New He, Nd, Pb, and Sr isotopic constraints on the constitution of the Hawaiian plume: Results from Koolau Volcano, Oahu, Hawaii, USA

New He, Nd, Pb, and Sr isotopic constraints on the constitution of the Hawaiian plume: Results from Koolau Volcano, Oahu, Hawaii, USA

- Geochimica et Cosmochimica Pergamon Acta. Vol. 58. No. 5. DD. 1431-1440. 1994 Copyright 0 1994k~ier Science Ltd Printed in the USA. All rights r...

1MB Sizes 0 Downloads 2 Views

-

Geochimica

et Cosmochimica

Pergamon

Acta. Vol. 58. No. 5. DD. 1431-1440. 1994 Copyright 0 1994k~ier Science Ltd Printed in the USA. All rights reserved

0016-7037194 $6.00 + .OO

New He, Nd, Pb, and Sr isotopic constraints on the constitution of the Hawaiian plume: Results from Koolau Volcano, Oahu, Hawaii, USA M. F. RODEN, ’ T. TRULL,~,* S. R. HART,~ and F. A. FREY 4 ‘Department of Geology, University of Georgia, Athens, GA 30602, USA ‘Department of Chemistry, Woods Hole Oceanographic Institution, Woods Hole, MA 02543, USA ‘Department of Geology and Geophysics, Woods Hole Oceanographic Institution, Woods Hole, MA 02543, USA 4Department

of Earth, Atmospheric,

and Planetary

Science, Massachusetts

Institute

of Technology,

Cambridge,

MA 02 139, USA

(Received February 19, 1993; accepted in revisedfimn November 2 I, 1993)

Abstract-Most analyzed tholeiitic basalts from Koolau Volcano, Oahu, USA, have strontium (“Sr/ *%r = 0.7040-0.7043), neodymium (‘43Nd/‘44Nd = 0.51270-0.51276), and lead (206Pb/204Pb = 17.817.9) isotopic compositions near that of the bulk silicate earth, and 3He/4He isotopic ratios of 1 l-14 times the atmospheric ratio. These helium ratios are higher than MORB, but lower than those of lavas from Loihi seamount. Moreover, the source for the Koolau tholeiites is inferred to have non-bulk earth abundance ratios of highly incompatible elements. Consequently, the source of the Koolau lavas is not primitive, undegassed mantle. The abundance ratios La/Nb, Zr/Nb, and Sr/Nb correlate with *‘Sr/*‘Sr has relatively and ‘43Nd/ ‘44Nd ratios in Hawaiian tholeiites. The enriched (Koolau) source component high La/Nb, Zr/Nb, and Sr/Nb ratios; in fact, Koolau tholeiites have higher Zr/Nb and La/Nb, and lower Th/Nb than most other OIB. These combined trace element and isotopic signatures ofthe enriched component are not consistent with derivation from primitive mantle, recycled crustal material, or a carbonatite metasomatized source. A simple explanation is that the enriched component is residual material, formed recently when a small amount of melt was extracted from primitive mantle, perhaps during the incorporation of the Koolau component into the plume. INTRODUCTION

plume because of their near-primitive Sr, Nd, Pb, and Hf isotopic ratios (STILLE et al., 1983; RODEN et al., 1984a). In this paper we report on the He, Sr, Nd, and Pb isotopic composition of tholeiitic lavas from a well-exposed stratigraphic section of the Koolau shield at Makapuu point, southeast Oahu. Our objectives were ( 1) to determine the temporal variability of isotopic ratios during growth of the Koolau shield, (2) to assess if Koolau shield lavas with near-bulk earth strontium, neodymium, and lead isotopic ratios have high 3He/4He ratios, and (3) to evaluate the correlation between intershield differences in isotopic ratios and abundance ratios of highly incompatible elements. FREY et al. ( 1994) reported on the major and trace element compositions of the same samples.

ISOTOPICDATA FOR HAWAIIAN lavas require three isotopically distinct source components (e.g., STILLE et al., 1986; WEST et al., 1987; TATSUMOTO et al., 1987; KURZ and KAMMER, 199 1). Alkalic lavas forming the rejuvenated stage of Hawaiian volcanoes, e.g., the Honolulu volcanics overlying the Koolau shield, are derived from a mantle component isotopically similar to the source of mid-ocean ridge basalts (MORB, Figs. 1, 2; KURZ et al., 1987; STILLE et al., 1986). Tholeiitic basalts from several shields (Koolau, Lanai, Kahoolawe) have relatively radiogenic strontium, and unradiogenie lead and neodymium isotopic ratios (Figs. 1, 2)) and thus require an “enriched” source component. Finally, the combination of unradiogenic strontium and radiogenic lead of tholeiitic basalts from the currently active shield volcanoes, Mauna Loa and Kilauea, and the active seamount, Loihi, requires a third source component for Hawaiian volcanism (Figs. 1, 2; the FOZO component of HART et al., 1992). Additionally, Hawaiian shield lavas have high 3He/4He ratios compared to MORB; the tholeiitic lavas from Loihi have the highest 3He/4He ratios of any oceanic island basalt (OIB; e.g., KURZ and KAMMER, 1991).

SAMPLING

of Tasmania,

PROCEDURES

Most of the samples (KOO- I-KOO-3 I ) analyzed are from a 2 IO m section at Makapuu Point (see Fig. 2 of FREY et al., 1993). This section contains at least forty-seven lava flows which erupted relatively late in the history of the shield and represent less than 10% of its total volume (M. 0. GARCIA,pers. commun.). Sample numbers indicate relative stratigraphic position; absolute position is indicated in Table I. KOO-53 and KOO-54 are flows collected from Kamehame Ridge and are stratigraphically above the Makapuu Point section (see Fin. 2 of FREY et al.. 1993). KOO-55 is a dike collected from the same ridge. Sample 69 TAN-2 was originally collected by E. D. Jackson from Moleka Stream, northwest of Puukakea, Oahu; this sample was previously analyzed by RODEN et al. ( I984a) and HART ( 1988) and has the lowest ‘43Nd/‘44Nd and 20’Pb/2”4Pb ratios ofany Hawaiian lava. Major and trace element data from this sample were also reported in RODEN et al. ( 1984a). Olivine-rich flows from the Makapuu Point section were collected specifically for helium isotopic analyses: these samples are indicated by KOO-##a in Tables 2 and 3. KOO-8a. KOO-19a, and KOO-30a are distinct flows collected from approximately the same stratigraphic position as KOO-8. KOO-

The Koolau and adjoining Waianae shield volcanoes form the island of Oahu in the Hawaiian archipelago (see Fig. 1 of FREY et al., 1994). Tholeiitic basalts of the Koolau volcano erupted between 2.7 and 1.8 Ma (MCDOUGALL and AZIZ(JR-RAHMAN, 1972; DOELL and DALRYMPLE, 1973); these lavas are crucial to understanding the nature ofthe Hawaiian

* Pre.tent uddress: Antarctic CRC, University Box 252C Hobart, Tasmania 700 I, Australia.

