400.010.28

99.141.8670.0330.1800.0390.2070.0060.8240.8050.0290.0000.008

4.000

mph-r51.25

1.402.511.298.710.26

14.6919.820.430.020.03

100.411.9030.0390.1100.0360.2710.0080.8130.7890.0310.0010.001

4.000

g

49.951.585.470.567.250.19

14.2120.810.370.040.27

100.701.8390.0440.2370.0160.2230.0060.7800.8210.0260.0010.008

4.000

UnitG12R-221-33

mph-c50.35

1.772.751.23

10.270.33

13.7619.330.410.020.04

100.261.8860.0500.1210.0350.3220.0110.7680.7760.0300.0010.001

4.000

Hole 770C

Unit A

2R-358-61

g

46.093.085.283.927.460.15

11.2321.480.700.000.00

99.391.7570.0880.2370.1130.2380.0050.6380.8780.0520.0000.000

4.005

mph-r49.53

1.555.160.656.280.18

14.4220.870.370.010.49

99.511.8410.0430.2260.0180.1950.0060.7990.8310.0270.0000.014

4.000

g

46.593.015.082.968.010.20

11.3721.340.560.000.00

99.121.7780.0860.2290.0850.2560.0070.6470.8730.0410.0000.000

4.000

mph-c49.72

1.503.001.50

11.570.35

13.4118.010.410.050.00

99.521.8840.0430.1340.0430.3670.0110.7570.7310.0300.0020.000

4.001

g

45.663.405.423.267.930.15

10.9921.090.640.000.00

98.541.7560.0980.2460.0940.2550.0050.6300.8690.0480.0000.000

4.000

mph-r50.58

1.233.450.857.520.20

14.8420.280.310.010.20

99.471.8860.0350.1520.0240.2350.0060.8250.8100.0220.0000.006

4.000

UnitB

4R-126-27

g

48.352.845.491.119.250.22

11.8220.570.750.030.01

100.441.8100.0800.2420.0310.2900.0070.6600.8250.0540.0010.000

4.000

mph-r50.59

1.162.990.849.130.23

14.4719.330.350.060.00

99.151.9010.0330.1320.0240.2870.0070.8100.7780.0260.0020.000

4.001

Abbreviations: c, crystal core; m, crystal middle; r, crystal rim; g, crystal in the groundmass; mph-c, mph-r, microphenocryst core and rim, respectively.

285

G. SERRI ET AL.

Celebes Sea

5 0

Mg 50 Fe

Unit AlC770B-18R-2,90-93cm\ 1770 -16R-4,95-96cm

Φ770B-18R-3.144-145cm$ - 4R-1,26-27cm

«1jθ>770B-20R-2,101 -102cmC1|<>77OC- 5R-7,70-71cm

Unit C2

Unit F

Unit G

Unit H

D77OC- 6R-3, 44-45cmD77OB - 21R-1,99- 100cmH770B-21R-4,34-35cm•77OB-21R-4,77 78cm

O770C-11R-1.96-98cm

-β•770C-12R-2.21-33cm

V770C-12R-3.66-67cm

Figure 7. Pyroxene compositions of Celebes basement rocks in the Ca-Mg-Fe diagram (atoms %). Large and small symbolsrepresent core and rim of the same crystal, respectively.

represent nearly primitive or weakly fractionated basalticmagmas, despite the relatively low MgO contents in manysamples, largely due to removal during halmyrolysis.

The Sr and Nd isotope ratios of the Celebes basalts plot onthe mantle array (87Sr/87Sr 0.70324-0.70348 with 143Nd/l44Nd0.51298-0.51291) outside the field of Atlantic and PacificMORBs. Similar Sr and Nd isotope ratios are typical of IndianOcean MORBs and also of some back-arc basalts, such asthose of Lau Basin.

The mantle source of Celebes basement basalts does notshow a detectable influence of a subduction-related component.Therefore, the geochemical and isotopic data so far obtained onthe Celebes basement rocks does not allow a clear discriminationbetween mid-ocean ridge and back-arc settings.

ACKNOWLEDGMENTS

We benefited from the comments of A. D. Saunders and ananonymous reviewer. We also thank E. Saccani for cooper-ation in the analytical work.

REFERENCES

Alt, J. C , Honnorez, J., Laverne, C , and Emmermann, R., 1986.Hydrothermal alteration of a 1 km section through the upperoceanic crust, Deep Sea Drilling Project Hole 504B: mineralogy,chemistry, and evolution of seawater-basalt interactions. /. Geo-phys.Res., 91:10309-10335.

Ayuso, R. A., Bence, A. E., and Taylor, S. R., 1976. Upper Jurassictholeiitic basalts from DSDP Leg 11. /. Geophys. Res., 81:4305-4325.

Beccaluva, L., Di Girolamo, P., and Serri, G., 1984. High-K Calk-alkalic, Shoshonitic and Leucitic Volcanism of Campania (RomanProvince, Southern Italy): Trace Elements Constraints on theGenesis of an Orogenic Volcanism in a Post-collisional ExtensionSetting: Pisa (Cursi C), 1-47.

