deep sea drilling project initial reports volume 58
TRANSCRIPT
33. GEOCHEMISTRY OF BASALTS FROM THE SHIKOKU AND DAITO BASINS,DEEP SEA DRILLING PROJECT LEG 58
Nicholas G. Marsh, Andrew D. Saunders,1 and John Tarney, Department of Geological Sciences,University of Birmingham, U.K.
andHenry J. B. Dick, Department of Geology and Geophysics, Woods Hole Oceanographic Institution,
Woods Hole, Massachusetts
ABSTRACTLeg 58 successfully recovered basalt at Sites 442, 443, and 444, in
the Shikoku Basin, and at Site 446 in the Daito Basin. Only at Site442 did penetration reach unequivocal oceanic layer 2; at the othersites, only off-axis sills and flows were sampled. Petrographic obser-vations indicate that back-arc basalts from the Shikoku Basin, withthe exception of the kaersutite-bearing upper sill at Site 444, are min-eralogically similar to basalts being erupted at normal mid-oceanridges. However, the Shikoku Basin basalts are commonly very vesic-ular, indicating a high volatile content in the magmas. Site 446 in theDaito Basin penetrated a succession of 23 sills which include bothkaersutite-bearing and kaersutite-free basalt varieties.
A total of 187 samples from the four sites has been analyzed formajor and trace elements using X-ray-fluorescence techniques. Chem-ically, the basalts from Sites 442 and 443 and the lower sill of Site 444are subalkaline tholeiites and resemble N-type ocean-ridge basaltsfound along the East Pacific Rise and at 22° N on the Mid-AtlanticRidge (MAR), although they are not quite as depleted in certain hy-gromagmatophile (HYG) elements. They do not show any chemicalaffinities with island-arc tholeiites. The basalts from Site 446 andfrom the upper sill at Site 444 show alkaline and tholeiitic tendencies,and are enriched in the more-HYG elements; they chemically resem-ble enriched or E-type basalts and their differentiates found alongsections of the MAR (e.g., 45°N) and on ocean islands (e.g., Icelandand the Azores).
Most of the intra-site variation may be attributed to crystal set-tling within individual massive flows and sills, to high-level fractionalcrystallization in sub-ridge magma chambers, or, where there is evi-dence of a long period of magmatic quiescence between units, tobatch partial melting. However, the basalts from Sites 442 and 443and from the lower sill at Site 444 cannot easily be related to thosefrom Site 446 and the upper sill at Site 444, and it is possible that thedifferent basalt types were derived from chemically distinct mantlesources. From comparison of the Leg 58 data with those alreadyavailable for other intra-oceanic back-arc basins, it appears that themantle sources giving rise to back-arc-basin basalts are chemically asdiverse as those for mid-ocean ridges. In addition, the high vesicu-larity of the Shikoku Basin basalts supports previous observationsthat the mantle source of back-arc-basin basalts may be contami-nated by a hydrous component from the adjacent subduction zone.
INTRODUCTION Shikoku back-arc or marginal basin, lying between theIwo Jima and Kyushu-Palau Ridges, sampled basalts
During Leg 58, basaltic rocks were recovered at four are no older than 21, 17, and 15 m.y., respectively, fromof the five sites occupied in the Shikoku and Daito both sides of the north-south-trending axial zone. OnlyBasins. At Sites 442, 443, and 444, in the intra-oceanic at Site 442 were basalts penetrated which belong un-
equivocally to oceanic layer 2; at Sites 443 and 444, re-covered basalts probably represent off-axis intrusive or
1 Present address: Department of Geology, Bedford College, extrusive activity. At Site 446, in the Daito Basin, sillsUniversity of London. of post-early Eocene age were penetrated, but basement
805
N. G. MARSH, A. D. SAUNDERS, J. TARNEY, H. J. B. DICK
was not reached; in this region it is thought to beoceanic (Karig, 1975).
It has been established by detailed geophysical inves-tigation and by dredging operations that most back-arcor marginal basins are of extensional origin, and arefloored by basaltic crust broadly comparable in struc-ture and composition to that of the major ocean basins(Karig, 1971, 1974; Hart et al., 1972; Barker, 1972;Hawkins, 1974; Gill, 1976). However, there is still noconsensus as to whether all marginal-basin basalts areencompassed by the range of basalt compositions pres-ently being erupted at mid-ocean ridges, or whether theyhave geochemical characteristics which are, to a smalldegree, transitional towards orogenic basalts, as in-dicated by detailed chemical investigations of lavasfrom back-arc centers in the East Scotia Sea (Saundersand Tarney, 1979) and Bransfield Strait (Weaver et al.,1979). Similar studies of basalts from the MarianaTrough (Hart et al., 1972) and Lau Basin (Hawkins,1976, 1977), however, do not corroborate the evidencefrom the Scotia Sea and Bransfield Strait basins. It isimportant, therefore, to assess the compositions ofmagmas being erupted in several back-arc basins beforeany general conclusions about their genesis can bemade. The prominent role of back-arc basins in the for-mation of many ophiolites (Dewey, 1976; Saunders etal., 1979) emphasizes the importance of being able tochemically characterize and distinguish back-arc-basinbasalts from ocean-ridge basalts, and this can be achievedonly when sufficient data on modern back-arc-basin ba-salts are available.
Comprehensive studies of basalts from the North At-lantic Ocean have illustrated the chemical variability ofmagmas being erupted along the Mid-Atlantic Ridge(Schilling, 1973, 1975a,b; Hart et al., 1973; White et al.,1976; Bougault et al., 1979; Tarney et al., 1978, 1979, inpress; Wood et al., 1979). It is therefore important, be-fore any detailed comparisons between back-arc-basinand ocean-ridge basalts are made, that the compositionof ocean-ridge basalts be clearly delineated. Sun et al.(1979) and Tarney et al. (in press) have defined threemain basalt types presently being erupted in the NorthAtlantic: normal, or N-type, depleted ocean-ridge ba-salts; enriched, or E-type, basalts; and T-type basalts,which have a composition transitional between N- andE-type basalts. These terms will be defined in detaillater. The variations in chemistry appear to be in partdue to chemical heterogeneities in the mantle source, inaddition to variations caused by partial melting andfractional crystallization (see discussions by Bougault etal., 1979; O'Nions et al., 1976; Tarney et al., 1978,1979, in press; Wood et al., 1979). It is necessary toassess whether or not back-arc-basin basalts exhibit asimilar range in chemistry.
Leg 58 has recovered basaltic material which affordsan excellent opportunity to undertake detailed petrolog-ical and geochemical studies of basalts from a "typical"Western Pacific intra-oceanic back-arc basin. In thispaper we shall be concerned essentially with X-ray-fluorescence data on major and trace elements. Basaltsfrom Leg 49 in the North Atlantic were analyzed using
the same analytical equipment (Tarney et al., 1978), al-lowing regional comparisons without inter-laboratoryerror. Instrumental conditions were arranged to providethe highest precision possible, especially for the traceelements. Thus, reported geochemical differences, bothwithin and between sites, are significant. However, wehave confined ourselves to an essentially qualitativeassessment of the results.
REGIONAL AND TECTONIC SETTING
The Shikoku Basin
The Shikoku Basin (Figure 1) has the form of anelongate trough some 1200 km long, 500 km wide, and4.5 km deep, bounded by the I wo Jima Ridge to the eastand by the Kyushu-Palau Ridge to the west. The north-ern end of the basin is delineated by the Nankai Troughand by the Island of Shikoku. Southward, the basincontinues without any physical barrier into the PareceVela Basin.
The Shikoku Basin has been identified as an inter-arcbasin separating a composite frontal arc system — com-prising the I wo Jima Ridge, Bonin Trough, and BoninRidge — from the remnant Kyushu-Palau Ridge. Sub-duction of Pacific Ocean floor beneath the Bonin Ridgeis presently occurring along the Izu-Bonin Trench, andit is apparent that the Bonin Trough is an incipientback-arc basin (Karig, 1971). The Nankai Trough prob-ably represents the site of subduction of Shikoku Basinfloor beneath southern Japan (Karig, Ingle, et al., 1975;Kobayashi and Isezaki, 1976).
Figure 1. Map of the northern Philippine Sea, showingthe location of Leg 58 sites in the Shikoku andDaitoBasins.
806
GEOCHEMISTRY OF BASALTS
Geophysical studies have shown that the ShikokuBasin is floored by oceanic crust, although the absenceof detectable shallow earthquakes beneath the axialzone of the basin (Katsumata and Sykes, 1969) suggeststhat spreading is not now taking place. However, the ax-ial zone has the high heat flow and topography charac-teristic of many presently active oceanic spreading cen-ters (Kobayashi and Isezaki, 1976), although the mag-netic lineations observed within the basin are of loweramplitude than those found in most major ocean basins.Linear, weak magnetic anomalies running parallel to theKyushu-Palau Ridge are well established for the west-ern side of the basin, but because of rough topographythe interpretation of anomalies in the eastern side of thebasin remains equivocal (Watts and Weissel, 1975; Ko-bayashi and Isezaki, 1976; Shih, this volume). Availabledata suggest that spreading in the Shikoku Basin beganbetween 25 and 27 Ma and ceased about 17 Ma (Wattsand Weissel, 1975; Kobayashi and Isezaki, 1976). Thiseffectively separated the continental crustal slivers ofthe Kyushu-Palau and Iwo Jima Ridges (Murauchi etal., 1976). Alternative interpretations of the data aresuggested by Tomada et al. (1975), who consider thatthe spreading occurred later, between 22 and 10 Ma,and by Murauchi et al. (1976), who have proposed thatspreading occurred in two distinct episodes, 24 to 19 Maand 12 to 16 Ma. The interpretations agree, however, insuggesting that spreading is not now occurring in theShikoku Basin. It would appear that the locus ofspreading has migrated westward into what is now theBonin Trough.
Northeast Philippine Basin
The northeast corner of the Philippine Basin isroughly triangular; it is bordered by the Ryukyu Trenchto the northwest and by the Kyushu-Palau Ridge to theeast (Figure 1). The area comprises a complex group ofbasins and topographic highs, including the Oki-Daitoand Daito Ridges, which apparently are structurallysimilar to island arcs (Murauchi et al., 1968). The inter-vening basins (e.g., Daito Basin) are sufficiently deep tobe floored by old ocean-crust, possibly of early-Tertiaryor late-Cretaceous age (Karig, 1975). However, a thickpile of sediments now obscures the oceanic layer 2, al-though heat-flow measurements (Watanabe et al., 1970)show values similar to those of old ocean crust.
Rocks previously dredged from the Daito Ridge in-clude greenschists, hornblende schists, serpentinites, an-desite, and diorite. Andesite, basalt, and granodioritehave been recovered from the Oki-Daito Ridge. Fromthe geophysical evidence and dredge hauls it has beensuggested that the Daito and Oki-Daito Ridges formedas island arcs (Murauchi et al., 1968; Karig, 1975;Mizuno et al., 1975, in press). These arcs and the ad-joining pieces of ocean crust drifted some 2000 kmnorthward during the last 52 m.y. as a result of sea-floorspreading in the West Philippine Sea, before being trappedagainst the Kyushu-Palau Ridge.