AND ANALYTICAL

GPO

1431

1432

M. F. Roden et al.

y DMM

MAUNA

LOA

*

HIMU

. EMl\ 0.5125

, 0.7030

/ I I 0.7035

I

I

I I 0.7040

I

, 0.7045

0.7050

07Sr/06Sr

lavas have probably lost alkali elements during alteration. Evidence for this conclusion is the excellent correlation between Rb/Sr and Kz0/Pz05 (Fig. 4) and the variable Ba/ Rb ( 14-94) and Rb/Cs ( 13- 10 1) ratios compared to the near-constant value for these ratios in fresh oceanic basalts (HOFMANN and WHITE, 1983; NEWSOM et al., 1986). Seven samples with K20/P205 > 1.3 have Ba/Rb ratios ( 13.916.5 ) only slightly higher than the average value for oceanic basalts ( 11.5 + 1.7, NEWSOM et al., 1986) and similar to the average Ba/Rb ratio, 14.1, of fresh Kilauean lavas ( HOFMANN et al., 1984); Koolau samples with lower KzO/P205 ratios have higher Ba/Rb ratios. Based on these criteria these seven samples are the least altered Koolau samples analyzed by isotope dilution, and they can be used to estimate the magmatic values for K/Rb, Rb/Sr, Rb/Cs, Ba/Rb, and K/Ba; mean ratios are K/Rb = 594, Rb/Sr = 0.015, Rb/Cs = 88, Ba/Rb = 15.2, and K/Ba = 39. These values are similar to those for fresh tholeiitic lavas from Kilauea (K/Rb = 501,

FIG. 1. Sr-Nd isotope correlation diagram for Koolau, Waianae, Kahoolawe, and Lanai tholeiites and the rejuvenated stage Honolulu Volcanics of the Koolau shield. Fields for tholeiites from Loihi, Kilauea and Mauna Loa are also shown (data for this figure and the following figure are taken from O’NIONS et al., 1977; TASUMOTO, 1978; STILLEet al., 1983; HOFMANNet al., 1984; RODENet al., 1984a; STAUDIGELet al., 1984; STILLEet al., 1986; WESTet al., 1987; HART, 1988; KURZ and KAMMER, 1991; LEEMANet al., 1993). Locations of various mantle components are taken from HART et al. ( 1992).

19, and KOO-30. KOO-17a is from an olivine-rich horizon of the same flow as KOO- 17; the major and trace element compositions of KOO-17a are also reported in FREY et al. ( 1994). Analytical methods followed established procedures: HART and BROOKS( 1977) for isotope dilution analyses of K, Rb, Ba, Cs. and Sr (Table 1) and for strontium isotopic measurements (Table 2), DOSS0 et al. ( 199 I ) for neodymium isotopic measurements (Table 2)) TARASand HART ( 1987) for lead isotopic measurements (Table 2), and TRLJLLet al. ( 1990a) for helium isotopic measurements (Table 3). Further analytical details are contained in footnotes to the tables. RESULTS AND PRIMARY MAGMATIC CHARACTERISTICS Alkali element abundances are quite variable in the Makapuu section lavas, for example, potassium ranges by a factor of 3 ( 1400-4500 ppm), rubidium by a factor of 10 (0.787.47 ppm), and cesium by a factor of 5 (0.024-o. 124 ppm, Table 1). These large ranges contrast with more limited ranges for other highly incompatible elements; e.g., niobium ranges by a factor of 1.5 (FREY et al., 1994). Almost certainly, much of the range in alkali element abundances is due to posteruptive element mobility as discussed by FREY et al. ( 1994). In contrast, the alkaline earth elements strontium and barium have more limited ranges (factors of 1.7 and 2.2, respectively) similar to that of niobium. FREY et al. ( 1994) showed that abundances of potassium, silicon, total iron, and in rare examples even the REE and Y in Koolau lavas were affected by postmagmatic processes. One criterion for recognizing altered Hawaiian tholeiites is a relatively low K20/P205 ratio. Tholeiitic glasses from Hawaiian shields have Kz0/Pz05 in the range of 1.3-2.6 ( GARCIA et al., 1989), but this ratio ranges from 0.7 to 1.7 in the lavas studied here (Table 1; Fig. 3); hence, many of these

MAUi’VA LOA

i

0.7050

I

-

/

EM 1

I

l

. KI:OLA, <>HLS * KAHOCLA’hE

MAUNA

i I

T

EM:

LOA

FIG. 2. Pb-Sr and Pb-Pb isotope correlation diagrams for tholeiites from Koolau, Waianae, Kahoolawe, and Lanai volcanoes as well as rejuvenated stage Honolulu volcanics from the Koolau shield. Fields for tholeiites from Loihi, Kilauea, and Mauna Loa volcanoes are also shown. Data sources as in Figure 1.

Nature

of the magma

Rb/Sr = 0.024, Rb/Cs = 95, BafRb = 14.1, and K/Ba = 36, HOFMANN et al., 1984). The lower K/Rb and higher Rb/Sr of Kilauea lavas is consistent with the relative enrichment of incompatible elements that is characteristic of Kilauea lavas (e.g., FREY and RHODES, 1993). We infer that the Koolau source had K/Rb -600 and that the Rb/Sr of relatively fresh Koolau lavas (mean Rb/Sr = 0.015, maximum = 0.0 17) is an upper limit to the source ratio. Therefore, the Koolau source had a Rb/ Sr ratio less than that (0.0 19) inferred for the Kilauea source by HOFMANN et al. ( 1984). Most Koolau tholeiites (Table 2; STILLE et al., 1983; RoDEN et al., 1984a; HART, 1988) exhibit a restricted range in strontium, neodymium, and lead isotopic composition: “‘Sr/ ‘43Nd/‘44Nd = 0.51270-0.51276, ‘%r = 0.7040-0.7043,

0.02

KOOLAU THOLEIITES FROM MAKAPUU SECTION 8 0 a

l

l

k

Y

0.01

$?

0.00

207Pb/204Pb = 15.43-15.46, and 206Pb/204Pb = 17.82-17.93. Because of the young ages of the lavas, these ratios are indistinguishable from initial ratios. Previously RODEN et al. ( 1984a) reported three samples with isotopic compositions outside these ranges (the samples were not from Makapuu Point), and LANPHERE and DALRYMPLE (1980) reported

II.5

FIG. 3. Plot of Rb/Sr vs. K20/P205 ratios for Koolau (Table

1433

source, Koolau volcano

tholeiites

1).

several samples (two from the Makapuu section) with 87Sr/ %r lower than 0.7040. We reanalyzed one of the latter sam-

Table 1. Alkali and Alkaline Earth Element Abundances (ppm) of Koolau Tholeiitic Basalts From The Makapuu Point Section, Oahu’

Sample KOO-3 1 KOO-30 KOO-26 KOO-20 KOO- 18 KOO-17 KOO- 16 KOO-15 KOO-10 KOO-9 KOO-8 KOO-7 KOO-5 KOO- 1

height (m)’ 210 192 155 103 76 73 71 61 34 31 23 20 14 1

U.S. Geol. Surv. Standard Rock This work Rautenschlein et al. (1985) Feigenson et al. (1983)