Bence, A. E., and Albee, A. L., 1968. Empirical correction factors forthe electron microanalysis of silicates and oxides. /. GeoL,76:382-403.

Bienvenu, P., Bougault, H., Joron, J. L., Treuil, M., and Dmitriev,L., 1990. MORB alteration: rare-earth element/non-rare-earthhygromagmaphile element fractionation. Chem. GeoL, 82:1-14.

Böhlke, J. K., Honnorez, J., and Honnorez-Guerstein, B. M., 1980.Alteration of basalts from Site 396B, DSDP: petrographic andmineralogic studies. Contrib. Mineral. Petrol., 73:341-364.

Böhlke, J. K., Honnorez, J., Honnorez-Guerstein, B. M., Muehlen-bachs, K., and Petersen, N., 1981. Heterogeneous alteration of theupper oceanic crust: correlation of rock chemistry, magneticproperties, and 0 isotope ratios with alteration pattern in basaltsfrom Site 396B, DSDP. J. Geophys. Res., 86:7935-7950.

Bryan, W. B., Thompson, G., and Ludden J. N., 1981. Compositionalvariation in normal MORB from 22°-25°N: Mid-Atlantic Ridge andKane fracture zone. J. Geophys. Res., 86:11815-11836.

Crawford, A. J., Beccaluva, L., Serri, G., and Dostal, J., 1986.Petrology, geochemistry and tectonic implications of volcanicsdredged from the intersection of the Yap and Mariana trenches.Earth Planet. Sci. Lett., 80:265-280.

286

CELEBES SEA IGNEOUS ROCKS

DePaolo, D., 1988. Neodymium Isotope Geochemistry: An Introduc-tion: Berlin (Springer-Verlag).

Faure, G., 1986. Principles of Isotope Geology: New York (Wiley).Gamble, R. P., and Taylor, L. A., 1980. Crystal/liquid partitioning in

augite: effects of cooling rate. Earth Planet. Sci. Lett., 47:21-33.Hart, S. R., 1988. Heterogeneous mantle domains: signatures, genesis

and mixing chronologies. Earth Planet. Sci. Lett., 90:273-296.Hart, S. R., Erlank, A. J., and Kable, J. D., 1974. Sea floor basalt

alteration: some chemical and Sr isotopic effects. Contrib. Min-eral. Petrol., 44:219-230.

Hawkins, J. W., and Melchior, J. T., 1985. Petrology of MarianaTrough and Lau Basin basalts. J. Geophys. Res., 90:11431-11468.

Hertogen, J., Janssens, M.-J., and Palme, H., 1980. Trace elements inocean ridge basalt glasses: implications for fractionations duringmantle evolution and petrogenesis. Geochim. Cosmochim. Acta.,44:2125-2143.

Hofmann, A. W., 1988. Chemical differentiation of the Earth: therelationship between mantle, continental crust, and oceanic crust.Earth Planet. Sci. Lett., 90:297-314.

Honnorez, J., 1978. Generation of phillipsites by palagonitazion ofbasaltic glass in sea water and the origin of K-rich deep-seasediments. In Sand, L. B., and Mumpton, F. A. (Eds.), NaturalZeolites: Occurrence, Properties and Use: New York, 245-258.

Jenner, G. A., Longerich, H. P., Jackson, S. E., and Fryer, B. J.,1990. ICP-MS: a powerful tool for high-precision trace-elementanalysis in earth sciences: evidence from analysis of selectedU.S.G.S. reference samples. Chem. Geol., 83:133-148.

Klein, E. M., and Langmuir, C H . , 1987. Global correlations of oceanridge chemistry with axial depth and crustal thickness. / . Geo-phys. Res., 92:8089-8115.

Lee, C. S., and McCabe, R., 1986. The Banda Celebes Sulu Basins:a trapped piece of Cretaceous Eocene oceanic crust? Nature,322:51-54.

Leg 124 Scientific Drilling Party, 1989. Leg 124 researchers drillmarginal basins. Geotimes, 34:15—17.

Le Roex, A. P., Dick, H.J.B., and Fisher, R. L., 1989. Petrology andgeochemistry of MORB from 25°E to 46°E along the SouthwestIndian Ridge: evidence for contrasting styles of mantle enrich-ment. J. Petrol., 30:947-986.

Le Roex, A. P., Dick, H.J.B., Reid, A. M., Frey, F. A., Erlank, A.J., and Hart, S. R., 1985. Petrology and geochemistry of basaltsfrom the American-Antarctic Ridge, Southern Ocean: implicationsfor the westward influence of the Bou vet mantle plume. Contrib.Mineral. Petrol., 90:367-380.

Le Roex, A. P., Erlank, A. J., and Needham, H. D., 1981. Geochem-ical and mineralogical evidence for the occurrence of at least threedistinct magma types in the "Famous" Region. Contrib. Mineral.Petrol, 77:24-37.

Le Roex, A. P., Reid, A. M., Frey, F. A., and Hart, S. R., 1983.Geochemistry, mineralogy and petrogenesis of lavas erupted alongthe southwest Indian Ridge between the Bouvet triple junction and11 degrees E. J. Petrol., 24:267-318.