ANALYTICAL TECHNIQUES
Sample Preparation
Minicores or quarter-core sections were washed indistilled water to remove surface contamination, anddried at 100°C prior to being ground to a fine powder(less than 60 µm) in the shipboard tungsten-carbideTEMA swing mill. Powder briquettes 46 mm in diame-ter were made by binding 15 grams of rock powder withabout 30 drops of a 7 per cent aqueous solution ofpolyvinyl alcohol, under a pressure of 15 tons againstthe polished, hardened faces of a steel die.
Representative samples from Sites 442, 443, and 444,and all the samples from Site 446, were made into fusionbeads for the purpose of major-element analysis. Thistechnique eliminates mineralogical effects and reducesmatrix absorption. The rock powders were dried over-night at 110°C, and ignition loss was determined byheating at 1000°C in an electric furnace for 1 hour.Beads were prepared by mixing sufficient rock powderand flux in a Pt/Au crucible to give an exact ratio of 1:5after fusion (0.9 gram rock, 4.5 gram flux, ignited mate-rial). Fusion was carried out at 1200°C, with frequentswirling, for 20 minutes in a vertical-tube electric fur-nace. The flux, 80 per cent Li metaborate, 20 per cent Litetraborate (Johnson Matthey Spectroflux 100B), is de-scribed by Bennett and Oliver (1976) and is designed todissolve the widest possible range of silicate and ceramicmaterials.
After fusion, the melt was pressed in aluminummolds, using equipment similar to that described byHarvey et al. (1973); the beads were transferred toheated insulating blocks for annealing at 200° C for 30minutes, and finally allowed to cool in the blocks. Thelithium borate beads are slightly hygroscopic and mustbe stored in sealed polypropylene bags in a desiccator toprevent atmospheric attack of the bead surface.
X-Ray-Fluorescence Analysis
The equipment used was a Philips PW1450 automaticspectrometer with a PW1466 60-position sample chang-er. Count data from the spectrometer were processedusing a 32K Digital PDP11-03 computer, and the Uni-versity of Birmingham ICL1906A computer. Trace ele-ments were determined on 46-mm powder briquettes,using molybdenum (Y, Sr, Rb, Th, Pb, Ga, Zn, Ba) andtungsten (Ni, Cr, Ce, La, Zr, Nb) anode X-ray tubes.Major- and minor-element analyses of the basalts fromSites 442, 443, and 444 were determined on 46-mm pow-der briquettes and representative fusion beads using achromium-anode X-ray tube. Major- and minor-ele-ment analyses of basalts from Site 446 were determinedon fusion beads, using a rhodium-anode X-ray tube. In-strumental repeatability was ensured by ratioing countsto a reference sample, re-analyzing every fourth sample(Rh and Cr tubes) or at the beginning of every loadingof 60 samples (Mo and W tubes).
807
N. G. MAR^H, A. D. SAUNDERS, J. TARNEY, H. J. B. DICK
Instrumental Conditions
Details of instrumental conditions for analysis oftrace elements are given in Tarney et al. (1978).
The Rh-anode X-ray tube, operated at 30 kV, 70 mA,provides equivalent or superior sensitivity for thoseelements previously analyzed using a Cr anode, with theexception of Ti, Ca, and K. However, these elements areextremely sensitive to Cr excitation; it is usually nec-essary to greatly reduce tube power to avoid saturatingthe detector, and therefore no problems are encounteredwith the reduced efficiency of the Rh anode. An addi-tional advantage is that the analysis of Mn is possiblewithout filtering out the Cr K radiation from the tubetarget.
The Kα wavelengths of all major elements were ana-lyzed using the flow-proportional detector (argon/meth-ane) with automatic pulse-height selection. The fine col-limator (0.15 mm) was used for all elements except Naand Al, for which the coarse collimator (0.55 mm) wasemployed. The crystals used were: LiF2Oo (Mn), LiF220
(Fe), PET (Ti, Ca, K, Si, Al), T1AP (Mg, Na), and G e m
(P). The germanium crystal was chosen for P to avoidthe overlap of the escape peak arising from the second-order CaKß on PKα. A symmetrical, 50 per cent widthwindow was used for the pulse-height discriminator forall elements, with the exception of Mg, for which a set-ting of 20 to 65 per cent is needed to reduce the overlapof the third-order CaKα. However, both these Ca in-terferences are considerably less than when using Cr ex-citation.
Calibration
Initial trace-element calibrations were prepared froma wide range of rock types spiked with an appropriatepure-element compound (Leake, Hendry, et al., 1969).These values were then related using either the WL^ orMoKα Compton scatter lines to provide correction fortotal mass absorption for all emission lines shorter thanthe iron absorption edge (Reynolds, 1967). Ce and Lacalibrations were prepared using basalts previously ana-lyzed by isotope dilution. The accuracy of the variouscalibration lines was monitored by analyzing severalinternational reference samples of suitable composi-tions.
For the major elements, primary calibrations and " α "(matrix-dependent constant) correction factors were cal-culated from binary and ternary mixtures of appropri-ate pure chemicals and referred to an average basaltcomposition. The validity of these calibrations waschecked against departmentally analyzed material andinternational reference standards. Blank values for mi-nor elements were checked using pure SiO2, A12O3, andBeO, and were found to be low but significant forNa2O, K2O, and P2O5. For these elements at very lowlevels the values produced from undiluted powder bri-quettes are preferred.
Precision and Accuracy
The precision of measurement and sample prepara-tion is illustrated by the standard deviation of six
separate beads of the internal ocean basalt referencesample BOB-1 (Table 1). The trace-element data inTable 1 were determined on powder briquettes. The in-strumental precision with the techniques used is high,even at relatively low trace-element concentrations.However, a potential problem of instrumental analysisis long-term drift when periods greater than 12 monthsare considered. This has been kept to a minimum by us-ing 30 samples from Leg 49 (Tarney et al., 1979) to-gether with other basaltic material, selected to give aswide a concentration range as possible for all elements.These samples are analyzed together with the unknowns,ensuring compatibility of data between different legs.For major-element analysis it is not desirable to re-analyze fusion beads over long periods of time, hence anumber of internal reference samples are fused and ana-lysed with each batch of unknowns.
Accuracy of analysis can only be judged against in-ternational reference samples which, for trace elements,do not cover the range of concentrations encountered inmost DSDP legs. Data for trace elements in basalt
TABLE 1Analyses of Standard Reference Samplesa
Component(% or ppm)
SiO2
TiO2
AI2O3F e 2 O 3
MnOMgOCaONa2OK 2 O
P2O5
Standard
BOB-1
51.02 ± 0.0781.28 ± 0.004
16.55 ± 0.0128.48 ± 0.0130.14 ± 0.0017.58 ± 0.030
11.39+ 0.0273.10+ 0.0740.37+ 0.0030.16 ± 0.001
AVG-1
60.191.01
17.446.980.101.564.874.402.990.49
JB-1
53.131.30
14.489.020.157.799.412.871.450.25
Total 100.07 100.03 99.85
BOB-1 BCR-1 JB-1
NiCrbRbSrYZ I L•
NbBa d
LaCeZnGa
106.4+ 2.1257± 10
4.8 ± 0.6202.1 ± 1.0
26.9 + 0.5107.6+ 3.0
5.3 ± 1.650.4+ 3.6
6.4 ± 1.916.6 ± 1.761.5 + 2.117.6 + 0.7
191848
32936
19112
6952655
12025
134359
42442
23150
35500
38668017
a F e 2 θ 3 is total iron as F e 2 θ 3 ; mean and 1 sigmastandard deviation of six fusion beads of BOB-1,and means of three fusion beads of internationalreference samples; all values are for ignited sam-ples; average concentration of BOB-1 determinedon 22 consecutive days (with 1 sigma variation);average concentration of two samples of interna-tional reference material.
t>Not corrected for V interference.cCorrected for Sr interference.^Corrected for Ce interference.
808
GEOCHEMISTRY Cg BASALTS
BCR-1 are given in Tarney et al. (1979) and for a widerrange of materials in Hendry (1975). Major elements forAGV-1 and JB-1 are given in Table 1.
SHIKOKU BASIN
Three sites, comprising a total of six holes, weredrilled in the Shikoku Basin during Leg 58. The igneousstratigraphy of the recovered sections from each ofthese sites is given in Figures 2, 17, and 27 in this paper,and although detailed petrographic descriptions aregiven in the relevant site chapters (this volume), briefresumes are included in the following sections.
Site 442
Site 442 (Figure 1) is approximately 50 km west of theaxial zone of the Shikoku Basin, at 28°59.00'N, 136°03.43 'E, on magnetic anomaly 6 (Kobayashi and Naka-ta, 1977; 19-20 m.y.: LaBrecque et al., 1977). Threeholes were drilled; Hole 442 was a test hole for the re-entry cone, and Holes 442A and 442B cored approxi-mately 158 meters of massive basalt flows and pillowbasalts beneath some 300 meters of sediment, at a waterdepth of about 4635 meters.
Site 442 is the only site of Leg 58 from which pillowbasalts typical of oceanic Layer 2 were definitelyrecovered. Two lithologic units have been recognized:Unit 1, comprising mainly massive flows or sills, andunit 2, the underlying pillow lava sequence (Figure 2).The basalts from both units are vesicular and sparselyolivine bearing. Unit 1 is separated from unit 2 by ap-proximately 2 meters of brown mudstone, for whichmicrofossils indicate an age between 18 and 21 m.y.This is in agreement with the magnetic-anomaly age
(19-20 m.y.) of the basement at this site. The oldestsediments overlying unit 1 are limestones containingmicrofossils giving an age between 15 and 17 m.y.
Unit 1 consists of approximately 59 meters of fine- tomedium-grained basalt, which has been divided into 16sub-units. Many of these sub-units have fine-grained oroccasionally aphanitic chill zones, and the absence ofglassy margins suggests that these basalts were emplacedas sills or as massive flows (Site 442 report, this vol-ume). However, the presence of glassy selvages and var-iolitic basalts in Core 6 of Hole 442B indicates that theunit may consist, at least in part, of pillow lavas.
The sub-units of unit 1 grade from fine-grained ba-salts to medium-grained dolerites with an inequigranu-lar, subophitic texture. Seriate textures are common. Inthe finer-grained samples, intersertal to intergranularaugite and plagioclase (An55_75) and interstitial magnet-ite are embedded in a cryptocrystalline groundmasswhich is often replaced by clays. Strongly zoned clino-pyroxene crystals, colored green in their centers tobrown at the edges, and "bow-tie" zoned clinopyrox-enes are present in the coarsest-grained sections of somesub-units. In these coarse-grained rocks, it is difficult toascertain whether such crystals are phenocrysts or mere-ly large groundmass crystals. However, they frequentlyform glomerocrystic intergrowths with zoned plagio-clase crystals (core compositions of An7O_85). Chromespinel is a common accessory phase throughout unit 1,generally appearing as inclusions in plagioclase pheno-crysts. Chrome-spinel inclusions also occur in clinopy-roxene phenocrysts. The basalts and dolerites of unit 1tend to be very vesicular, often with two distinct popula-tions of vesicles. Most sub-units have 10 to 30 per centsmall vesicles ( < l mm) and generally 1 to 5 per cent
Core
No.