K 2120 1971 1415 4041 4021 3570 1480 2980 2088 1516 3907 4221 4413 3515

Rb

CS

Sr

3.31 2.95 1.87 6.99 7.46 6.26 0.78 1.45 2.98 2.11 6.95 6.32 7.47 5.48

0.041 0.035 0.024 0.071 0.079 0.062

462 454 354 465 443 401 383 389 352 290 480 426 449 430

78.0 76.0 64.7 113 108 93.2 73.3 110 65 5 52.9 115 87.7 107 89.6

398 396 404

134 131 132

0.114 0.039 0.028 0.078 0.072 0.124 0.067

Ba

K,O&‘,O,’ 0.85 0.88 0.71 1.35 1.73 1.53 0.69 1 28 1.14 0.87 1.52 1.64 1.74 1.41

BHVO-1 4168 4303 4330

9.22 9.25 10.1

0.104

Notes *Sample powders were prepared using agate grinding vessels by F.A. Frey. Separate aliquots of the same sample split were analyzed for major and trace elements (Frey et al., 1994), and for K, Rb, Cs, Sr and Baby isotope dilution (this table). Precisions for K, Ba and Sr arc less than + l%, and approximately & 1% for Rb (Roden et al., 1984b; Taras and Hart, 1987); Precision for Cs has not been rigorously evaluated although the ratio “‘CS/“~CS was generally measured to better than + 1%; by comparison to the internal precisions for the other elements, we conservatively estimate the external precision for Cs to be approximately + 5%. Procedural blanks for the isotope dilution analyses were negligible for all elements except for Cs where the maximum blank contribution to the measured concentration was approximately 1.4% (Blanks: K, 9 ngm; Rb < 0.09 ngm; Cs, 20 pgm; Sr, 0.3 ngm; Ba, < 2.9 ngm, this work and Roden et al., 1984b). ‘Height (m) above sea level @LO. Garcia, written communication). *P205 from Frey et al. (1994). the two samples, KOO-53, KOO-54, from Kamehame of 0.73 and 0.44 respectively.

Ridge, have K20/P205 ratios

1434

M. F. Roden et al. Table 2. Sr, Nd and Pb Isotopic Composition Vicinity, Oahu.’

of Tholeiitic Basalts From the Koolau Shield at Makapuu Point and

Samples From The Vicinity of Makapun Point and Stratigraphically KOO-54

0.704 16’ 0.70416’

KOQ-53

0.70410’ 0.70416’

above Lavas of Makapuu Point

Makapuu Point Lavas ww 10403’ 5X767-ld KOO-3 1 KOO-30

0.70423b 0.70404’ 0.70425 0.70439 0.70440 0.70439’ 0.70430” 0.70412 0.70417 0.70411’ 0.70432’ 0.704 13

KOO-26 KOO-20 KOO-17A KOO- 16

0.70417 0.70419 0.70417” 0.70447 0.70459 0.70408 0.70410 0.70411 0.70411*

KOO-10 KOO-9 KOO-8

KOO-7 KOO-1

[email protected]

0.51261 0.51258 0.51259

17.860

15.435’

37.93 1”

0.51271

17.821”

15.433’

37.7658

17.879”

15.441’

37.826”

17.852”

15.450”

37.792”

17.751” [email protected]

15.392’ 15.406*

37.792’ 31.735s

0,51275’ 0.5 1272’ 0.51271

0.51271

0.51274 0.51274

Sample 69 TAN-2 .7045 1a .70455b

.51254’ .51267b

‘Sr and Pb isotopic compositions (with one exception) were measured on a single collector, 9-inch radius, mass spectrometer at MIT. Sr isotopic compositions were measured on the same sample dissolution as elemental concentrations reported in table I or in the case of leached samples, new unspiked sample dissolutions. The leaching consisted of 1 hour in hot (approximately lOO”C) 6.2 N HCI. All 8’Sr/s6Sr ratios were normalized to ?#?ilr = 0.1194 and are reported relative to Eimer and Amend “Sr/% = 0.70800. Internal and external pm&ions based on run statistics and repeated analyses of the standard are consistent and indicate a precision of measurement of + 0.005%. Pb isotopic ratios were measured on distinct sample dissolutions of leached powders following Taras and Hart (1987). The reproducibili~ of the measured ratios is 0.05% a.m.u. -‘., the isotopic ratios were corrected for mass fractionation based on results for NBS 981. Nd isotopic ratios were measured on Finnegan Mat 261 Mass spectrometer equipped with 5 collectors at IFREMER (Brest) following sample dissolution of unspiked powders at University of Clermont-Ferrand II. All i?ld/“‘Nd ratios are normalized to ‘?Nd/‘*bNd = 0.7219 and are reported as analyzed; La Jolla Nd was yielding a mean “sNd/‘*“Nd of 0.511840 at the time of the analyses. Run statistics as well as repeated analyses of La Jolia Nd indicates a precision oft 0.002% for the measured ratios. Procedural blanks were 0.3 ngm Sr, 200 ngm Pb and 0.15 ngm Nd (F. Vidal, written commnnication); thus sample/blank ratios for a nominal 60 mgm (Sr and Nd) or 500 mgm (Pb) sample were 80000 (Sr), 8000 (Nd) and 2500 (Pb, assuming 1 ppm Pb concentration, Tatsumoto, 1978). No blank corrections were applied to the analyses. a.

Leached

b.

Roden et al. (1984a)

c.

Collected by Wentworth & Winehell (1947) from summit of Makapuu Point

d.

“Top flow of Makapuu Beach section.“, Lanphere & Dahymple

e.

Analyst: L. G&en

f.

Analyst: L. Doss0

g.

Hart (1988)

(1980)

1435

Nature of the magma source, Koolau volcano Table 3. Helium Isotopic Compositions of Koolau basal&.

‘HerHe normalized to air

[HeI Sample

ccSTP/g

KOO 55dike

4.6 x lo-” 3.4 x 1O”O

15.8 11.5

+ 2.8 k 2.3

2.0 x 2.3 x 2.3 x 1.3 x 1.3 x 2.1 x

lo-9 lO-9 lO-9 1o-9 1o.9 1o.9

9.3 2.9 2.4 3.3

lo-9 lo+’ lo-* lo-9

6.51 10.56 10.92 12.73 15.0 14.23 14.16 14.29 10.75 14.25 10.61

+ + + 2 + + + + + + +

KOO 49 melt powder KOO 49 KOO 30a

KOO 30 KOO 19a KOO 17a KOO 8a

x x x x

0.70 0.24 0.58 0.35 1.5 0.79 0.18 0.29 0.25 0.12 0.16

KOO-8A, 19A and 30A are distinct flows from approximately the same stratigraphic positron as KOO-8, 19 and 30. KOO-17A is from an olivine- rich portion of KOO-17 flow. All analyses on olivine separates by crushing, except as noted. Helium blanks for the crusher and the furnace varied in the range of 4-5 x 10.” ccSTP He; results were corrected for blank contribution by assuming atmospheric isotopic composition for the blank. The results are given in the R/R, notation where WR, = (‘He/‘He),,,J(‘He/“He), and the atmospheric ratio is 1.39x 10d. Errors given for ‘HerHe ratios are 1 sigma instrumental uncertainties, and depend primarily on sample size.