Lofgren, G. E., 1979. Dynamic Crystallization and DevetrificationStudies 1970-1978: Houston (NASA, Lyndon B. Johnson SpaceCenter).

Mahoney, J. J., Natland, J. H., White, W. M., Poreda, R., Fisher, R.L., and Baxter, A. N., 1989. Isotopic and geochemical provincesof the western Indian Ocean spreading centers. J. Geophys. Res.,94:4033-4052.

McGoldrick, P. J., Reid, R. K., and Scott, B. B., 1979. Thallium: asensitive indicator or rock/seawater interaction and of sulfur satura-tion of silicate melts. Geochim. Cosmochim. Acta, 43:1303-1311.

Menzies, M. A., and Seyfried, W. E., 1979. Basalt-seawater interac-tion: trace element and strontium isotopic variation in experimen-tally altered glassy basalt. Earth Planet. Sci. Lett., 44:463-472.

Mottl, M. J., and Holland, H. D., 1978. Chemical exchange duringhydrothermal alteration of basalt by seawater. I. Experimentalresults for major and minor components of seawater. Geochim.Cosmochim. Acta, 42:1103-1115.

Papike, J. J., Cameron, K. L., and Baldwin, K., 1974. Amphibolesand pyroxenes: characterization of other than quadrilateral com-

ponents and estimates of ferric iron from microprobe data. Geol.Soc. Am. Abstr. Programs, 6:1053-1054.

Pearce, J. A., 1982. Trace element characteristics of lavas from destruc-tive plate boundaries. In Thorpe, R. S. (Ed.), Andesites: OrogenicAndesites and Related Rocks: New York (Wiley), 525-548.

Perfit, M. R., and Fornari, D. J., 1983. Geochemical studies of abyssallavas recovered by DSRV Alvin from eastern Galapagos Rift, IncaTransform, and Ecuador Rift. 2. Phase chemistry and crystalliza-tion history. J. Geophys. Res., 88:10530-10550.

Price, R. C , Johnson, L. E., and Crawford, A. J., 1990. Basalts of theNorth Fiji Basin: the generation of back arc basin magmas bymixing of depleted and enriched mantle sources. Contrib. Mineral.Petrol, 105:106-121.

Rangin, C , 1989. The Sulu Sea, a back arc basin setting within aNeogene collision zone. Tectonophysics, 161:119-141.

Rangin, C , Silver, E. A., von Breymann, M. T., et al., 1990. Proc.ODP, Init. Repts., 124: College Station, TX (Ocean DrillingProgram).

Saunders, A. D., and Tarney, J., 1984. Geochemical characteristics ofbasaltic volcanism within back-arc basins. In Kokelaar, B. P., andHowells, M. F. (Eds.), Marginal Basin Geology. Volcanic andAssociated Sedimentary and Tectonic Processes. London (Black-well), 59-73.

Saunders, A. D., Norry, M. J., and Tarney, J. 1988. Origin of MORBand chemically depleted mantle reservoirs: trace elements con-straints. J. Petrol. (Spec. Lithospheric Iss.), 415-445.

Seyfried, W. E., Janecky, D. R., and Mottl, M. J., 1984. Alteration ofoceanic crust: implications for geochemical cycles of lithium andboron. Geochim Cosmochim. Acta, 48:557-569.

Seyfried, W. E., Jr., and Bischoff, J. L., 1979. Low temperaturebasalt alteration by seawater: an experimental study at 70°C and150°C. Geochim. Cosmochim. Acta, 43:1937-1947.

Seyfried, W. E., Jr., and Mottl, M. J., 1982. Hydrothermal alterationof basalt by seawater under seawater-dominated conditions.Geochim. Cosmochim. Acta, 46:985-1002.

Staudigel, H., and Hart, S. R., 1983. Alteration of basaltic glass:mechanisms and significance of the oceanic crust-seawater bud-get. Geochim. Cosmochim. Acta, 47:337-350.

Staudigel, H., Muehlenbachs, K., Richardson, S. H., and Hart, S. R.,1981. Agents of low temperature ocean crust alteration. Contrib.Mineral. Petrol, 77:150-157.

Sun, S. S., 1980. Lead isotopic study of young volcanic rocks frommid-ocean ridges, ocean islands and island arcs. Philos. Trans. R.Soc. London, A297:409-445.

Sun, S.-S., and McDonough, W. F., 1989. Chemical and isotopicsystematics of oceanic basalts: implications for mantle composi-tion and processes. In Saunders, A. D., and Norry, M. J. (Eds.),Magmatism in the Ocean Basins. Geol. Soc. Spec. Publ. London,42:313-345.

Sun, S.-S., Nesbitt, R. W., and Sharaskin, A. Y., 1979. Geochemicalcharacteristics of mid-ocean ridge basalts. Earth Planet. Sci.Lett., 44:119-138.