Lithologic
Unit
Magnetic
InclinationZ f { p p m ) ( p p m ) ( p p m ) F e / M g Y / Z r T j / Z r
280 η
300-
320-
3. 3 4 0 -
Q.
Qc 3 6 0 -
Su
b-b
ott
oi
CO
4 0 0 -
4 2 0 -
4 4 0 -
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
Mudstone and
Unit 1(sub-units A to P)
Massive aphyricvesicular basalts
Unit 2Aphyric vesicular
pillow basalts
+90°r . . . . . . .
— —
f -90°
'I
II"
r
80 100 120
I 1 1
.V
*
-
•*i r
20 40 601 1 1
V
.V•* •"•• *.
t
f
100 200 300 400I I I I
• •••*3»
••-V
•
*•i
I
*
1 2 3
I I 1
1
<V•
\
m
*
0.1 0.2 0.3 0.41 I 1 I
•*••*••*
•s
60 70 80 90
1 1 1 1
. 9
•
.
t
u
1
2
3
4
Figure 2. Down-hole variation in lithology, magnetic inclination, and selected geochemical parameters at Site 442.
809
N. G. MARSH, A. D. SAUNDERS, J. TARNEY, H. J. B. DICK
large vesicles (2-5 mm). Pipe vesicles are also foundnear the chill margins of several sub-units.
The pillow basalts of unit 2 have glassy margins,generally less than 1 cm thick, and contain 1 to 2 percent small plagioclase laths at their inner margins. Theglassy zones grade abruptly into variolitic zones towardthe pillow interiors. The varioles consist of small anhe-dral plagioclase laths and clinopyroxene microlites in aglassy matrix. The cores of the pillows are usually finegrained, with an intersertal texture comprising 25 percent plagioclase, (An55_7O), 30 per cent augite, 2 to 3 percent granular magnetite, and a glassy or (more com-monly) a clay-rich matrix. Vesicles are again common,composing up to 40 per cent of the rock, and are usuallyless than 1 mm across. However, larger vesicles up to 10mm across do occur, and pipe vesicles are common,usually in a zone between 1 and 10 cm from the glassymargin.
Low-temperature alteration is a general feature of allthe basalts recovered at this site. The cryptocrystallineor glassy matrixes are commonly replaced by clay min-erals. The basalts are commonly cut by calcite- and clay-lined fractures or veins, and the vesicles are generallyfilled or lined by similar material. Zeolites are rare, andpyrite was found in small amounts in vesicles and frac-tures of sub-unit IK only, which also contained somechlorite and talc-like alteration products (Site 442report, this volume).
Major- and trace-element data for the basalts recov-ered at Site 442 are listed in Tables 2 and 3. In thefollowing discussion, emphasis will be placed on thebasalts recovered from Hole 442B, the range of com-positions of which generally encompass those fromHole 442A. The basalts have been grouped in differentgeochemical units according to variations in their mi-nor- and trace-element contents, as illustrated in Figure2. These variations broadly coincide with the lithologicand magnetic units, although lithologic unit 2 (Hole442B) has been subdivided into three geochemical units.The geochemical units and the lithologic units do notnecessarily coincide.
Site 442 basalts are hypersthene and olivine normative(subalkaline), with practically no normative nepheline orquartz (Figure 3); the minor transgressions into the Neand Q fields on Figure 3 may be due to the presence ofsecondary carbonate and clays, respectively. The basaltsalso have a high normative-plagioclase content, eventhough plagioclase phenocrysts are not a major phase,and this is shown in Figure 4, where the basalts plot asplagioclase tholeiites on the pyroxene-plagioclase-olivinefield of Shido et al. (1971). Geochemical unit 442-1 isgenerally more olivine normative than the underlyingpillow lavas (units 442-2 to 442-4), and this is alsoreflected in the lower Fe/Mg ratios, and higher Cr andNi abundances in Unit 442-1 (Figures 5 and 6).
There is considerable downhole variation in absoluteelement abundances at Site 442. The essential featuresof this variation are illustrated in Figure 2. There is aclear correlation between units based on magnetic in-clination and geochemistry, although geochemical unit
TABLE 2Analyses, Element Ratios, and CIPW Norms
for Basalts from Hole 442A&
CoreSectionInterval (cm)
SiO2 (%)TiO2
A12O3
Fe 2 O 3
MnOMgOCaONa2OK 2 Op 2 θ 5
Total
Ni (ppm)CrZnGaRb
SiYZr
Nb
BaLaCe
PbTh
Zr/NbTi/ZrY/ZrCe/ZrBa/ZrCeN/YN
Fetot/MgK/RbBa/Sr
QOrAbAnNeDi
HyOIMt11
Ap
442A-311
85
49.61.24
16.28.660.146.12
13.462.700.260.18
98.55
45
220
78
174
186
3384
337
4
10
5< l
28
89
0.390.120.440.741.64
5460.20
0.01.6
23.231.7
0.028.5
5.3
4.61.5
2.4
0.4
322
48.61.12
15.88.380.145.38
15.192.670.400.16
97.83
57234
6317
9
18328
864
28
211
3< l
21
780.330.130.330.961.81
3710.15
0.02.4
18.630.5
2.437.2
0.0
3.91.5
2.2
0.4
322
80
49.41.16
16.07.290.106.77
13.792.780.250.15
97.72
130
2406917
4182
26
86< l29
4
13
51
_
810.300.150.341.231.25
5150.16
0.0
1.524.131.1
0.030.7
1.0
7.1
1.32.30.4
a3i
124
47.51.09
15.47.960.146.16
16.712.400.240.16
97.77
5 3216
11616
4
18628
82
321
313
7< l
27
800.340.160.261.141.50
5060.11
0.01.5
12.131.2
4.7
42.80.03.11.42.1
0.4
332
57
49.91.23
15.99.660.136.53
11.472.770.380.15
98.13
49
2336318
7189
29
861
27
619
< l< l
86
860.340.220.311.611.72
4450.14
0.02.3
23.930.4
0.0
21.813.5
2.81.72.4
0.4
333
68
49.81.22
15.89.130.127.17
11.602.820.330.14
98.20
77245
6618
6182
26
89
330
313
3< l
30
820.290.150.341.231.48
451
0.16
0.02.0
24.330.2
0.022.410.7
5.31.62.4
0.3
334
111
49.91.20
15.210.09
0.136.55
11.742.840.520.14
98.32
35214
6519
26178
27
88
331
512
3< l
2982
0.310.140.351.091.79
167
0.17
0.03.1
24.427.8
0.0
25.08.4
5.9
1.8
2.3
0.3
341
78
49.81.15
16.18.620.117.04
12.372.720.230.14
98.30
6 4245
68
18
2
18627
87
322
511
2< l
29
790.310.130.251.001.42
9710.12
0.01.4
23.431.7
0.0
24.39.5
4.81.5
2.2
0.3
aNormative values based on Fe2θ3/FeO ratio of 0.15; Fe 2 θ3 is total iron as Fe2C>3.represents ratio of chondrite-normalized values.
442-1 exhibits a large amount of within-unit variation:Cr ranges from 193 to over 300 ppm and one lithologicsub-unit (442-IK) contains 412 ppm Cr and 201 ppm Ni.Zr varies between 67 and 106 ppm, and the Fe/Mg ratiovaries from 0.97 to 2.41. The singularly high Ni, Cr, andnormative 01 contents (Figures 3,5, and 6) and the lowFe/Mg ratio and Zr levels in lithologic sub-unit 442-IK(sample 442-8-1, 32 cm) may be due to the developmentof a within-flow cumulus assemblage. The rock at thisinterval has a doleritic appearance, and there is evidenceof a concentration of crystals where the sample was ta-ken.
Geochemical units 442-2 and 442-4 have significantlyhigher Zr contents and Fe/Mg ratios, and complemen-tary lower contents of Ni and Cr, than 442-1. Geo-chemical units 442-3 and 442-4 have very low Cr (35-64ppm) and Ni (32-63 ppm) levels, and a high Fe/Mg ratio(1.47-2.32), indicative of considerable fractionation ofolivine, pyroxene, and (or) chrome spinel prior to erup-tion and emplacement of the magma. Indeed, none ofthe geochemical units of Site 442 has the high MgO(>9.0%) and Ni (>200 ppm) indicative of the"primary" basaltic liquid of Kay et al. (1970), apart
810
GEOCHEMISTRY OF BASALTS
from 442-8-1, 32 cm, which as we have seen could easilybe a cumulus assemblage derived by within-flow dif-ferentiation.
The chemical data for Site 442 may be convenientlyplotted against an index of fractionation to illustratevariations within and between the four distinct geo-chemical units. The Fe/Mg ratio is one such index thathas been used in the past, but the strong dependence ofthis ratio on the composition of the fractionating phases,rather than their quantity, makes it less desirable as afractionation index than a suitable incompatible traceelement. Incompatible or hygromagmatophile elementsare those with bulk-distribution coefficients (D) signif-icantly less than unity; ideally, D should approach zero.Hygromagmatophile (HYG) elements may be groupedaccording to their relative bulk-distribution coefficientsinto more-HYG (D 0.01) and less-HYG types (D ~0.1); in ocean-ridge basalts, the more-HYG elements in-clude Rb, K, Th, La, Ce, Nb, Ta (Zr, Hf); and the less-HYG elements include (Zr, Hf), Sm, Ti, P, Y, and theheavy rare-earth elements (REE) (Wood et al, 1979).Zirconium is thus considered to have a low D value inocean-ridge basalts, and probably back-arc basinbasalts also (Saunders and Tarney, 1979; Weaver et al.,1979), and in consideration also of its apparent im-mobility during the hydrous alteration of basalts (Cann,1970; Pearce and Cann, 1971), it therefore provides asuitable index of fractionation for the study of basalticsuites (cf. Weaver et al., 1972).