ples (5X767-1, Table 2) and found that its 87Sr/86Sr (0.70404) is similar to other Koolau tholeiites and marginally within error of the previously reported value (0.7037 + 3).Also, one sample (KOO-8) with an apparently high 87Sr/86Sr ratio (ca. 0.70459; Table 2) had a lower *‘Sr/*‘Sr ratio (0.70408) following 1 h of leaching in 6.2N HCl. In contrast, five other

samples (including 69 TAN-2) exhibit no significant differences in measured *‘Sr/*%Srratios between leached and unleached aliquots. The case of 69 TAN-2 is particularly interesting, the reanalysis of a leached aliquot confirms the relatively radiogenic strontium and unradiogenic lead and neodymium of this sample. Additionally, the results for the leached aliquots of KOO- 17A and KOO-30 confirm the existence of lavas with relatively radiogenic strotium within the Makapuu section. KOO-30 is also characterized by a relatively low ‘43Nd/ ‘44Nd ratio, compared to the other tholeiites (Fig. 1). Leaching of 69 TAN-2 resulted in a lower ‘43Nd/ ‘44Nd ratio compared to the unleached powder (Table 2); as discussed by FREY et al. ( 1994) this sample contains a REErich phosphate in the groundmass and has elevated REE contents reflecting secondary enrichment in these elements. Thus the neodymium isotopic composition of this sample may have been increased by a secondary process. Although most analyzed Koolau samples have neodymium and strontium isotopic compositions within narrow ranges of 0.5127 l-O.5 1276 and 0.70404-0.70425 (Table 2, STILLE et al., 1983; RODEN et al., 1984a), when all Koolau samples

are considered, they define an inverse 87Sr/86Sr-‘43Nd/‘44Nd trend with samples 69 TAN-2 and KOO-30 defining the high *‘Sr/*%r extreme for Hawaiian shield lavas (Fig. 1). In general, the Koolau lavas have some of the least radiogenic Pb

20 3He/ 4He

I

Kohala

.7035

RCunion

.7045

.7055

85r/86sr FIG. 4. He-Sr isotope diagram comparing the Koolau results (Tables 2 and 4) with other Hawaiian volcanoes (grey fields; KURZ and KAMMER,1991; KLJRZet al., 1987; CHEN and FREY, 1985; STAUDIGEL et al., 1984; KURZ et al., 1983). The Kohala and Mauna Kea rectangles outline ranges obtained in separate helium and strontium studies (FEIGENSONet al., 1983; GRAHAMet al., 1990a; FREYet al., 199 I ). Fields for basalts from Reunion Island (GRAHAM et al., 1990b), Samoa ( FARLEYet al., 1992) and MORB (e.g., KURZ et al., 1982) are also shown.

143h

M. F. Roden et al.

of any Hawaiian lavas and plot close to the geochron (Fig. 2). As for strontium and neodymium isotopes, there is isotopic heterogeneity although most samples have 206Pb/204Pb ratios of approximately 17.8-17.9 and 207Pb/204Pb ratios of 15.43-15.45. Helium isotopic ratios for the crushed olivine separates range from 10.6 to 14.3 times the atmospheric value for 3He / 4He (Table 3; TRULL et al., 1990b), and hence are distinctly higher than the helium isotopic composition of MORB. Crushing releases only helium contained within fluid inclusions, and therefore provides reliable estimates of magmatic helium isotopic compositions (e.g., KURZ, 1986; TRULL et al., 1990a). The lower value, 6.5 R,, for the melted olivine powder, KOO-49, compared to the values 10.6 and 10.9, measured for crushed olivine separates from the same sample, probably reflects addition of radiogenic 4He to the bulk samplesinceeruption. Theconstantheliumcompositionobtained on replicate crushing of olivine from this sample argues against radiogenic helium contamination of the gas evolved from the olivine. as does the relatively small amount of radiogenic helium present in the melted powder (8.1 X 1Om’o ccSTP/g) in comparison to helium released on crushing (2.3 X 1O-9 ccSTP/g; see TRULL et al., 1990a for a discussion). Samples KOO-17A and KOO-30 are distinctive compared to other Koolau lavas because they have relatively high strontium and helium isotopic ratios (Tables 2 and 3). The 3He/4He ratios of Koolau tholeiites are higher than those of MORB but less than those of Loihi lavas (Fig. 4 ). Tholeiitic lavas from Kilauea, Haleakala, and Koolau have similar 3He/4He. The He-Sr isotopic ratio of lavas from Koolau and Reunion are similar, and together with lavas from Samoa they define a trend to high 87Sr/86Sr at moderate 3He/4He (Fig. 4). Although there is evidence for isotopic heterogeneity within the Koolau shield, there is no systematic correlation between stratigraphic height and isotopic composition for the Makapuu section. The lowermost and the uppermost flows have indistinguishable 87Sr/86Sr ratios and are similar to most of the lavas from Makapuu Point. Interspersed with these normal lavas are three flows with more radiogenic strontium (KOO-17A, KOO-30, and KOO-31; KOO-3 1 was not leached). Similarly, the helium isotopic data show that lavas with relatively high 3He/4He ratios are interspersed with lavas with lower 3He/4He ratios; the lowermost and uppermost samples analyzed for helium isotopic composition have relatively low and indistinguishable 3He/4He ratios. The lead isotopic data show limited and random variation compared to stratigraphic height (Table 2). These results contrast with the temporal isotopic variations observed at Mauna Loa (KURZ and KAMMER, 1991). DISCUSSION Does Primitive Mantle Dominate the Koolau Source? Because the Koolau tholeiites have heavy isotope ratios similar to those of the bulk silicate earth, it is possible that primitive, undifferentiated mantle was a source component (e.g., STILLE et al., 1986). FARLEY et al. (1992) found HeSr isotopic systematics in Samoan lavas similar to those of Koolau lavas (Fig. 4) and inferred that a primitive reservoir with a 87Sr/86Sr ratio of approximately 0.705 contributed to

Samoan volcanism. Thus the high 3He/4He ratios in some Hawaiian lavas compared to MORB (e.g., KURZ et al., 1983; KURZ and KAMMER, 199 1) may reflect undegassed, primitive mantle. However, there are other possible explanations for the origins of high 3He/4He ratios. Uncertainty over the relative solid/liquid partitioning of uranium, thorium, and helium during melting led GRAHAM et al. ( 1990a) to suggest that high 3He/4He ratios may not require a primitive source. For example, some differentiated reservoirs may be characterized by high 3He/4He ratios compared to MORB if under certain conditions helium is more compatible than uranium or thorium. Such a differentiated reservoir, however, would have very low helium abundances and nonprimitive strontium. neodymium, and lead isotopic ratios. Additionally, recent experiments indicate that helium is highly incompatible in basaltic systems; mineral/melt partition coefficients are very small ( <0.008), and thus similar to those for uranium and thorium ( LATOURETTE and BURNETT, 1992; MARTY and LUSSIEZ, 1993; KURZ, 1993). Thus, the association of near-bulk earth lead, strontium, and neodymium isotopic ratios with high 3He/4He ratios would be strong evidence for a primitive source. More precisely, if primitive, undegassed mantle is the cause of the near-bulk earth lead, strontium, and neodymium isotopic ratios of the Koolau tholeiites and also the source of high 3He/4He ratios typical of some Hawaiian tholeiites, then the Koolau lavas should have very high 3He/4He ratios. The 3He/ 4He ratios for the Koolau tholeiites are higher than those of MORB but much lower than those for Loihi (Fig. 4). Thus the highest 3He/4He ratios at Hawaii are associated with a clearly non-bulk earth “Sr/“Sr ratio of approximately 0.7035. Moreover, the Koolau sample, 69 TAN-2, which is most similar to the bulk earth in strontium and lead isotopes, has a lower ‘43Nd/ ‘44Nd ratio than chondrites. Another relevant feature of the Koolau tholeiites is the relatively low Rb/ Sr ratio inferred for their source ( ~0.0 17). This value is lower by a factor of 2 than the value (approximately 0.03) required for a primitive reservoir, and thus the *‘Sr/“Sr ratios of the Koolau tholeiites are unsupported even for a model age of 4.6 Ga. Consequently, the source has had a multistage history which included a significant period of time in a high Rb/Sr environment followed by a more recent period characterized by the current Rb/Sr ratio. The obvious conclusion is that the component with nearbulk earth heavy isotope ratios is not undegassed primitive mantle. HOFMANN ( 1986) and FREY et al. ( 1994) presented other trace element data for Hawaiian lavas that support this contention: namely, incompatible element ratios that are not easily modified by the melting process, e.g.. Nb/Th and La/ Th ratios, are not chondritic in Koolau lavas as would be expected if the source was primitive mantle. On the Nature of the Koolau or “Enriched” Source The origin of the enriched Hawaiian component which dominates the Koolau lavas is an intriguing question. HART ( 1988 ) argued that the isotopic composition of Koolau lavas reflects derivation from the enriched mantle component EM 1 (see Figs. 1,2). Originally, the EM mantle components were defined solely on the basis of isotopic composition ( ZINDLER