Taylor, R. S., and Gorton, M. P., 1977. Geochemical application ofspark source mass spectrography — III. Element sensitivity preci-sion and accuracy. Geochim. Cosmochim. Acta., 41:1375-1380.

Volpe, A. M., Macdougall, J. D., and Hawkins, J. W., 1988. LauBasin basalts (LBB): trace element and Sr-Nd isotopic evidencefor heterogeneity in back arc basin mantle. Earth Planet. Sci.Lett., 90:174-186.

Weaver, B. L., and Tarney, J., 1984. Estimating the composition of thecontinental crust: an empirical approach. Nature, 310:575-577.

Wood, D. A., Tarney, J., Varet, J., Saunders, A. D., Bougault, H.,Joron, J.-L., Treuil, M., and Cann, J. R., 1979. Geochemistry ofbasalts drilled in the North Atlantic by IPOD Leg 49: implicationsfor mantle heterogeneity. Earth Planet. Sci. Lett., 42:77-97.

Date of initial receipt: 4 June 1990Date of acceptance: 17 June 1991Ms 124B-160

287

G. SERRI ET AL.

Table 4. Representative microprobe analyses (wt% oxide) and atomic proportions of plagioclases from Celebes basement rocks.

Hole 770B

Core-SectionInterval (cm)

SiO2

TiO2

A12O3

Fe2O3

MnOMgOCaONa2OK2O

Total

SiTiAlF e + 3

MnMgCaNaK

Total

Unit A16R-495-96

Am

48.920.04

31.330.480.000.21

15.882.380.01

99.25

2.2560.0011.7030.0170.0000.0140.7850.2130.001

4.990

Hole 770B

Core-SectionInterval (cm)

SiO2

TiO2

A12O3

Fe2O3

MnOMgOCaONa2OK2O

Total

SiTiAlF e + 3

MnMgCaNaK

Total

UnitCl

20R-2101-102

Ac

47.770.03

33.340.370.000.22

16.751.910.03

100.42

2.1830.0011.7960.0130.0000.0150.8200.1690.002

4.998

Ar

54.130.11

28.070.890.010.40

12.384.330.11

100.43

2.4440.0041.4940.0300.0000.0270.5990.3790.006

4.983

Ar

50.830.07

31.240.510.000.28

14.893.080.03

100.93

2.2980.0021.6650.0170.0000.0190.7210.2700.002

4.994

g

55.060.28

27.130.800.000.25

10.905.160.20

99.78

2.4940.0101.4490.0270.0000.0170.5290.4530.012

4.991

g

52.720.09

29.500.820.030.18

12.094.720.06

100.21

2.3890.0031.5760.0280.0010.0120.5870.4150.004

5.015

18R-290-93

Ac

47.430.01

33.800.420.000.19

17.011.630.00

100.49

2.1660.0001.8190.0140.0000.0130.8320.1440.000

4.989

g

52.810.08

29.620.890.020.32

12.664.340.07

100.81

2.3810.0031.5740.0300.0010.0220.6120.3790.004

5.006

Ar

55.590.01

27.670.970.030.23

10.475.350.13

100.45

2.4980.0001.4650.0330.0010.0150.5040.4660.008

4.990

Ac

47.120.02

32.850.460.000.21

16.322.100.01

99.09

2.1830.0011.7940.0160.0000.0150.8100.1890.001

5.007

Be

UnitB18R-3

144-145Bm Br Ac

52.93 49.39 52.92 47.200.11 0.06 0.10 0.02

29.21 31.51 29.43 33.030.980.020.29

12.63 14.330.08

0.50 0.69 0.410.00 0.02 0.020.28 0.22 0.245.37 12.81 17.582.70 4.29 1.370.02 0.09 0.01

100.58 99.83 100.57 99.88

2.3920.0041.5560.0330.0010.0200.6120.3790.005

5.001

Ar

49.450.06

31.000.570.000.26

14.493.180.02

99.03

2.2810.0021.6850.0200.0000.0180.7160.2840.001

5.008

2.262 2.391 2.1720.002 0.003 0.0011.701 1.567 1.7920.017 0.024 0.0140.000 0.001 0.0010.019 0.015 0.0170.754 0.620 0.8670.240 0.376 0.1220.001 0.005 0.001

4.997 5.001 4.986

Unit C221R-199-100

Be Br

51.08 57.940.07 0.14

29.88 25.330.69 0.820.02 0.000.27 0.08

13.11 7.803.91 7.010.05 0.17

99.08 99.29 1

2.346 2.6160.002 0.0051.618 1.3480.024 0.0280.001 0.0000.019 0.0050.645 0.3770.348 0.6140.003 0.010