Figures 5 to 14 present a visual summary of the data,using zirconium as an index of fractionation. Figure 15shows down-hole variation in Zr, P2O5, and K2O.Although there is only a moderate range of Zr contentsin the Site 442 basalts, the salient points may be sum-marized as follows. TiO2, Ce, Nb, and Y (Figure 16) ex-hibit strong colinearity with Zr, and the ratios do notchange significantly from unit to unit. By inference, itmay be expected that the other hygromagmatophile ele-ments P, K, Rb, and Ba would also show a colinear rela-tionship with Zr, but in fact this is not the case (Figures10 to 14). P2O5 varies independently of Zr (and of theother incompatible elements) especially in geochemicalunits 442-2, 442-3, and 442-4 (Figure 10). This is clearlyillustrated in Figure 15, where P2O5 is shown to vary on-ly slightly, and colinearly with Zr, in unit 442-1, butshows erratic behavior in the other units. We suggestthat this variation is due to alteration of the pillowbasalts, possibly at very low temperatures (<25°C).K,Rb, Ba, and Sr show similar erratic behavior, which isprobably also due to alteration by sea water: the mobili-ty of these elements in oceanic basalts is well docu-mented (Hart, 1971; Hart et al., 1974).
Included on Table 4 are representative element ratiosof basalts being erupted at a present-day ocean ridge,and each biaxial plot in this paper contains ratio lines toaid comparison with Leg 58 basalts. Three distinct typesare presented for comparative purposes.
Normal, or N-type ocean ridge basalts, which are theso-called mid-ocean-ridge basalts (MORB) now beingerupted along most ocean ridges, have low initial 87Sr/
ratios (0.7023-0.7027; Hart, 1971), have low abun-
dances of many HYG elements (K, Rb, Ba, Th, U, lightREE, Nb and Ta; White and Schilling, 1978; Bougaultet al., 1979), and are depleted in the more-HYG ele-ments relative to chondrites (Schilling, 1971; Frey et al.,1974; Wood et al., 1979). Representative analyses ofN-type ocean-ridge basalts were taken from Rhodes etal. (1976).
Enriched, or E-type tholeiitic and alkalic basalts, arepresently being erupted along certain sections of themid-Atlantic Ridge (e.g., the Azores, parts of Iceland,and MAR 45° N). They are characterized by variable butgenerally high abundances of most HYG elements; ra-tios of the more-HYG elements are approximately chon-dritic (Wood et al., 1979), and initial 87Sr/86Sr ratios aremoderately high (0.7035; White et al., 1976).
Transitional or T-type ridge segments lie between thenormal and enriched areas, and include most of Iceland,the Reykjanes Ridge (60-64° N) and the FAMOUS area(MAR 35-39° N). These segments are erupting basaltswith trace-element abundances and isotope chemistryintermediated between N- and E-type basalts. In addi-tion to ratios for ocean-ridge basalts, representativeratios from basalts from the East Scotia Sea (Saundersand Tarney, 1979) and from the Bransfield Strait (Wea-ver et al., 1979) back-arc basins are presented in Table5. The data for the MAR 45° N (E-type) and MAR 63° N(T-type) basalts are from Tarney et al. (1979) and weredetermined using the same XRF spectrometer as the pre-sent Leg 58 samples, as were the Scotia Sea and Brans-field Strait lavas. Inter-laboratory error is thus avoided.
In Zr/Nb ratios (Figure 9), the basalts of Site 442closely resemble N-type ocean-ridge basalts recoveredduring DSDP Leg 34 from the Nazca plate (Rhodes etal., 1976). Ce/Zr and TiO2/Zr ratios are also similar toN-type ocean-ridge basalts, and also closely resemblethose from the East Scotia Sea (Figures 7 and 8). Thus,in relative abundances of Ti, Zr, Nb, and Ce, the Site442 basalts most closely resemble N-type ocean-ridgebasalts.
Full rare-earth-element data are not yet available forthe Site 442 basalts, althoughWood et al. (this volume)have presented some preliminary patterns for Sites 443,444, and 446. However, by using Y as a heavy REE, ap-proximately equivalent to Ho, a plot of Ce versus Y(Figure 16) shows that the Site 442 basalts have chon-drite-normalized Ce/Y ratios (represented as CeN/YN)of about 1 (see also Table 2).
Other chemical criteria frequently used to character-ize and discriminate between different basalt types arethe relative abundances of K, Rb, U, Th, and Ba. Un-fortunately, apart from Th and Ba, these elements areeasily mobilized during alteration by sea water (Bass etal., 1973; Hart et al., 1974; Mitchell and Aumento,1977) and from Figures 11, 12, and 13 it is apparent thatalteration precludes their use in the present study.However, Ba levels in the Site 442 basalts are low(generally less than 40 ppm), giving the low Ba/Zr ratios(Figure 13) typical of N-type ocean-ridge basalts. Woodet al. (this volume) have also shown that the Th/Hfratios of the Site 442 basalts are similar to those ofN-type basalts.
811
N. G. MARSH, A. D. SAUNDERS, J. TARNEY, H. J. B. DICK
TABLE 3Analyses, Element Ratios, and CIPW Norms for Basalts from Hole 442B a
CoreSectionInterval (cm)
Siθ2 (%)TiC 2AI2O3Fe2θ3MnOMgOCaONa2θK 2 0P2O5Total
Ni (ppm)CrZnGaRbSrYZrNbBaLaCePbTh
Zr/NbTi/ZrY/ZrCe/ZrBa/ZrCeN/YN
Fetot/MgK/RbBa/Sr
QOrAbAnNcDiMy01Mt11Ap
33
40
50.21.38
15.99.230.176.08
12.152.720.360.18
98.44
44193
7620
81923298< l40
414
32
840.330.140.411.071.76
3700.21
0.02.1
23.430.7
0.024.113.9
0.21.62.70.4
33
140
50.01.23
16.59.070.136.63
12.492.760.290.15
99.19
57247
8318
5186
2989
124
216
7< l
89830.330.180.271.361.59
4760.13
0.01.7
23.532.0
0.024.1
8.45.11.62.40.4
34
13
50.01.27
15.99.980.146.39
11.692.910.330.15
98.83
52221
85187
1803096
137
512
11
9679
0.310.130.390.981.81
3940.21
0.02.0
24.929.80.0
22.910.44.61.82.40.4
41
57
49.51.18
16.58.870.156.00
12.822.880.310.17
98.40
60258
7018
7189
3089
422
31762
22790.340.190.251.391.71
3710.12
0.01.9
24.831.6
0.026.1
4.46.21.62.30.4
42
110
50.21.19
16.08.910.116.46
11.983.030.330.14
98.33
42222
6620
8187
2985
328
3123
< l
2884
0.340.140.331.021.60
3440.15
0.02.0
26.129.6
0.024.5
7.55.31.62.30.3
43
21
49.51.18
15.99.170.126.64
12.423.270.290.14
98.58
53216
67176
1842890
226
214
4< l
4579
0.310.160.291.231.60
4050.14
0.01.8
26.528.2
0.927.3
0.010.4
1.62.30.3
51
120
50.01.29
16.08.740.117.49
11.602.890.250.14
98.54
84222
6718
3188
2589<l33
513
< l2
870.280.150.371.281.35
6950.18
0.01.5
24.830.5
0.021.910.2
5.91.52.50.3
521
49.91.23
15.98.480.107.63
11.432.970.250.14
98.00
103237
7019
2183
2391
132
212
< l<l
9181
0.250.130.351.281.29
10210.17
0.01.5
25.630.00.0
21.99.36.81.52.40.3
52
60
50.21.26
15.98.660.107.93
11.112.800.220.13
98.37
157232
8419
2179
2791
435
211
1< l
2383
0.300.120.381.001.27
8970.20
0.01.3
24.130.7
0.019.815.2
3.81.52.40.3
52
106
49.81.15
15.88.390.117.32
12.302.790.340.13
98.16
80244
59155
1762684
330
211
1<l
2882
0.310.130.361.041.33
5660.17
0.02.1
24.030.2
0.025.3
7.06.71.52.20.3
53
94
49.61.18
15.68.590.126.71
12.842.780.310.14
97.86
74226100
155
1792790
332
3941
3079
0.300.100.360.821.48
5230.18
0.01.9
24.029.7
0.028.0
5.55.81.52.30.3
54
79
49.61.18
16.09.010.126.54
12.802.700.230.14
98.37
60233
7914
3185
2884
229
316
32
4284
0.330.190.351.401.60
6360.16
0.01.4
23.231.4
0.026.3
8.14.51.62.30.3
The colinearity exhibited by the hygromagmatophileelements (excluding those which have been subject tomobilization during alteration by sea water) in the Site442 basalts suggests that the magmas giving rise to thefour geochemical units were derived from similar paren-tal magmas, or by fractional crystallization of a com-mon parental magma. However, unit 442-1 could nothave been derived from the same magma that producedthe other units, because it is separated from them by 2meters of mudstone. This could have taken as long as 2m.y. to have been deposited (Site 442 report, this vol-ume). In addition, units 442-3 and 442-2 are separatedby a magnetic reversal, which suggests that these two
units may not be comagmatic, although they are prob-ably cogenetic.
Site 443
Site 443 (Figure 1) lies approximately 95 km east ofthe axial zone of the Shikoku Basin, at 29° 19.65'N,137°26.43 'E, at a water depth of 4372 meters. Approx-imately 116 meters of basaltic flows and 460 meters ofsediment were drilled in the single hole. The site is on apositive magnetic lineation tentatively identified asanomaly 6A (21-22 m.y.; LaBrecque et al., 1977; Koba-yashi and Nakata, 1977). The oldest microfossils recov-ered at this site indicate an age of 14 to 17 m.y., but it is
812
GEOCHEMISTRY OF BASALTS
TABLE 3 - Continued
6153
49.71.19
15.68.980.126.89
12.942.780.250.15
98.63
6423775182
13028883293103
<l
29810.320.110.330.881.51
10420.22
0.01.5
23.829.70.0
28.05.56.41.62.30.4
6232
49.81.16
15.69.050.136.9712.882.730.260.15
98.73
71277106187
1522786237388
<l
43810.310.090.430.731.51
3110.24
0.01.6
23.429.80.027.76.66.01.62.20.4
62
147
49.81.19
15.88.490.118.19
10.972.720.200.13
97.64
822477120<l19328834284121
<l
21860.340.140.341.051.20_0.15
0.01.2
23.631.10.019.315.84.21.52.30.3
6328
50.11.24
16.38.030.137.14
11.902.990.290.15
98.20
4923276183
195268933241261
30840.290.130.361.131.30
8050.16
0.01.8
25.830.70.0
23.17.76.11.42.40.3
7146
50.31.20
15.79.920.167.94
10.192.820.200.16
98.61
5020975192
20231943444143
<l
31770.330.150.471.111.45
8220.22
0.01.2
24.230.20.016.419.63.21.82.30.4
71
110
50.21.35
15.810.050.167.5810.512.780.170.14
98.68
5921180182
16130934403105
<l
23870.320.110.430.821.54
6930.25
0.01.0
23.830.40.0
17.619.42.11.82.60.3
8132
47.60.9114.99.880.1811.839.382.090.110.12
96.96
2014125919<l13521676145131
<l
11820.310.190.211.520.97_0.10
0.00.618.231.80.0
12.419.612.61.81.80.3
8574
49.41.16
15.59.510.158.88
10.342.640.110.13
97.86
10025558162
15928841293134
<l
84830.330.150.351.141.24
4690.18
0.00.7
22.830.90.016.817.95.81.72.30.3
86
141
49.51.24
15.29.380.139.2810.132.700.150.16
97.84
1052586318<l169279113431252
91820.300.130.371.091.17—0.20
0.00.9
23.329.60.0
16.817.56.61.72.40.4
87
104
50.21.21
16.08.900.137.43
10.733.240.370.14
98.38
6622841175
15420944192136
<l
23770.210.140.201.601.39
6180.12
0.02.2
27.928.60.020.07.68.71.62.30.3
9134
49.31.14
16.18.310.147.7810.972.590.260.12
96.74
137300591811167247843551451
19880.310.180.451.431.24
1920.21
0.01.6
22.732.70.018.416.83.11.52.20.3
91
107
49.11.08
14.89.170.137.6412.502.660.520.13
97.80
11828261149
16425796352104
<l
13820.320.130.440.981.39
4780.21
0.03.1
23.027.60.0
28.52.9
10.01.62.10.3
9237
49.21.13
15.38.420.126.7313.982.720.330.14
98.07
10027085177
173288445221262
21810.330.140.621.051.45
3960.30
0.02.0
22.529.20.5
33.10.08.01.52.20.3
9354
50.21.47
15.011.080.147.8510.032.970.300.17
99.21
11825199203
1773110623132031
53830.290.190.291.581.64
8300.18
0.01.8
25.326.90.018.116.25.51.92.80.4
possible that some older sediments were not cored (Site443 report, this volume). Sites 442 and 443 lie on thesame spreading line, on opposite sides of the spreadingaxis of the basins.