1437

Nature of the magma source, Koolau volcano Table

4.

Incompatible

Trace

Element

Ba/Nb

La/Nb'

Zr/Nb

Abundance

Ratios

of Koolau

B&k'

Ba/Th

Tholeiites

Th/Nb

Th/La'

Ba/La'

53

77

53

76

32

50

51

49

Meall

16.1

1.3

10.1

11

205

0.052

0.04

Range

13.9-18.6

1.1-1.6

7.4-15.5

0.027-0.095

0.018-0.062

3.1-14.2

n

Ranges

of mean values

from Weaver

7.8-15.5

108-394

8.2

3.2- 5.0

0.66-0.77

4.9-6.5

49-77

0.08-0.10

0.11-0.13

6.8-8.7

EM1 OIB

4.2-11.5

0.86-1.19

11.4-17.8

103-154

0.11-0.12

0.11-0.13

13.2-16.9

EM2 OIB

4.5-7.3

0.89-1.09

7.3-11.0

67-84

0.11-0.16

0.12-0.16

8.3-11.3

Mean values

from Hofmann

(= DMM) 29.7

Prim. Mantle

15.7

26 9.1 6.3-14.2

(1991)

HIMU OIB

N-MORB

Ba/La'

(1988) 1.11

3.95

74.1

0.053

0.048

3.56

0.99

9.80

74.4

0.132

0.132

9.85

Notes : n

=

number

of samples

used to calculate

KoolaudatafromFreyetal. 'Does not include 'Ba/Nb and Ba/La

(1994) andRodeneta1.

two samples ratios

average

showing

for samples

evidence

(1984a) for secondary

REE mobility.

with K20/P205 > 1.5

and HART, 1986 ) . WEAVER ( 199 1) recognized trace element ratio differences between HIMU, EM 1, and EM2 OIB. The high La/Nb ratio typical of the Koolau tholeiites compared to other OIBs was initially recognized by SUN and MCDONOUCH ( 1989) and new data confirm that this ratio is higher in Koolau tholeiites than for most other OIBs (Table 4). In addition to higher La/Nb, the Koolau lavas differ from the EM 1 and EM2 OIB tabulated by WEAVER ( 199 1) in having relatively low thorium and high zirconium contents compared to other incompatible elements (i.e., lower Th/Nb, Th/La, and higher Zr/Nb, Table 4). In fact, Koolau lavas have higher Zr/Nb, and lower Th/Nb and Th/La ratios than any of the OIB tabulated by WEAVER ( 199 1). The low Th/Nb and Th/ La ratios stem from the relatively low thorium contents of Koolau and other Hawaiian tholeiites; the high Zr/Nb and La/Nb ratios result from the relative depletion of elements more incompatible than the LREE in the Koolau tholeiites (see Fig. 12 of FREY et al., 1994). FREY and RHODES (1993) noted that Sr/Nb and Zr/Nb were effective intershield discriminants for tholeiitic lavas in the Hawaiian archipelago. Moreover, because Zr/Nb and Sr/Nb increase in the order Kilauea-Mauna Loa-Koolau, there is a correlation between these elemental ratios and *‘Sr/ “Sr ratios. The more comprehensive dataset in Fig. 5 shows that there is a correlation between Zr/Nb, Sr/Nb, and La/ Nb, and strontium and neodymium isotopic ratios in Hawaiian lavas. Lavas dominated by the enriched source component (i.e., Koolau, Lanai, Kahoolawe) are characterized by relatively high Zr/ Nb, Sr / Nb, and La/Nb. An exception to this generalization are some Lanai lavas with low La/Nb; however, highly variable La contents (ranges by a factor of 3.8) and depletion of the REE compared to other relatively

immobile elements suggest the possibility that variable La/ Nb ratios in Lanai lavas is a secondary effect (REE mobility in Hawaiian lavas was discussed by FREY et al., 1994, and FODOR et al., 1989). The correlation between isotopic composition and these trace element ratios is strong evidence for varying trace element ratios in Hawaiian source components. However, during partial melting, abundance ratios between a highly incompatible trace element and a moderately incompatible trace element, such as Zr/Nb and Sr/Nb, may be fractionated. Indeed, HOFMANN et al. ( 1984) found that such a ratio, Rb/Sr, correlates with rubidium content in Kilauean lavas, and apparently was fractionated by up to 25% from the inferred source value. However, the intershield differences for La/Nb, Sr/Nb, and Zr/Nb are much larger (ranging up to a factor of 3) and FREY and RHODES (1993) showed that these intershield differences are independent of MgO content. Nonetheless, these ratios may have been fractionated upon melting; their relative compatibilities (see FREY et al., 1994) indicate that the measured trace element ratio is lower than the source ratio. Thus, the enriched source component is characterized by even higher Sr/Nb, La/Nb, and Zr/Nb than those in Koolau lava% Therefore, the correlation between isotopic composition and trace element ratio is strong evidence for concluding that the EM1 source component is characterized by high Sr / Nb, La/ Nb, and Zr / Nb compared to the FOZO component of HART et al. ( 1992) which dominates the lavas of Mauna Loa, Kilauea, and Loihi. It is tempting to attribute the high LREE/Nb signature to recycled continental crust-a possibility that has been posed before (e.g., HOFMANN and WHITE, 1982) and is plausible given the evidence for subduction of sediments in currently

1438

M. F. Roden et al.

HAWAIIAN KILA UE.4

KILA IlEA

MAUNA

LOA

THOLEIITES

)‘*.

MAUNA a

LOA

t’

*

[email protected] LANA,

LANAI

t*..

* LANAI

*

“’ %4L _*

9:a=r

MAUNA

. .

*

_

.

b

LOA

.