5.006 5.003

Ar

54.190.07

26.760.890.010.35

12.943.920.06

99.19

2.4770.0021.4420.0310.0000.0240.6340.3470.004

4.960

Ac

47.580.03

33.530.400.000.24

17.071.780.00

100.63

2.1710.0011.8040.0140.0000.0160.8350.1580.000

4.998

g

52.470.10

29.130.880.020.32

13.393.850.04

100.20

2.3830.0031.5590.0300.0010.0220.6520.3390.002

4.990

Am

51.800.06

30.760.650.030.24

13.953.450.05

100.99

2.3340.0021.6340.0220.0010.0160.6740.3020.003

4.988

19R-198-99

Ac

47.470.03

33.040.430.010.27

17.101.730.01

100.09

2.1790.0011.7870.0150.0000.0190.8410.1540.001

4.997

21R-477-78

Ar

56.640.13

26.840.870.000.129.395.960.13

100.08

2.5450.0041.4220.0290.0000.0080.4520.5190.008

4.988

Am

49.250.05

32.130.420.020.31

15.882.460.01

100.53

2.2420.0021.7240.0140.0010.0210.7750.2170.001

4.996

Be

52.140.08

30.150.670.000.26

13.443.800.04

100.58

2.3570.0031.6070.0230.0000.0180.6510.3330.002

4.993

g

52.230.14

29.611.190.020.54

13.033.930.04

100.73

2.3610.0051.5780.0410.0010.0360.6310.3440.002

4.999

Br

56.600.15

27.160.870.010.159.845.840.13

100.75

2.5300.0051.4310.0290.0000.0100.4710.5060.007

4.991

g

53.350.12

28.841.060.030.50

12.354.330.04

100.62

2.4070.0041.5340.0360.0010.0340.5970.3790.002

4.994

Cc

54.230.12

28.480.890.000.24

11.684.810.07

100.52

2.4430.0041.5120.0300.0000.0160.5640.4200.004

4.994

Abbreviations: c, crystal core; m, crystal middle; r, crystal rim; g, crystal in the groundmass; mph-c, mph-r, microphenocryst core and rim, respectively.

288

CELEBES SEA IGNEOUS ROCKS

Table 4 (continued).

Core-SectionInterval (cm)

SiO2

TiO2

A12O3

Fe2O3

MnOMgOCaONa2OK2O

Total

SiTiAlF e + 3

MnMgCaNaK

Total

Core-SectionInterval (cm)