The basalts recovered at Site 443 appear to be frommassive flows or sills (Figure 17). Essentially all thebasalts are olivine-bearing, and in general vesicles arescarce (although in lithologic unit 3 they reach 30%).On the basis of shipboard petrography, the basalt se-quence has been divided into six lithologic units. Unit 1consists of one small piece of glassy basalt immediatelyunderlying the sediments in Core 49. Unit 2 compriseseight sub-units, each of which is considered to be an in-
dividual, massive basalt flow. The various sub-unitshave similar mineralogies and textures, containing botholivine and plagioclase phenocrysts, having less than 5per cent vesicles, and ranging in texture from cryp-tocrystalline at chilled margins to intergranular andeven subophitic at the centers of the thickest flows. Themost massive sub-unit, 2E, shows accumulation of oli-vine in its middle and lower sections, indicative of grav-itational differentiation.
Unit 3 has been divided into nine sub-units, the sixuppermost comprising three pillow-basalt sub-units(3A, 3C, 3E) each with an underlying massive-basaltsub-unit (3B, 3D, 3F) of similar lithology. The pillow
813
N. G. MARSH, A. D. SAUNDERS, J. TARNEY, H. J. B. DICK
TABLE 3 - Continued
CoreSectionInterval (cm)
101
110
111
146
112
70
113
24
121
52
131
34
132
92
133
19
151
40
161
24
161
47
161
111
SiO2 (%)TiO2
AI2O3Fe2θ3MnOMgOCaONa2OK2O
P2O5Total
Ni (ppm)CrZnGaRbSrYZrNbBaLaCePbTh
Zr/NbTi/ZrY/ZrCe/ZrBa/ZrCeN/YN
Fetot/MgK/RbBa/Sr
QOrAbAnNeDiHy01Mt11Ap
49.91.62
16.010.94
0.175.26
12.182.890.340.21
99.52
50150
9420
7162
35105
325
613
51
3592
0.330.120.240.912.41
4040.15
0.02.0
24.629.8
0.024.410.2
2.51.93.10.5
48.81.60
16.411.890.194.52
12.432.660.410.29
99.18
58192108
239
18538
1143
306
153
<l
38840.330.130.260.973.05
3790.16
0.02.4
22.731.9
0.023.7
9.52.92.13.10.7
49.81.55
16.110.340.165.89
11.972.830.330.22
99.18
70195105
215
17537
1164
244
1871
2980
0.320.160.211.192.04
5530.14
0.02.0
24.130.5
0.022.911.1
3.11.83.00.5
49.81.59
16.310.310.185.56
12.172.860.330.23
99.31
71207110
215
17839
1155
394
144
< l
23830.340.120.340.882.15
5410.22
0.01.9
24.430.9
0.023.410.1
3.01.83.00.5
49.11.60
16.311.480.194.66
12.122.590.500.32
98.85
67194
9924
8181
40120
336
417
2< l
4080
0.330.140.301.042.86
5210.20
0.03.0
22.231.6
0.022.513.4
0.42.03.10.8
49.91.74
16.110.940.195.60
12.012.750.360.25
99.82
66201100
206
17143
1245
286
196
< l
25840.350.150.231.092.27
4950.16
0.02.1
23.330.6
0.022.713.0
1.51.93.30.6
49.21.66
16.111.450.164.66
12.452.730.390.33
99.16
48197103
216
18641
1144
406
1832
2887
0.360.160.351.082.85
5400.22
0.02.3
23.330.8
0.024.410.8
1.42.03.20.8
49.51.58
16.110.790.145.20
12.272.790.400.35
99.11
51195110
227
18239
1154
318
1821
29820.340.160.271.132.41
4730.17
0.02.4
23.830.6
0.023.610.6
2.31.93.00.8
49.71.23
16.39.880.165.52
12.652.450.390.40
98.73
32498321
7219
3188
333
715
3< l
2984
0.350.170.381.192.08
4640.15
0.42.3
21.032.8
0.023.214.3
0.01.72.40.9
50.01.31
16.710.670.165.58
11.702.830.480.19
99.55
3935882221
2083395
245
517
32
4883
0.350.180.471.272.22
1880.22
0.02.8
24.131.5
0.021.110.7
4.01.92.50.5
49.71.23
16.69.830.175.51
12.672.390.380.22
98.67
42527919
8210
3388
231
413
4< l
4484
0.380.150.350.972.07
3960.15
0.32.3
20.533.9
0.023.414.1
0.01.72.40.5
50.21.13
16.38.610.126.80
12.172.580.300.14
98.31
59508220
4200
2781
330
416
54
27840.330.200.371.461.47
6230.15
0.01.8
22.232.5
0.022.814.7
1.21.52.20.3
basalts are sparsely olivine- and plagioclase-phyric, andchrome spinel occurs as an accessory phase. Vesicles arecommon to abundant (15-30% ). The flows are severalmeters thick, and the thickest again shows subophitictextures and olivine accumulation in its middle andlower sections. The bottom three sub-units of unit 3 aresparsely phyric massive flows or sills, in lithologysimilar to sub-units 3A to 3F, although they do contain2 to 3 per cent resorbed plagioclase phenocrysts.
Unit 4 is a single 15-m-thick flow or sill consisting ofintersertal to subophitic basalt. Plagioclase and olivinephenocrysts are common, and pilotaxitic alignment ofgroundmass pyroxene and feldspar is observed. Unit 5,on the other hand, consists of five flows which are either
aphyric or sparsely plagioclase-phyric. Again, however,vesicles are scarce. The last unit drilled, unit 6, com-prises two olivine- and plagioclase-phyric basalt sub-units, both of which could be flows or sills. Only the topfew centimeters of subunit 6B were recovered.
The Site 443 basalts are generally fine to mediumgrained with intersertal to intergranular textures. Por-phyritic varieties contain 1 to 10 per cent plagioclasephenocrysts (An70 or greater), and 2 to 5 per cent olivinephenocrysts (probably Fo80_9o) Due to alteration, thepresence of groundmass olivine is uncertain. Plagioclase(Ang<µyo) and augite are the two major groundmassphases, with about 2 to 5 per cent interstitial magnetite.Finer-grained sections have a groundmass containing
814
GEOCHEMISTRY OF BASALTS
TABLE 3 - Continued
162
15
171
14
171
64
171
142
181
46
191
45
192
36
192
51
201
50
49.41.26
15.89.390.136.26
11.312.610.420.15
96.71
3942
10920
8204
3093< l34
3184
<l
810.320.190.371.471.74
4400.17
0.02.6
22.831.2
0.021.116.4
0.51.72.50.4
49.61.24
16.510.15
0.155.54
12.662.370.370.21
98.79
43577820
6211
3394
224
416
52
4779
0.350.170.261.192.12
5060.11
0.22.2
20.333.8
0.023.414.5
0.01.82.40.5
aNormative values
50.11.21
17.19.490.155.38
12.952.700.480.19
99.78
5058
14920
8210
3187
339
515
4< l
2983
0.360.170.451.192.05
4950.19
0.02.8
22.933.2
0.024.7
7.63.61.72.30.4
based on
50.21.65
15.511.09
0.175.75
11.952.680.330.25
99.61
47599420
6185
37121
330
617
52
4082
0.310.140.251.132.24
4570.16
0.22.0
22.829.5
0.023.415.5
0.01.93.10.6
FeoO^/F<
50.21.63
16.010.210.165.53
12.312.820.290.26
99.42
48598822
3189
39123
131
516
1<l
12379
0.320.130.251.012.14
8000.16
0.01.7
24.030.4
0.024.213.0
0.41.83.10.6
:O ratio
50.71.65
15.810.330.156.05
11.413.240.320.18
99.81
5958
10622
5173
36123
542
317
1< l
2580
0.290.140.341.161.98
5250.24
0.01.9
27.527.7
0.022.9
9.44.41.83.10.4
of 0.15: F<
50.41.65
16.410.110.165.38
12.263.080.320.21
99.92
57598122
7189
38129
434
419
3< l
3277
0.290.150.261.232.18
3810.18
0.01.9
26.129.9
0.024.4
8.23.31.83.10.5
50.21.59
16.210.380.195.46
12.133.130.380.20
99.86
6360842414
18638
1233
344
1573
4177
0.310.120.280.972.20
2250.18
0.02.2
26.529.0
0.024.7
6.15.31.83.00.5
:oθ^ is total iron as
50.11.62
15.910.960.175.47
12.042.760.470.22
99.70
4164792010
17939
1241
296
19< l< l
12478
0.310.150.231.202.32
3870.16
0.02.8
23.429.8
0.023.712.1
1.81.93.10.5
23Ce^/YN represents ratio of chondrite-normalized values.
plagioclase and pyroxene microlites, with granular orskeletal magnetite set in a cryptocrystalline matrix,often partially altered to clays. Chrome spinel is an ac-cessory phase in units 3 and 4.
Veins and fractures lined by calcite and clays arecommon throughout the recovered section, as arecarbonate- and clay-filled amygdules. In unit 4 (Core60), native copper occurs in a carbonate vein togetherwith well-formed pyrite octahedra. Other veins in thisunit are lined by a talc-like mineral which has alsoreplaced the olivine phenocrysts. Olivine in other unitsis invariably partially or totally replaced by pseudo-
morphs of clay and chloritic material, with occasionalzeolites or calcite. In unit 2, the cores of some largeplagioclase phenocrysts are saussuritized.