LANAI

*

LANAI

*

LANAI

n

MAUNA

L_l

0

*

q

MAUNA

LOA

LOA . ,

KILAUEA

KILAUEA

KILAUEA

m o n

,‘I’ LO

o.7030 + Sr/N'b"

r,,

Z;JNb

7

0.5

I’,“”

“#‘I”

1.0

La/Nb

iiOOLAU KAHOOLAWE HALEAKALA LOIHI MAUNA YEA

1.5

2.0

FIG. 5. Plots of strontium and neodymium isotopic composition vs. trace element ratios. Data sources for isotopic composition are as in Fig. I; additionally, Mauna Kea and Haleakala data are from FREY et al., ( 1991) and CHEN and FREY( 1985), respectively. Trace element data are from RODEN et al. (1984a),FREY et al. ( 1991, 1994), LEEMAN et al. (1993), BVSP (1981), FREYand CLAGUE(1983), WESTet al. (1992) and CHEN and FREY(1985). Sample 69 TAN-2 is not plotted because of the evidence for REE mobility in this sample.

active subduction zones (e.g., MORRIS et al., 1990). Moreover, the decoupling between trace element ratios (e.g., Rb/ Sr) and isotopic composition could be explained by trace element fractionation during subduction zone processing. However, the high 3He/4He ratios in Koolau lavas compared to MORB is inconsistent with the low 3He/4He ratios inferred for subducted material and observed in some arc magmas (e.g., TRULL et al., 1990a). In contrast, the HIMU component has been ascribed to recycled material and at least in one case, HIMU lavas (from St. Helena) exhibit low 3He/4He ratios consistent with the involvement of recycled material (GRAHAM et al., 1992). The high Ba/Nb ratios typical of some EM 1 OIB (Table 4) was one of the main lines of evidence cited in support of recycled material in EM1 (e.g., WEAVER, 199 1; CHAUVELet al., 1992). However, our data for Koolau (Table 4) as well as new data for Kerguelen lavas ( WEIS et al., 1993 ) show that EM 1 lavas do not always have high Ba/Nb ratios. In general, our arguments against a subduction related source are consistent with arguments based on Nb/U and Pb/Ce ratios in MORB and OIB and the HfNd isotopic correlation in oceanic basalts ( HOFMANN, 1986; PATCHETTet al., 1984; SUN and MCDONOUGH, 1989; BEN OTHMAN et al., 1989). RICHARDSONet al. ( 1982) originally suggested that the EM 1-dominated Walvis Ridge basalts formed by melting of metasomatized mantle and WRIGHT ( 1984) argued for a

metasomatized lithospheric source for Hawaiian tholeiites. Recent experimental evidence (e.g., WALLACEand GREEN, 1988) indicates that carbonatitic metasomatism may be common in the upper mantle; GREEN and WALLACE( 1988) predicted that the carbonatitic phase is likely to have high REE/HFSE ratios and LILE/HFSE ratios and to fractionate the HFSE from each other. Thus, the high La/Nb and Sr/ Nb ratios of Koolau lavas could be indicative of carbonatitic metasomatism. Recent experimental studies, however, indicate that carbonatitic melts in equilibrium with garnet, clinopyroxene, or amphibole will not have anomalously high La/Nb and Sr/Nb (GREEN et al., 1992; SWEENEYet al., 1992). This result is consistent with analyses of peridotite xenoliths which are interpreted to contain a carbonatitic component ( RUDNICKet al., 1993). In addition, these studies indicate that carbonatites have unusually low Zr/Nb and high Zr/Hf; in contrast Koolau lavas have relatively high Zr/ Nb (Table 4) and Zr/Hf (43.8 t 1.8, FREY et al., 1994) typical of OIB. Thus, the Koolau data are not consistent with derivation from a source affected by carbonatitic metasomatism. The trace element signature (high Sr/Nb, Zr/Nb, La/Nb, and low Rb/Sr) of the Koolau component is consistent with this component being residual after the extraction of a partial melt from a primitive source. The relative compatibilities of incompatible elements during this process were similar to

Nature of the magma source, Koolau volcano

those during the formation of the MORB source, except thorium is more depleted than rubidium and barium in Koolau lavas (Table 4). That is, moderately incompatible elements strontium, zirconium, and lanthanum were preferentially retained in the residue relative to the highly incompatible elements niobium and rubidium. This depleted trace element signature contrasts with the isotopic evidence which requires a relatively enriched or primitive trace element signature for a long time. Perhaps extraction of a small increment of a melt was associated with the incorporation of the Koolau component into the plume. Melting of a recently depleted source then produced the Koolau lavas as the plume approached the base of the lithosphere. Prior to the extraction of the small melt increment the initial material may well have been undifferentiated, primitive mantle.

Most Koolau tholeiites are characterized by 87Sr/86Sr = 0.7040-0.7043, ‘43Nd/‘44Nd = 0.51270-0.5 1276, *“Pb/ *04Pb 17.8-17.9, 207Pb/204Pb = 15.43-15.46, and 3He/4He = 11-14 times the atmospheric value, although a few tholeiites define the lower ‘43Nd/ ‘44Nd (to 0.5 1254) and higher 87Sr/86Sr (to 0.70455) limits for Hawaiian lavas. The 3He/ 4He ratios are higher than those of MORB but lower than those of lavas from Loihi. Moreover, the estimated source Rb/Sr ratio ( ~0.0 17 ) is clearly lower than primitive mantle. As a consequence it is unlikely that primitive, undegassed mantle is the source of the Koolau lavas. Some incompatible element abundance ratios correlate with neodymium and strontium isotopic compositions of Hawaiian tholeiites and thus provide strong evidence for differences in trace element ratios in the source regions. The source for the Koolau tholeiites is characterized by high Zr/ Nb, Sr/Nb, and La/Nb ratios compared to the sources for the Kilauea and Mauna Loa shields. This relative depletion in the highly incompatible element niobium, compared to more compatible elements, can most simply be explained by prior extraction of a partial melt from a primitive source. The near-bulk earth strontium, neodymium, and lead isotopic ratios require that this event occurred recently, perhaps during ascent within the plume. Lavas erupted earlier in the history of Koolau volcano may have incorporated this initial partial melt; thus the oldest lavas forming the Koolau shield may have very high 3He/4He ratios (>Loihi) as well as near-bulk earth strontium, neodymium, and lead isotopic ratios. Acknowledgments-This research was supported by the National Science Foundation. MFR thanks Laure Dosso for the use of the mass spectrometer at IFREMER (Brat). He analyses were performed by Tom Trull in the WHOI lab with support from NSF to Mark Kurz. We thank M. 0. Garcia and P. Schroeder for comments on an early version of the manuscript and F. Albarede and W. McDonough for journal review comments. Thanks to Patti Gary and Mary Grimes for help in manuscript preparation. Editorial handling: A. N. Halliday

REFERENCES BVSP ( 198 I ) Basaltic Volcanism on the Terrestrial Planets. Pergamon. D., WHITE W. M., and

PATCHET-T P. J.

geochemistry of marine sediments, island arc magma genesis, and crust-mantle recycling. Earth Planet. Sci. Lett. 94, I-2 1. CHAUVEL C., HOFMANNA. W., and VIDALP. ( 1992) HIMU-EM: The French Polynesian connection. Earth Planet. Sci. Lett. 110, 99-l 19.