SiO2

TiO2

A12O3

Fe2O3

MnOMgOCaONa2OK2O

Total

SiTiA 1 ,F e + 3

MnMgCaNaK

Total

Hole 770C

Unit B4R-1

26-27Ac

49.460.04

31.380.470.010.31

14.922.730.01

99.33

2.2730.0011.7000.0160.0000.0210.7350.2430.001

4.990

Hole 770C

UnitD7R-53-6Be

47.650.01

33.240.390.020.23

17.331.720.01

100.60

2.1770.0001.7900.0130.0010.0160.8480.1520.001

4.998

Ar

53.930.08

28.210.890.000.48

12.394.330.05

100.36

2.4360.0031.5020.0300.0000.0320.6000.3790.003

4.986

Br

55.220.14

27.680.990.000.27

10.575.330.11

100.31

2.4860.0051.4690.0340.0000.0180.5100.4650.006

4.993

Be

47.410.04

32.830.410.000.27

16.271.760.01

99.00

2.1930.0011.7900.0140.0000.0190.8060.1580.001

4.983

g

54.030.10

27.861.110.000.45

12.794.110.06

100.51

2.4400.0031.4830.0380.0000.0300.6190.3600.004

4.977

Bm

46.120.04

33.900.380.000.20

17.251.230.01

99.13

2.1370.0011.8520.0130.0000.0140.8570.1110.001

4.985

Uni tF11R-196-98

Ac

48.290.04

32.850.390.020.28

16.491.940.00

100.30

2.2060.0011.7690.0130.0010.0190.8070.1720.000

4.988

Br

54.310.09

27.590.570.040.38

11.794.350.07

99.19

2.4730.0031.4810.0200.0020.0260.5750.3840.004

4.968

Am

48.400.02

32.710.370.000.22

16.331.980.01

100.04

2.2150.0011.7640.0130.0000.0150.8010.1760.001

4.984

g

52.800.10

29.070.840.010.39

12.634.190.05

100.08

2.3%0.0031.5550.0290.0000.0260.6140.3690.003

4.995

Ar

52.340.07

29.890.700.000.19

13.063.930.07

100.25

2.3720.0021.5960.0240.0000.0130.6340.3450.004

4.991

UnitCl5R-7

70-71Ar

53.690.10

28.630.800.000.19

11.764.880.08

100.13

2.4310.0031.5280.0270.0000.0130.5700.4280.005

5.005

mph-c

54.730.12

27.920.790.010.10

10.685.330.13

99.81

2.4770.0041.4900.0270.0000.0070.5180.4680.008

4.998

UnitC26R-3

44-45g

52.500.06

29.060.750.000.28

13.273.940.06

99.92

2.3890.0021.5590.0260.0000.0190.6470.3480.004

4.992

mph-r

51.430.07

29.680.800.010.11

13.014.100.07

99.28

2.3580.0021.6040.0280.0000.0080.6390.3650.004

5.008

Ac

48.460.04

32.930.390.000.26

16.822.020.01

100.93

2.2030.0011.7640.0130.0000.0180.8190.1780.001

4.997

UnitG12R-221-33

Ac

47.460.05

33.520.410.000.24

16.871.760.00

100.31

2.1710.0021.8080.0140.0000.0160.8270.1560.000

4.994

Ar

52.980.06

28.910.780.010.29

13.244.030.05

100.35

2.4000.0021.5440.0270.0000.0200.6430.3540.003

4.992

Am

50.580.04

31.420.450.000.28

14.992.970.01

100.74

2.2910.0011.6770.0150.0000.0190.7270.2610.001

4.992

Uni tD7R-53-6Ac

49.190.03

31.440.430.020.31

15.762.510.01

99.70

2.2580.0011.7010.0150.0010.0210.7750.2230.001

4.996

Ar

60.610.09

24.330.730.020.086.197.940.21

100.20

2.6970.0031.2760.0240.0010.0050.2950.6850.012

4.998

Ar

52.430.07

29.120.880.000.36

13.483.780.05

100.17

2.3820.0021.5590.0300.0000.0240.6560.3330.003

4.989

Br

53.670.08

28.210.900.000.23

11.824.760.07

99.74

2.4390.0031.5110.0310.0000.0160.5760.4200.004

4.999

Be

51.000.04

30.650.530.000.27

14.303.330.02

100.14

2.3210.0011.6440.0180.0000.0180.6970.2940.001

4.995

289

G. SERRI ET AL.

Celebes Sea

Or

Unit AlC770B-18R-2,90-93cmI 1770 -16R-4,95-96cm

Unit B

Unit Cl

770B-19R-1.98-99cm <s>J770B-18R-3, 144-145cm *J77OC- 4R-1.26-27cm I

0770 B-20R-2,101 -102cm OO770C- 5R-7.7O-71cm O

Unit C2

Q77OC- βR-3, 44-45cmQ770B-21R-1,99-100cmiB770B-21R-4.34-35cm

•77OB-21R-4.77 78cm

Unit D •77OC- 7R-5,3-6cm

Unit F O770C-11 R-1.96-98cm

Unit G •θ770C-12R-2.21 -33cm

Unit H V77OC-12R-3.66-67 cm

Figure 8. Plagioclase compositions of Celebes basement rocks in the Ab-An-Or diagram (mol %). Large and small symbols represent core andrim of the same crystal, respectively.

290

.60

<

ü 4°

ü

.20

.80

£ 60

.40

.20

CELEBES SEA IGNEOUS ROCKS

Celebes SeaS i t e 770

UNIT B { Φ 770-18R-3.144-145cm

UNIT C2 4 B 770-21 R-434-34cm

UNIT D 4 • 770- 7R-5,3-6cm

.10 .30 .50 .70 .10 .70.30 .50

Mg/Mg+Fe2+ Mg/Mg+Fe2+

Figure 9. Cr-spinel composition of Celebes basement rocks in the Cr/(Cr+Al) vs. Mg/(Mg+Fe2+) and Fe 3 + +Al+Cr) vs. Mg/(Mg+Fe2+) diagrams. For

comparison, fields for Cr-spinels from Salomon Island picrites, Grenada basanites, and MORB are also reported (after Crawford et al., 1986).

Table 5. Representative major-element analyses (wt% oxide) of the Celebes basement rocks. Analyses carried out by XRF; mg = 100Mg/(Mg + Fe2 +)

calculated with Fe2θ3/FeO = 0.15. A and C represent Udine University and-aboard ship analyses, respectively.

Hole 767C

Core-SectionInterval (cm)

SiO 2

TiO 2

A12O3

F e 2 O 3

MnOMgOCaONa 2 OK 2 O

P2O5Total

mg

12R,CC36-39

A

49.081.59

16.769.900.315.35

12.832.280.160.28

98.54

52.2

Hole 770C

Core-SectionInterval (cm)