The basalts recovered at Site 443 (Table 6) areolivine- and hypersthene-normative sub-alkaline tholei-ites; no nepheline- or quartz-normative varieties havebeen found (Figure 3). On the basis of their chemistry,the basalts have been divided into five geochemical units(Figure 17): units 443-1 and 443-2 correspond to litho-logic unit 2; geochemical unit 443-3 corresponds to mostof lithologic unit 3; geochemical unit 443-4 consists ofthe lower part of lithologic unit 3, and lithologic units 4
815
N. G. MARSH, A. D. SAUNDERS, J. TARNEY, H. J. B. DICK
Site 443
Figure 3. Normative nepheline (Ne), olivine (Ol), diopside (Di), hypersthene (Hy),quartz (Q) tetrahedra for basalts from Sites 442 and 443. Normative values basedon Fe2O3/FeO ratio of 0.15.
Site 443
20
10
90 80
Figure 4. Normative plagioclase, pyroxene, olivine diagrams for basaltsfrom Sites 442 and 443. The Fo-An cotectic (C) is from Shido et al.(1971).
and 5; and geochemical unit 443-5 corresponds to litho-logic unit 6. Geochemical unit 443-4 cannot be disting-uished from 443-5 using the present XRF data, butbecause they are petrographically distinct and becausethe two units have different Th/Hf ratios (Wood et al.,this volume) we have retained the division for the pur-pose of the present discussion.
On the diagram of normative plagioclase-pyroxene-olivine, the basalts range in composition from olivinetholeiites (443-3, 443-4 and 443-5), through olivine-pla-gioclase tholeiites straddling the 1-atmosphere Fo-Ancotectic (443-2), to plagioclase tholeiites (443-1)(Figure4). This division broadly corresponds with the observedphenocryst assemblages, although many of the basaltsfrom Site 443 are only sparsely phyric or even aphyric.
Thus, the phenocryst assemblage of lithologic unit 2 isdominated by plagioclase, whereas in the lower litho-logic units, olivine phenocrysts are more abundant.
As with the Site 442 basalts, there is considerabledown-hole variation in chemistry at Site 443 (Figure 17).From the Fe/Mg ratio and Zr, Cr, and Ni abundances(Figures 17 and 18), it is evident that geochemical units443-4 and 443-5 contain the most-primitive basalts (withthe lowest Fe/Mg ratio and Zr, and highest Ni and Cr)and that units 443-1 and 443-3 contain the most-evolvedbasalts recovered at this site. Unit 443-2 occupies an in-termediate position. However, there is considerablewithin-unit variation. Thus, unit 443-3 basalts contain54 to 156 ppm Ni, one sample (Core 58-3, 29 cm) con-taining 405 ppm Ni, 428 ppm Cr, and 14.01 per cent
816
GEOCHEMISTRY OF BASALTS
200 -
3 100 -
2.0
Zr (ppm)
Figure 5. Ni versus Zr, Site 442. The numerals indicategeochemical units in down-hole order (see Figure 2).
' ' ' ' 1 I
Site 442
• 442-1A 442-2x 442-3O 442-4
-
l,
l,
' . ' | i i i i
•
-
V; :
A
K. x αS>oX XX —
0 50 100 150Zr (ppm)
Figure 6. Cr versus Zr, Site 442.
MgO. Ni and Cr also show strong variations in Units443-1, 443-4, and 443-5. Whereas fractional crystalliza-tion (involving removal of pyroxene, olivine, and (or)chrome spinel from the melt) may be responsible forsome of the observed variations in Ni and Cr, it is evi-dent that some within-flow differentiation has also oc-curred, particularly in the thickest flows of lithologicunit 3. Thus, the high Ni and Cr levels observed in 58-3,29 cm are probably due to the presence of cumulus oli-vine and chrome spinel. However, the olivine accumula-
1.0 -
1 1 1 1—
• Site 442
• 442-1A 442-2x 442-3O 442-4
~S *
-I 1 1 1 1 1 1 r
X
A
•• *
•
i i i i i i i i
/
A
A
i
' ' '
d>
i i
<ß× -o
-
1 P
100
Zr (ppm)
Figure 7. TiO2 versus Zr, Site 442.
15 -
* Site 442
• 442-1A 442-2x 442-3O 442-4
-
-
•
X «XXX•
• x*
m» •«
xX
X
•• A
1••
i 1 i
àAà.
A
A
<A O
AOO
O
O J>
* ^ -
-
-
100Zr (ppm)
150
Figure 8. Ce versus Zr, Site 442.
10 I I , .
- Site 442
• 442-1- A 442-2
x 442-3O 442-4 *
_
-
-
-―
— i 1-
•
×
1
—i 1—
/
/
/
• x#_•_•_•
~1—1
••XX
A
•
i
A
AA4
A
CA
0 0
* Site 443 ^
|l/ _
/O /
x^_ • * '
-
50 100Zr (ppm)
150
Figure 9. Nb versus Zr, Sites 442 and 443. Individualdata points for Site 443 omitted for clarity.
tion which was observed petrographically in unit 2E hasnot been geochemically confirmed. In general, it is ap-parent that the basalts from Site 443 have undergone
817
N. G. MARSH, A. D. SAUNDERS, J. TARNEY, H. J. B. DICK
1 1 1—i—
Site 442
ΦA
X
o
_
442-1442-2442-3442-4
^ Φ
X
X
ΦΦΦΦ
x m mm• ΦΦ Φ
Φ _,
1
X
X
Φ
ΦX ΦΦ>
1 '
A
Φ
A
A
A
A
A
,/-
y^ —
O A 9^-^^
oo
° ,,6/ivi L-^—
o —
-
-
100Zr (ppm)
Figure 10. P2O5 versus Zr, Site 442. The scatter of datapoints along the P2O5 axis in geochemical units 2, 3and 4 appears to be a function of alteration of thesebasalts. See also Figure 15.
0.5 -
- Site 442
Φ 442-1~ A 442-2
X 442-3O 442-4
-y×^^—
Φ
Φ
*
X
<
Φ
Φ
X
X
Φ
' • Φ
1
X
X
•Φ
Φ•
A AA
Φ
_
O ^-"^
A o
°o oo -
-
-
50 100Zr (ppm)
Figure 11. K2O versus Zr, Site 442. Scatter of datapoints is indicative of sea-water alteration.
20
10
n
Site 442
-
AX
O
442-1442-2442-3442-4
Φ
ΦΦXX
ΦΦXΦ
ΦX
ΦΦ
Φ«Φ
ΦΦ Φ
1 1
x
x ΦΦ
XΦ
. 1
A
Φ
A
A
A
A t
-
-
-
o
o
AO
O A
o
o
100Zr (ppm)
150
Figure 12. Rb versus Zr, Site 442. Scatter of data pointsalong Rb axis in geochemical units 3 and 4 is in-dicative of sea-water alteration.
Site 442
ΦA
X
o
-
442-1442-2442-3442-4
Φ
Φ
X
ΦΦ t
V. *•
Φ Φ
y
Φ
I
Φ
A
' ' ' ' ' \ ^
—
O
A
o ot °°2 ja£^°2^-A
50 100Zr (ppm)
Figure 13. Ba versus Zr, Site 442. As in figures 11 and12, scatter of data points may be a result of sea-wateralteration.
200
150
100
1
Site 442
ΦA
~ X
o
_
442-1442-2442-3442-4
Φ
//
/
1
X
ΦΦ
1 1 1
X
y
x× xX
Φ
Φ Φ
Φ>*
Φ *
Φ
Φ
I '
A
Imn
-1 T—7
/
A
1 '
k oA
i i i i
;
"V -o
•
-
-
-
I 1 1 1
150
Zr (ppm)
Figure 14. Sr versus Zr, Site 442.
considerable fractionation prior to their eruption: theirNi and Cr contents are too low, and Fe/Mg ratios toohigh, for the liquids to have been in equilibrium withupper-mantle olivine compositions (Fo88_90; Green,1970; O'Hara, 1973).
Using Zr as an index of fractionation, hygromagmat-ophile trace elements have been plotted in Figures 19 to25. TiO2, P2O5, Ce, La, Nb, and Y all show strong posi-tive correlations with Zr, similar to the Site 442 data. Ingeneral, the ratios of these elements between the dif-ferent geochemical units are constant, particularly forthe more-HYG elements (Ce, La, Nb, and Zr). The plotof Zr versus Y (Figure 20), however, suggests that geo-chemical units 443-2 and 443-5 have slightly higherY/Zr ratios than units 443-3 and 443-1; unit 443-4shows a range of Y/Zr ratios covering both extremes. Astrong correlation exists between P2O5 and Zr (Figure21): the incoherent scatter of P2O5 in lithologic unit 2 atSite 442 is not observed at Site 443.
818
GEOCHEMISTRY OF BASALTS
Core P2O5 (wt. %)
0.1 0.2 0.3 0.4
K2O (wt. %)
0.1 0.2 0.3 0.4 0.5
• 442-1A 442-2x 442-3O 442-4 Λ•
•
60—i r
Zr (ppm)
80 100
\ .
J I I I I L
Figure 15. Down-hole variation at Site 442 in the abundancies ofP2O5, K2O, and Zr.P2O5 and Zr show a broad correlation in unit 442-1 but the distribution ofP2O5 isincoherent in units 2, 3 and 4. K2O also exhibits considerable scatter, probably in-dicative of sea-water alteration.
1 ' ' ' 1
- Site 442
• 442-1A 442-2
1 x 442-3O 442-4
^ — " " "
i • | i
. 1 1 , 1
• xx • A
Y (ppm)
Figure 16. Ce versus Y, Site 442. CeN/YN indicateschondrite normalized Ce/Y ratios (approximatelyequivalent to CeN/HoN ratio).
K2O, Rb, and Sr do not show clear correlations withZr (Figures 22 and 23) and, as with the Site 442 basalts,we suspect that this is essentially because of alterationby sea water. The least-altered samples show low K/Zr,Ba/Zr, and Rb/Zr ratios, similar to those of N-typeocean-ridge basalts (Table 4; Erlank and Kable, 1976).
CeN/YN ratios for the basalts of Site 443 are, withinthe error of XRF analysis at such low levels of Ce, con-stant throughout the five units, at about 1 (Figure 26).This is in agreement with preliminary, more-compre-hensive REE data which indicate that the CeN/YbN ratiois approximately 1.5 in unit 443-3, and 1.2 in unit 443-4(Wood et al., this volume; unpublished data of theauthors). The CeN/YN ratios for the Site 443 and 442basalts are practically indistinguishable, and are similarto the low values reported for N-type ocean-ridge ba-salts (e.g., Kay et al., 1970; Sun et al., 1979). Other
trace-element ratios, in addition to the K/Zr, Ba/Zr,and Rb/Zr ratios already mentioned, indicate closesimilarities between the Site 443, 442, and N-type ocean-ridge basalts. Zr/Nb ratios are high (25-40; Figure 9),Ce/Zr ratios are low (Figure 26), and Y/Zr ratios ap-proach chondritic values (Figure 20). However, neitherthe Site 442 nor Site 443 basalts show the strong deple-tion of all more-HYG elements that is observed inN-type basalts. Thus, CeN/YN is approximately 1, andnot significantly less than 1, as in N-type ocean-ridgebasalts (Sun et al., 1979).