CHEN C. Y. and FREYF. A. ( 1985 ) Trace element and isotope geochemistry of lavas from Haleakala volcano, East Maui: Implications for the origin of Hawaiian basalts. J. Geophys. Rex 90, 87438768.

DOELLR. R. and DALRYMPLEG. B. ( 1973) Potassium-argon ages and paleomagnetism of the Waianae and Koolau Volcanic Series, Oahu, Hawaii. Geol. Sot. Amer. Bull. 84, 1217-1242. Doss0 L., HANANB. B., BOUCAULTH., SCHILLING J.-G., and JORON J.-L. ( 1991) Sr-Nd-Pb geochemical morphology between 10” and I7”N on the Mid-Atlantic Ridge: A new MORB isotope signature. Earth Planet. Sci. Lett. 106, 29-43. FARLEYK. A., NATLANDJ. H., and CRAIGH. ( 1992) Binary mixing of enriched and undegassed (primitive?) mantle components (He, Sr, Nd, Pb) in Samoan lavas. Earth Planet. Sci. Lett. 111, 183199.

CONCLUSIONS

BEN OTHMAN

1439

(1989) The

FEIGENSONM. D., HOFMANNA. W., and SPERAF. J. (1983) Case studies on the origin of basalt II. The transition from tholeiitic to alkalic volcanism on Kohala volcano, Hawaii. Contrib. Mineral. Petrol. 84, 390-405. FODORR. V., MALTAD. V., BAUERG. R., and JACOBSR. S. ( 1989) Microbeam analyses of rare-earth element phosphate in basalt from Kahoolawe Island, Hawaii. In Proc. 24th Ann. Conf: Microbeam Analyt. Sot. (ed. P. E. RUSSELL), pp. 554-558. San Francisco Press. FREYF. A. and CLAGUED. A. ( 1983) Geochemistry of diverse basalt types from Loihi Seamount, Hawaii: Petrogenetic implications. Earth Planet. Sci. Lett. 66, 337-355. FREY F. A. and RHODESJ. M. ( 1993) Intershield geochemical differences among Hawaiian volcanoes: Implications for source compositions, melting process, and magma ascent paths. Phil. Trans. Roy. Sot. London A, 342, 121-136. FREY F. A., GARCIAM. O., WEST H., KENNEDY A., GURRIET P., and ALBAR~DEF. ( 199 1) The evolution of Mauna Kea Volcano, Hawaii: Petrogenesi of tholeiitic and alkalic basalts. J. Geophys. Res. 96, 14,347-14,375. FREY F. A.. GARCIAM. O., and RODENM. F. ( 1994) Geochemical characteristics of Koolau volcano: Implications of intershield geochemical differences among Hawaiian volcanoes. Geochim. Cosmochim. Acta 58, 1447-1468. GARCIAM. O., MUENOWD., and AGGLEYK. E. ( 1989) Major element, volatile, and stable isotopic geochemistry of Hawaiian submarine tholeiitic glasses. J. Geophys. Res. 94, 10,525-10,538. GRAHAM D., LUPTON J., ALBAR~DEF., 2nd CONDOMINESM. ( 1990a) Extreme temporal homogeneity of helium isotopes at Piton de la Fournaise, Reunion Island. Nature 347, 545-548. GRAHAMD., LU~~ONJ., and GARCIAM. (1990b) He isotopes in olivine phenocrysts from submarine basalts of Mauna Kea and Kohala, Island of Hawaii. Eos 71, 657. GRAHAM D., HUMPHRISS. E., JENKINSW. J., and KURZ M. D. ( 1992) Helium isotope geochemistry of some volcanic rocks from Saint Helena. Earth Planet. Sci. Lett. 110. 12 l-13 1. GREEN D. H. and WALLACEM. E. (1988)‘Mantle metasomatism by ephemeral carbonatite melts. Nature 336, 459-462. GREEN T. H., ADAMJ., and SIE S. H. ( 1992) Trace element partitioning between silicate minerals and carbonatite at 25 kb and application to mantle metasomatism. Mineral. Petrol. 46, 179184. HART S. R. ( 1988) Heterogeneous mantle domains: Signatures, genesis, and mixing chronologies. Earth Planet. Sci. Left. 90, 273296. HART S. R. and BROKS C. ( 1977) The geochemistry and evolution of early Precambrian mantle. Contrib. Mineral. Petrol. 61. 109HART S. R., HAURI E. H., OSCHMANNL. A., and WHITEHEADJ. A. ( 1992) Mantle plumes and entrainment: Isotopic evidence. Science 256,5 17-520. HOFMANNA. W. ( 1986) Nb in Hawaiian magmas: Constraints on source composition and evolution. Chem. Geol. 57? 17-30. HOFMANNA. W. ( 1988) Chemical differentiation of the Earth: The

I440

M. F. Roden et al.

relationship between mantle, continental crust, and oceanic crust. Earth Planet. Sci. Lett. 90, 297-314. HOFMANNA. W. and WHITE W. M. ( 1982) Mantle plumes from oceanic crust. Earth Planet. Sci. Lett. 57, 421-436. HOFMANNA. W. and WHITE W. M. (1983) Ba. Rb. and Cs in the Earth’s mantle. Z. Natwforsch 38a,~256-i66. HOFMANNA. W., FEICENSONM. D., and RACZEK1. ( 1984) Case studies on the origin of basalt: III. Petrogenesis of the Mauna Ulu eruption, Kilauea, 1969- t971. Curttrib. Mineral. Petrol. 88, 2435. KURZ M. D. ( 1986) Cosmogenic helium in a terrestrial igneous rock. Nature 320, 435-439. KLJRZM. D. (1993) Mantle heterogeneity beneath ocean islands: some inferences from isotopes. Phil. Trans. Roy. SIX. London A 342,91-103. KURZ M. D. and KAMMERD. P. ( 199 1) Isotopic evolution of Mauna Loa volcano. Earth Planet. Sci. Lett. 103, 257-269. KURZ M. D., JENKINSW. J., SCHLLING J. G., and HART S. R. ( 1982) Helium isotopic variations in the mantle beneath the central North Atlantic Ocean. Earth Pkmet. Sci. Lett. 58, l-14. KURZ M. D., JENKINSW. J.. HART S. R., and CLACUE D. (1983) Helium isotopic variations in volcanic rocks from Loihi Seamount and the Island of Hawaii. Earth PIanef. Sci. Letf. 66, 388-406. K~IRZM. D., GARCIAM. 0.. FREYF. A., and O’BRIENP. A. ( 1987) Temporal helium isotopic variations within Hawaiian volcanoes: Basalts from Mauna Loa and Haleakala. Geochim. Cosmochim. Acta 51, 2905-29 14. LANPHEREM. A. and DALRYMPLECi. B. ( 1980) Age and strontium isotopic composition of the Honolulu volcanic series, Oahu, Hawaii. Amer. J. Sci. 280-A, 736-751. LATOURZTTE1‘. 2. and BURNETTD. S. (1992) Ex~~men~l determination of U and Th partitioning between clinopyroxene and natural and synthetic liquid. Earth Planet. Sci. Left. 110, 227244. LEEMANW. P., GERLACHD. C., GARCIAM. O., and WEST H. B. ( 1993) Geochemical variations in lavas from Kahoolawe Volcano, Hawaii: Evidence for open system evolution of plume-derived magmas. Coritrib. ~~~zera/. Pefrol. (in press). MARTYB. and LUSSIEZP. ( 1993) Constraints on rare gas partition coefficients from analysis of olivine-glass from a picritic mid-ocean ridge basalt. Chem. Ged. 106, 1-7. M~DO~JGAL.L1. and AZIZ-UR-RABMAN( 1972) Age of the GaussMatuyama boundary and the Kaena and Mammoth event. Earfh Planet. Sci. Mt. 14, 367-380. MORRISJ. D., LEEMAN W. P.. and TERA F. ( 1990) The subducted component in island arc lavas: Constraints from Be isotopes and B-Be systematics. Nature 344, 3 i-36. NEWSOM H. E., WHITE W. M., J~CHUM K. P,, and HOFMANN A. W. ( 1986) Siderophile and chalcophile element abundances in oceanic basal& Pb isotope evolution and growth of the Earth’s core. Earth Planet. Sci. Lett. 80, 299-3 13. O’NIONS R. K., HAMILTONP. J., and EVENSEN,N. M. ( 1977) Variations in 14’Nd/‘@Ndand ?Sr/%r ratios in oceanic basal&. Earth Planef. Sci. Lnt. 34. 13-22. PATCHETTP. J., WHITE W. M.. FELDMANH., KIELINCZUKS., and HOFMANNA. W. ( 1984) Hafniumirare earth element fractionation in the sedimentary system and crustal recycling into the Earth’s mantle. Earth Planet. Sci. Lett. 69, 365-378. RAUTENSCHLEINM., JENNER G. A.. HERTOGENJ., HOFMANN A. W., KERRIC~~R., SCWMINCKEH. U., and WHITEW. M. ( 1985) Isotopic and trace element composition of volcanic glasses from the Akaki Canyon, Cyprus: Implications for the o&in of the Troodos ophiolite. Earth Planet. Sci. Leti. 75. 369-383. RHODESJ. M., WENZ K. P.. NEAL C. A., SPAR& J. W., and LOCKWOOD J. P. ( 1989) Geochemical evidence for invasion of Kilauea’s plumbing system by Mauna Loa magma. Nature 337, 257-260. RICHARDSON5. II.. ERLANKA. J., DUNCAN A. R., and RFID D. L. ( 1982) Correlated Nd, Sr, and Pb isotope variation in Walvis Ridge basalts and implications for the evolution of their mantle source. Earth Planet. Sci. Left. 59, 327-342.