SiO2

TiO 2

A12O3

F e 2 O 3

MnOMgOCaONa 2 OK 2 OP 2 O 5

Total

mg

Unit A2R-3

59-62A

49.902.20

16.109.800.315.37

11.703.150.190.25

98.97

52.6

Hole 770B

Unit A16R-484-87

A

49.752.15

15.6010.950.235.65

11.303.150.290.26

99.33

51.1

3R-24-7

A

49.802.21

16.0510.680.175.30

11.303.050.290.25

99.10

50.1

17R-3134-137

A

49.402.24

16.0510.800.195.20

11.653.250.190.27

99.24

49.3

3R-3124-128

A

49.552.14

16.2010.770.195.55

11.552.950.160.24

99.30

51.0

18R-290-93

C

47.922.23

15.7610.840.215.26

11.442.510.310.24

96.72

49.5

3R-3125-128

A

50.002.24

16.709.650.175.25

11.453.250.210.24

99.16

52.4

Unit B18R-3

141-143C

49.581.32

18.618.210.185.74

12.932.600.180.10

99.45

58.6

UnitB4R-127-30

C

48.691.23

17.448.700.146.32

12.822.760.070.10

98.27

59.5

19R-199-102

C

49.161.28

19.238.600.166.09

11.902.220.430.15

99.22

58.9

UnitCl5R-239-41

C

48.421.25

17.279.230.147.35

12.401.870.330.07

98.33

61.7

19R-242-45

A

48.801.18

18.108.530.156.00

12.922.600.140.10

98.52

58.7

UnitC26R3

45-48C

48.671.53

16.648.660.136.53

12.432.450.380.13

97.55

60.4

UnitCl20R-2

102-105C

49.541.23

18.747.120.127.29

12.232.250.140.08

98.74

67.4

U n i t D7R-53—6

C

49.061.29

17.479.220.157.41

12.412.330.360.10

99.80

61.9

20R-310-15

C

49.471.22

18.137.180.157.42

12.453.160.160.08

99.42

67.6

7R-644-47

A

49.701.62

16.909.920.145.70

12.272.850.360.16

99.62

53.7

20R-4135-140

C

46.931.23

17.638.230.186.83

14.012.310.270.09

97.71

62.7

Unit E9R-2

42-48A

49.001.84

15.8010.600.256.52

12.202.850.270.20

99.53

55.4

UnitC221R-196-99

C

47.451.38

17.129.040.147.67

13.452.130.400.11

98.89

63.2

Unit F10R-243-47

A

50.702.04

15.909.170.176.10

11.783.250.240.21

99.56

57.4

21R-430-34

C

49.581.32

18.618.210.185.74

12.932.600.180.10

99.45

58.6

11R-191-93

C

50.012.10

16.277.940.186.55

11.572.670.190.20

97.68

62.5

21R-474-77

C

48.191.58

16.579.420.147.08

12.271.960.380.15

97.74

60.3

U n i t G12R-221-33

A

48.301.37

15.2010.650.279.05

11.872.650.200.12

99.68

63.2

21R-615-18

C

48.851.49

17.458.540.136.69

13.002.400.390.13

99.07

61.3

12R-238-41

A

48.901.33

15.5210.630.208.67

11.702.450.240.12

99.76

62.3

Hole 770C

Unit A2R-3

58-61A

49.352.19

16.2510.500.285.45

11.733.100.220.24

99.31

51.2

UnitH12R-364-70

A

48.701.24

17.209.700.157.30

12.402.350.320.11

99.47

60.3

291

G. SERRI ET AL.

Table 6. Representative trace-element analyses (ppm element), ^Sr/^Sr and143Nd/144Nd ratios of Celebes basement rocks. Columns A and C referto analyses carried out by XRF and ICP (Y and REE, Nancy) methods at Udine University and aboard ship, respectively; columns B refer toanalyses carried out by the ICP-MS (St. Johns) method.

Hole 767C

Core-Section 12R,CCInterval (cm) 30-33

SeVCrNiTIBeLiCsRbBaSrZrHfThUPbNbYICPYXRF

LaCePrNdSmEuGdTbDyHoErTmYbLu

87Sr/‰1

39330322180

5

11891

233.833

3.44

8.813.711.324.44

5.15

2.96

2.830.38

A

0.70332(1)t>

143Nd/144Nd

Core-SectionInterval (cm)

SeVCrNiTIBeLiCsRbBaSrZrHfThUPbNbYICPYXRF

LaCePrNdSm

21R-615-18

C

244238152

6

14195

331.531

4.51

8.993.63

Hole 770B

B

42.0392

0.050.55

10.60.094.579.26

12381.82.540.150.240.522.45

28.3

3.209.401.628.863.221.284.530.785.231.133.250.473.110.46

Unit A16R-484-87

A

38324220117

5

192164

646.742

8.18

15.75.501.826.05

6.89

3.88

3.700.57

0.70348(1)0.70345(1)0.51291(3)

Hole 770C

B

33.9245

0.020.12

61.00.295.689.61

13783.62.420.180.090.743.01

26.7

3.9011.0

1.789.303.14

Unit A2R-3

58-61A

319207

7

158

7

37

B

39.9341

0.820.75

18.70.143.83

15.7192149

3.900.340.111.015.89

38.0

7.4420.83.20

16.04.891.746.331.086.961.424.130.593.810.58

UnitD7R-53-6A

117

6

13096

434.932

4.6912.8

9.813.89

17R-3134-137

A

41329233116

4

185170

7

43

B

37.5301

0.010.30

37.50.214.34

10.113292.2

2.670.180.070.613.43

28.5

3.8911.81.89

10.13.48

Unit B19R-1

99-102C

281404150

3

11170

3

31

7R-644-47

A

36

30594

9

138112

435.734

5.31

10.74.22

19R-242-45

C

270383144

3

10767

230.829

3.22

7.143.051.123.69

4.44

2.56

2.400.36

0.70340(2)0.70335(2)0.51292(3)