With the exception of the slight variation in Y/Zrratio, the constant ratios of most hygromagmatophileelements among the units of Site 443 indicate that mostof the basalts could be derived from a common basalticparent by varying degrees of low-pressure fractionalcrystallization of olivine, plagioclase, and (or) chromespinel. The rapid decrease in Ni and Cr relative to Zr(Figure 18) supports this suggestion. However, to varythe Y/Zr ratio (and the Th/Hf ratio; Wood et al., thisvolume) requires either variable degrees of partial melt-ing, or open-system crystal fractionation (O'Hara, 1977),in order to retain Y in the refractory mineral assem-blage. We suggest, judging from Y and Zr distributionsin basalts recovered on DSDP Leg 49 in the NorthAtlantic Ocean (Tarney et al., 1978), that units 443-1and 443-3 represent magma batches derived by smallerdegrees of partial melting of the mantle source, and thatunits 443-2 and 443-5 represent slightly higher degreesof partial melting of the same mantle source. Unit 443-4occupies an intermediate position, and it is apparentthat most of the units subsequently underwent con-
819
N. G. MARSH, A. D. SAUNDERS, J. TARNEY, H. J. B. DICK
TABLE 4Representative Analyses and Element Ratios of Basalts from the Daito Basin and from the Atlantic Oceana
Component
SiO2
TiO2
A12O3
F e 2 O 3
MnOMgOCaONa2OK2OP2O5L.O.I.
Total
NiCrRbSiYZrNbBaLaCe
Fe/MgZr/NbCe/ZrCe N /Y N
K/RbBa/SrP/ZrBa/Zr
446-1
45.83.67
14.710.880.364.80
11.143.162.040.372.53
99.40
291836
68927
20627
3132245
2.6380.224.09
4700.457.81.52
Daito Basin,Alkalic Group
446-4
44.84.9
13.311.850.196.25
12.442.811.5.50.990.48
99.53
456922
108826
18832
3622558
2.2060.315.48
5860.33
23.01.93
446-5
42.44.07
10.312.770.22
12.4311.25
1.490.740.732.88
99.33
2931137
16480
19150
24212
2242
1.1960.285.43
3830.44
21.21.41
446-10a
47.13.72
14.99.760.244.97
11.532.971.950.352.03
99.45
283030
75628
21125
3322448
2.2880.234.21
5390.447.21.57
446-6
49.63.61
13.013.01
0.176.339.882.690.270.341.07
100.02
72141
1367
31194
19121
1842
2.3810
0.223.33
22170.337.70.62
Daito Basin,Tholeiitic Group
446-8
47.84.16
13.114.490.215.82
10.162.500.200.430.94
99.86
3323
2447
37302
28143
3260
2.8911
0.203.98
8220.326.20.47
446-9
46.34.95
13.714.970.235.568.303.071.030.511.11
99.70
2039
45 341
37832
2253470
3.1212
0.194.19
9490.505.90.60
446-14
49.63.17
13.312.5 30.236.19
10.862.560.280.371.09
100.08
94175
11426
33229
241022650
2.3510
0.223.72
2090.247.10.45
1
50.011.37
16.1810.18—7.71
11.332.790.220.130.87
100.79
100200
1124
3390
31239
1.5330
0.100.67
18260.106.30.13
Atlantic Ocean Basalts
2
49.301.74
14.0112.130.198.09
11.652.300.180.220.83
100.64
106317
4103
35110
1370
619
1.7480.171.33
3740.688.70.64
3
47.635.22
12.5416.710.235.619.233.040.830.56n.d.
101.6
156616
44536
29031
1802857
3.4690.203.90
4310.408.40.62
4
46.452.94
16.6611.000.075.30
10.383.751.240.481.32
99.59
97022
693-
220-
3022951
2.41-0.23—
4680.449.51.37
aAnalyses of Daito Basin basalts from Tables 6 and 7. Atlantic Ocean basalts: 1, typical "depleted" N-type ocean-ridge basalt; major ele-ments from Engel et al. (1965); trace elements from Sun et al. (1979) and Bougault et al. (1979), with typical values for Ni and Cr; 2, tran-sitional or "T-type" ocean-ridge basalt, Reykjanes Ridge, Leg 49, Site 409, unit 3 (Wood et al., 1979); 3, "enriched" ferrobasalt, Eldgjá,Iceland, sample ISL79 (Wood et al., in press); 4, "enriched" alkali basalt, Fayal, Azores, sample FA47 (Joron et al., in prep.). F e 2 θ 3 istotal iron as F e 2 θ 3 . Ce^/Y^ represents ratio of chondrite-normalized values.
siderable fractional crystallization (particularly units443-3 and 443-1) prior to eruption.
Site 444
Site 444 (Figure 1) lies approximately 100 km southof Site 443, and along the same depositional, structural,and topographic trends. Located at 28°38.25'N, 137°41.03' E, in 4843 meters of water, Hole 444A drilled twobasaltic units, some 50 meters thick, and 285 meters ofsediments (Figure 27). The sediments above and be-tween the two basaltic units are approximately 15 m.y.old, and Waples (this volume), judging from changes inthe sedimentary organic chemistry profile, has sug-gested that the basalt of unit 1 was emplaced as a sillsome 13 m.y. ago. In addition, the sediment immediate-ly above unit 2 appears to have undergone thermal alter-ation, suggesting sub-sediment emplacement of unit 2also. Hence, it appears that both units were emplaced assills and are not true basement, which, from magnetic-anomaly data, has been tentatively dated as 21 to 22m.y. at this site.
Unit 1 is an aphyric dolerite sill with petrography dis-tinct from the other rocks recovered from the ShikokuBasin. The principal constituents are plagioclase (An60_82)and augite; olivine varies from less than 1 per cent at thetop of the sill to 15 per cent in the middle and lower sec-tions. The groundmass is essentially replaced by cryp-tocrystalline chloritic or clayey matrix. Opaque mineralsinclude magnetite and elongate, saw-toothed aggregatesof ilmenite. Apatite needles are found as occasional in-clusions in plagioclase, and chrome spinel occurs as in-clusions in some relict olivines. Small, green-brown,platey crystals have been tentatively identified as kaer-sutite amphibole (Dick et al., this volume). The inter-val 20-1, 53-81 cm is distinct from the remainder of theunit in containing less pyroxene (which has a strongpink-brown color indicative of a high TiO2 component),more magnetite and ilmenite, and aggregates of kaer-sutite up to 1 cm across. Although the very alterednature of the groundmass in this section suggests thatthe different mineralogy may be due to alteration, late-stage magmatic differentiation may also have producedthe observed mineralogy. In fact, the chemistry sup-
820
GEOCHEMISTRY OF BASALTS
TABLE 5Representative Analyses and Element Ratios of Basalts from the Shikoku, East Scotia Sea,
and Bransfield Strait Back-Arc Basinsa
Component
Siθ2TiO2
A12O3
F e 2 O 3
MnOMgOCaONa2OK 2 OP2O5
Total
NiCrRbSrYZrNbBaLaCe
Fe/MgZr/NbCe/ZrCeN/YN
K/RbBa/SrP/ZrBa/Zr
442-1
50.21.26
15.98.660.107.93
11.112.800.220.13
98.37
157232
2179
2791
435
211
1.27230.121.0
8970.206.20.38
442-3
50.21.13
16.38.610.126.80
12.172.580.300.14
98.31
5950
4200
2781
330
416
1.4727
0.201.46
6230.157.50.37
Shikoku Basin
443-3
50.31.83
15.59.420.107.52
10.693.090.220.18
98.83
73225
3178
40126
750
717
1.4518
0.131.04
5950.286.20.40
443-4
49.71.22
16.39.130.137.95
11.622.710.100.11
98.99
113297
1154
2887
331
48
1.3329
0.090.70
8550.205.50.36
444-1
50.31.94
16.27.960.156.258.593.862.310.39
97.92
26127
16391
38209
31177
1842
1.4870.202.71
11980.458.10.85
444-2
50.21.61
16.110.150.176.87
11.362.820.160.23
99.73
53245
2160
40141
771
820
1.7120
0.141.23
6640.447.10.50
East Scotia Sea
D20.43B
50.71.29
16.68.40.167.67
11.123.150.340.19
99.63
63270
5193
29107
349
_16.1
1.2736
0.151.37
5640.257.80.46
D24.14
53.80.61
14.59.20.177.71
10.801.790.240.08
98.99
42295
4123
1440<l55
-6.45
1.38>40
0.161.13
4980.458.71.38
Bransfield Strait
B138.2
51.91.5
16.29.460.186.11
10.074.070.280.21
99.96
35141
3340
26144
288
822
1.8072
0.152.08
7750.266.40.61
P640.1b
52.90.64
17.77.430.136.14
10.303.540.470.06
99.27
40130
11332
1058< l70
28
1.40>58
0.140.97
3550.214.51.21
aAnalyses of shikoku Basin basalts taken from Tables 2 to 5. Analyses of East Scotia Sea basalts from Saunders andTarney (1979), those for Bransfield Strait basalts from Weaver et al. (1979). F e 2 θ 3 is total iron as F e 2 θ 3 •Cej\[/Y]S[ represents ratio of chondrite-normalized values.
CoreNo.
LithologicUnit
MagneticInclination
-90° 0° +90°
Zr (ppm)
80 120
Ba (ppm)
20 40 60
Cr (ppm)
100 400
Fe/Mg Y/Zr
1 2 3 0.1 0.2 0.3 0.4
Ti/Zrε ~
P60 70 80 90 O
460- 49
Claystone and Volcanicash
Unit 1
4 8 0 -Unit2
Plagioclase—olivinephyric basalts
~ 500-54
520
Unit 3Aphyric olivine-
bearing basalts andpillow basalts I
= 540- Unit4Plagioclase—olivine basalts
5 6 0 -
580-
UnitδAphyric and sparsely
phyric plagioclase and
pillow basalts
- ^ Unit 6 ^ - –Olivine-plagioclase basalts
X
ci
»
V\
•TI
_J k t• ~4rFigure 17. Down-hole variations in lithology, magnetic inclination, and selected geochemical parameters at Site 443.