RODENM. F., FREYF. A., and CLAGUE D. A. ( t 984a) Geochemistry of tholeiitic and alkalic lavas from the Koolau Range, Oahu, Hawaii: implications for Hawaiian volcanism. Earth Planet. Sci. Letf. 69, 141-158. RODENM. F., FREYF. A., and FRANCISD. M. ( 1984b) An example ofconsequent mantle metasomatism in peridotite inclusions from Nunivak Island, Alaska. .I. Petrol. 25, 546-577. RUDNICKR. L., MCDONO~GH W. F., and CHAPPELL8. W. (1993) Carbonatite metasomatism in the northern Tanzanian mantle: Petrographic and geochemical characteristics. Earth Planet. Sci. L,etf. 114, 463-475. STAUDIGEL H., ZINDLER A., HART S. R., LESLIET., CHEN C. Y., and CLAGUED. A. ( 1984) The isotope systematics of a juvenile

intraplate volcano: Pb, Nd. and Sr isotope ratios from Loihi seamount, Hawaii. Earth Planet. Sci. Left. 69, 13-29. STILLE P.. UNRLJH D. M., and TATSUMOTOM. (1983) Pb, Sr, Nd, and Hf isotopic evidence of multiple sources for Oahu. Hawaii basalts. Nature 304, 25-29. STILLEP., UNRUH D. M., and TATSUMOTOM. ( 1986) Pb, Sr, Nd, and Hf isotopic constraints on the origin of Hawaiian basalts and evidence for a unique mantle source. Geochim. Cosmochim. Aria 50, 2303-23 19. SU~VS.-S. and McDovOucH W. F. f 1989) Geochemical and isotopic systematics of oceanic bar&s: implications for mantle composition and processes. Geol. Sot. Spec. Pub. 42, pp. 313-345. SWEENEY R. J., GREEND. H., and SIE S. H. ( 1992) Trace and minor element partitioning between garnet and amphibole and carbonatitic melt, Earth Planet. Sci. Left. 113, I-14. TARAS B. D. and HART S. R. ( 1987) Geochemical evolution of the New England Seamount Chain: Isotopic and trace-element constraints. Chem. Gd. 64, 35-54. TATSUMOTOM. ( 1978) isotopic compositions of lead in oceanic basalt and its implication to mantle evolution. Earth Planet. Sci. Lxtf. 38, 63-87. TATSUMOTOM., HEGNERE., and UNRUH D. M. (1987) Origin of the West Maui volcanic rocks inferred from Pb. Sr, and Nd isotopes and a multicomponent model for oceanic basalts. VSGSPrqt Paper 1350, pp. 723-744. TRULLT., PERL~TM. R., and KURZ M. D. ( 1990a) He and Sr isotopic constraints on subduction contributions to Woodlark Basin volcanism. Solomon Islands. Geochim. Cosmochim. Acta 54. 441454. TRULL T., RODEN M., HART S., FREY F., GARCIAM., and KURZ M. ( 1990b) He-Sr-Nd-Pb isotopic comaositions of Koolau Basalts: More precise definition ofthe Hawaiian Plume. Eos 71,657. WALLACEM. E. and GREEN D. H. (1988) An ex~~mental determination of primary carbonatite magma composition. N’ature335, 343-346. WEAVERB. L. ( 199 I ) The origin of ocean island basalt end-member compositions: trace element and isotopic constraints. Earth Planet. Sci. Lett. 104, 38 l-397. WEISD., FREYF. A., LEYRITH., and GAU’IXR 1. ( 1993) Kerguelen archipelago revisited: Geochemical and isotopic study of the southeast province lavas. Earth P/anef. Sci. Lett. 118, 10 I - 119. WENTWORTHC. K. and WINCHELLH. ( 1947) Koolau basalt series. Oahu, Hawaii. Gwl. S’oc.Amer. Bull. 58, 49-72. WESTH. B., GEmACt- D., LEEMANW. P., and GARCIAM. 0. ( 1987) Isotopic constraints on the origin of Hawaiian lavas from the Maui Volcanic Complex, Hawaii. Nature 330, 2 16-219. WESTH. B., GARCIAM. O., GERLACHD. C., and ROMANOJ. ( 1992) Geochemistry of tholeiites from Lanai, Hawaii. Conrn‘h. ~~~rzer~~, Petrol. 112, 520-542. WRIGHTT. L. ( 1984) Origin of Hawaiian tholeiite: A metasomatic model. J. Geophy-s.Re.9. 89, 3233-3252. YAXLEYG. M., CRAWFORDA. J., and GREEND. H. ( 1991) Evidence for carbonatite metasomatism in spine1 peridotite xenoliths from western Victoria, Australia. Earth Planet. S’ci. Letf. 107,305-3 17. ZINVLERA. and HART S. R. ( 1986) Chemical geodynamics. Ann Rev. Earth Pkmef. Sci. 14,493-57 1.