B

39.4317

0.010.32

37.90.528.00

10.7141100

2.880.230.120.753.62

30.3

4.4613.22.11

11.23.60

B

36.7263

0.170.248.830.061.954.55

10561.0

2.010.08

<0.020.341.89

25.1

2.667.711.307.492.721.114.010.704.711.002.910.422.660.40

Unit F10R-243-47

A

42307263110

5

160136

544.438

7.56

13.85.11

Unit Cl20R-2

102-105C

278338154

1

11169

2

26

B

40.5332

0.950.569.520.053.25

17.8163106

2.970.260.090.684.08

31.6

5.4515.22.39

12.33.88

20R-310-15

C

286302205

1

10869

227.928

2.85

6.762.991.083.46

4.11

2.38

2.260.35

0.70324(1)0.51298(4)

UnitG12R-221-33

A

35254375193

6

9678

3

25

B

43.8319

0.010.21

15.50.031.325.5

U064.5

1.750.100.030.431.99

25.3

2.146.851.166.802.520.983.370.624.000.852.460.372.260.34

12R-238-41

A

33250418132

5

6852

330.117

3.88

7.823.39

Unit C221R-196-99

C

263244249

9

12986

3

31

B

36.4284

0.010.32

39.40.152.945.94

10235.7

1.360.040.050.182.49

23.8

3.349.751.598.803.08

21R-474-77

C

260263190

6

13897

3

33

Unit H12R-364-70

A

35233313130

8

9867

3

24

292

CELEBES SEA IGNEOUS ROCKS

Table 6 (continued).

Core-SectionInterval (cm)

EuGdTbDyHoErTmYbLu

87Sr/86Sr ab

21R-615-18C

1.244.11

4.62

2.67

2.550.40

0.70349(2)

Hole 770C

B

1.174.210.754.851.023.030.422.790.42

Unit A2R-3

58-61A

Unit D7R-53-6A

1.344.53

5.06

3.01

2.850.54

0.70483(1)0.70344(2)

B

1.284.610.815.321.133.260.453.000.45

7R-644-47A

1.444.88

5.08

3.01

2.860.46

0.70350(1)

B

1.364.950.865.611.183.440.493.100.47

UnitF10R-243-47A

1.725.74

6.38

3.66

3.510.67

0.70346(1)

B

1.425.140.895.831.243.530.503.230.49

UnitG12R-221-33

A

12R-238-41A

1.144.00

4.38

2.56

2.430.47

0.70343(1)

B

1.154.190.714.640.982.710.392.500.36

Unit H12R-364-70

A

100 T

CD

co

O

üoer

10

10 1 Site767{ T 12RCC,30-33cmi 16R4.84-87cm

Δ19R2,42-45cm+ 21R6,15-18cm* 10R2.43-47cm

Site 770

CD

oI

oo

H 1 1 hLa Ce Pr Nd Sm Eu Gd Tb Dy Ho Er TmYb Lu K Rb Ba ThNbCe P Zr Hf SmTi Y Yb Se Cr

Figure 10. A. Chondrite normalized REE and B. N-MORB normalized trace-element patterns for Celebes basement rocks. Normalizing valuesaccording to Taylor and Gorton (1977), and Pearce (1982). For Ba, the normalizing value (6.30 ppm) is taken from Sun and McDonough (1989).

293

G. SERRI ET AL.

2 0

10

NI

OO

0.5

0.1

Δ SULU

°£}CAGAYAN

A CELEBES

Nb 100/Zr

Figure 11. Th/Zr vs. Nb/Zr diagram for Celebes basement (filledtriangles), Sulu basement rocks (open triangles) and Cagayan Ridgetholeiitic (open squares) and calcalkaline (half-filled squares) basalts.Intraplate/mid-ocean ridge and convergent plate margin basic lavasare clearly discriminated (after Beccaluva et al., 1984). NM and EM,normal and enriched MORB, respectively.

294

CELEBES SEA IGNEOUS ROCKS

7 -

5 -

N

1 -

P- Morb SWIR

N-Morb AAR

N - MorbATL 22°-25°

N-Morb SWIR

T+ P Morb ATL N-Morb ATL

10 30 50

Zr/Nb70 90

Figure 12. Zr/Y vs. Zr/Nb for Celebes Sea basaltic rocks. Field for P-, T-, and N-MORB from Southwest Indian Ridge (SWIR, Le Roex et al.,1983), American Antarctic Ridge (AAR, Le Roex et al., 1985), Atlantic Ridge (ATL, Wood et al., 1979, Le Roex et al., 1981), and N-MORBAtlantic Ridge 22°-25° N (Bryan et al., 1981) are also reported for comparison.

295

G. SERRI ET AL.

-30 -20 + 10 +-20 30

D

CO

0.5134

0.5132

0.5130

0.5128

0.5126

0.5124

0.5122

M id- At lantic Ridge

1 iHawaiian

N. E. Seamounts

East Pacif icRise Small~ Seamounts '

IWalvis Ridge ~" - 'Kerguelen

-4

-6

-8

0.702 0.703 0.704 0.705 0.706

87Sr/86srFigure 13. 87Sr/86Sr vs. 143Nd/144Nd diagram for basaltic rocks from Celebes Basement, Sulu Basement, and Cagayan Ridge. The mantle arrayand some compositional fields of MORB, within oceanic plate and arc volcanism are reported (after De Paolo, 1988; Faure, 1986; Hart, 1988).

296