821
N. G. MARSH, A. D. SAUNDERS, J. TARNEY, H. J. B. DICK
TABLE 6Analyses, Element Ratios, and CIPW Norms for Basalts from Hole 443 a
CoreSectionInterval (cm)
SiO2 (%)TiO2AI2O3Fe2O3MnOMgOCaONa2OK 2 OP2O5Total
Ni (ppm)CrZnGaRbSrYZrNbBaLaCePbTh
Zr/NbTi/ZrY/ZrCe/ZrBa/ZrCeN/YN
Fetot/MgK/RbBa/Sr
QOrAbAnNeDiHy01Mt11Ap
493
127
50.31.55
16.89.290.205.83
12.372.910.180.18
99.62
52219
7322
2162
35116
334
516
1<l
39800.300.140.291.121.85
7470.21
0.01.1
24.732.4
0.023.011.1
1.91.63.00.4
494
90
49.81.56
16.410.580.236.31
11.782.680.150.18
99.67
49216
8222
1155
34112
443
6146
< l
28840.300.130.381.011.94
12530.28
0.00.9
22.832.4
0.020.615.3
1.91.83.00.4
502
14
49.51.56
16.710.410.186.54
11.562.690.170.18
99.50
50198
7922
2151
35121
542
716
3< l
24770.290.130.351.121.85
7010.28
0.01.0
22.933.1
0.019.214.6
3.11.83.00.4
503
80
50.01.60
16.49.640.255.76
12.732.780.160.18
99.53
54224
7721
1161
37120
535
516
7< l
24800.310.130.291.061.94
13030.22
0.00.9
23.631.9
0.024.911.2
1.41.73.10.4
504
74
49.91.56
16.410.840.176.30
11.582.790.200.18
99.90
56231
9419
3152
35115
440
716
6<l
29810.300.140.351.122.00
5530.26
0.01.2
23.631.5
0.020.414.1
2.91.93.00.4
5110
49.71.58
14.710.160.137.77
11.422.650.440.18
98.73
107249
76186
16438
1167
587
144
< l
17820.330.120.500.901.52
6120.35
0.02.6
22.727.3
0.023.512.2
5.41.83.00.4
5212
50.11.58
16.59.130.355.77
12.792.860.160.18
99.40
60244488
192
16537
1143
384
136
< l
38830.320.110.330.861.84
6680.23
0.01.0
24.331.8
0.025.3
9.62.11.63.00.4
521
90
49.41.32
15.99.570.147.09
12.432.760.270.14
98.98
99249
9419
3168
3092
322
313
1<l
31860.330.140.241.061.57
7420.13
0.01.6
23.630.5
0.025.2
5.78.01.72.50.3
522
120
49.91.36
16.310.170.167.11
12.112.650.070.14
99.94
94263482
17<l
1643098
117
2124
<l
98830.310.120.170.981.66-0.10
0.00.4
22.432.4
0.021.913.04.31.82.60.3
524
16
49.81.35
16.410.000.167.04
12.042.700.080.15
99.75
86256
7617
< l163
33101
327
31351
34800.330.130.270.971.65-0.17
0.00.5
22.932.4
0.021.712.54.51.72.60.3
532
117
50.31.47
16.49.430.146.30
12.232.900.210.15
99.62
107263
9819
2171
35107
549
4154
< l
21820.330.140.461.051.74
8800.29
0.01.3
24.631.4
0.023.410.5
3.31.62.80.4
533
64
49.91.40
16.29.930.146.94
12.012.800.140.14
99.53
102263131
181
16534
1014
235
1352
25830.340.130.230.941.66
11540.14
0.00.8
23.831.3
0.022.610.7
5.11.72.70.3
ports the latter explanation, because this section hashigher contents of TiO2, P2O5, Zr, Ce, and Y than theremainder of lithologic unit 1.
Lithologic unit 2 consists of plagioclase-phyric basalt(10-15% plagioclase phenocrysts, composition aboutAn70), grading from aphanitic in the chill zone at its up-per margin to subophitic at its center. The uppermost 2meters are amygdaloidal, with 5 to 20 per cent clay- andzeolite-filled vesicles. Relict olivine phenocrysts are seenonly in the lower parts of the unit, and rarely exceed 10per cent of the rock. Augite and magnetite occur throughoutthe unit, and in the coarsest sections the augite ophiti-cally encloses plagioclase laths. Unit 2 appears to bequite distinct from unit 1, the former not containingsignificant ilmenite or apatite and no kaersutite ortitaniferous augite.
Major- and trace-element analyses of basalts recov-ered at Site 444 are given in Table 7. Geochemically, the
two lithologic units 1 and 2 are quite distinct; they aretherefore treated as two separate geochemical units (Fig-ure 27). The basalts of geochemical unit 1 are nepheline-normative and with a high normative olivine/diopsideratio (Figure 28), which is a function of their low CaOcontents. Geochemical unit 2, however, comprises sub-alkaline hypersthene- and olivine-normative tholeiiteswhich plot mainly in the plagioclase-tholeiite field on thenormative plagioclase-olivine-pyroxene diagram (Figure29). This is in agreement with the observed phenocrystassemblage (10-15% plagioclase phenocrysts, An70); thetwo samples which plot in the olivine-tholeiite field (Core23-2, 84 cm, and Core 27-5, 68 cm) both contain alteredgroundmass olivine. Core 27-5, 68 cm also contains rel-ict olivine phenocrysts, possibly in part of cumulusorigin.
Geochemical unit 444-1 comprises basalts which havelow absolute abundances of SiO2 and CaO compared
822
GEOCHEMISTRY OF BASALTS
TABLE 6 - Continued
542
44
5434
544
106
545
31
54636
547
71
551
81
551
104
551
145
552
115
561
112
562
103
563
115
571
129
49.51.32
16.310.090.167.4911.862.630.080.14
99.53
1122549321
116130953
2541333
32830.320.140.261.061.56
6640.16
0.00.522.432.60.0
20.912.26.01.82.50.3
49.61.29
16.310.050.157.6511.742.740.080.15
99.76
1172536419<l16031943
266144
<l
31820.330.150.281.111.52_0.16
0.00.523.232.10.020.710.87.3
1.82.50.3
49.51.31
16.310.180.157.6111.762.700.110.14
99.75
110275107172
15830953
182
1261
32830.320.130.190.981.55
4400.11
0.00.622.932.00.0
20.910.97.21.82.50.3
49.61.33
16.510.080.157.43
11.782.780.100.14
99.84
10126271191
16130982
24410<l
1
49810.310.100.240.821.57
8130.15
0.00.623.632.20.0
20.810.07.41.82.50.3
49.61.32
16.39.890.157.1412.072.660.100.13
99.43
1032677219<l1653095424417<l<l
24830.320.180.251.391.61_0.15
0.00.6
22.632.50.0
22.012.04.91.72.50.3
49.71.33
16.310.030.157.2412.162.690.060.13
99.82
9927311218<l16131972
304126
<l
48820.320.120.310.951.61—0.19
0.00.422.832.30.022.311.15.71.72.50.3
49.71.6314.811.200.137.1511.382.630.390.17
99.24
8024481205
178361182
4571351
59830.310.110.380.891.82
6420.25
0.02.3
22.427.70.023.014.33.82.03.10.4
49.11.57
14.810.710.147.5211.872.480.360.17
98.72
11125279185
17033
1175374131
<l
23800.280.110.320.971.65
5930.22
0.02.1
21.328.60.0
24.511.85.51.93.00.4
49.41.66
14.810.980.147.3711.492.570.350.17
98.90
9824181194
16838120
7269193
<l
17830.320.160.221.231.73
7180.15
0.02.1
22.028.20.023.213.64.51.93.20.4
49.31.60
14.710.640.137.7911.602.540.320.17
98.78
12423479203
167361164314193
<l
29830.310.160.271.301.58
8800.19
0.01.9
21.828.00.0
23.813.05.3 .1.93.10.4
49.61.48
14.210.510.138.4411.222.530.400.17
98.67
15627876165
162341095
454134
<l
22810.310.120.410.941.44
6610.28
0.02.4
21.726.50.023.414.25.8
1.92.80.4
48.71.62
15.29.460.136.6513.882.550.240.18
98.57
8621975213
176381223
3161841
41800.310.150.251.161.65
6610.18
0.01.4
21.929.60.0
31.92.07.11.73.10.4
49.91.74
15.210.470.146.6111.412.850.470.19
99.05
54213821914
167391313
565162
<l
44800.300.120.431.011.84
2790.34
0.02.824.327.70.0
23.111.63.91.83.30.5
49.51.70
14.810.610.137.2811.672.720.330.18
98.92
10322883235
17139122
73651871
17840.320.150.301.131.69
5500.21
0.02.0
23.327.60.0
24.310.85.51.93.30.4
with unit 444-2 and the other Shikoku Basin basalts. Inaddition, the abundances of Na2O, K2O, P2θ5, Rb, Sr,Nb, Ba, La, and Ce are significantly higher than in unit444-2, and this has resulted in high Ba/Sr, Ba/Zr,Ce/Zr, and CeN/YN ratios in the 444-1 basalts.Although both K2O and Rb are high in these basalts(over 1.5% and 25 ppm, respectively), the strong correl-ation between these two elements, and between themand Zr (Figure 30), suggests that alteration is not entire-ly responsible for this enrichment. Indeed, the very highK/Rb ratio (generally greater than 1000) testifies againstlow-temperature alteration by sea water, which is knownto reduce the K/Rb ratio in oceanic basalts (e.g., Hart etal., 1974). The alkaline affinity of the basalts of geo-chemical unit 444-1 is also supported by the low Zr/Nb,Ti/Zr, and Y/Zr ratios (Table 7; Figures 31, 32, and
33), and of course by the presence of kaersutite andgest that significant fractionation of these phases hastitaniferous augite. In many respects, the basalts of geo-chemical unit 444A-1 closely resemble E-type ocean-ridge basalts from the North Atlantic (Tarney et al., inpress).
The high Ni and Cr contents of the unit 444-1 basaltsindicate that the overall high abundances of many of theHYG elements are not due to high-level fractional crys-tallization involving pyroxene, olivine, or chrome spin-el. However, within-unit variation (Figure 34) does sug-occurred locally, resulting in the low Ni and Cr and highHYG-element abundances observed in Core 20-1, 73cm. This interval is also mineralogically distinct fromthe remainder of lithologic unit 1, containing titan-augite, ilmenite, and abundant kaersutite, and it possi-
823
N. G. MARSH, A. D. SAUNDERS, J. TARNEY, H. J. B. DICK
TABLE 6 - Continued
CoreSectionInterval (cm)
SiO2 (%)TiO2
A12O3
F e 2 O 3
MnOMgOCaONa2θK2O
P2O5
Total
Ni (ppm)CrZnGaRbSrYZrNbBaLaCePbTh
Zr/NbTi/ZrY/ZrCe/ZrBa/ZrCeN/YN
Fetot/MgK/RbBa/Sr
QOrAbAnNeDiHy01Mt11Ap
5728
49.91.69
15.210.360.146.73
12.202.680.390.17
99.39
92237
982210
16640
1248
486
167
< l
15820.320.130.390.981.79
3210.29
0.02.3
22.828.3
0.025.710.7
3.81.83.20.4
572
109
49.41.71
14.811.340.186.46
12.282.510.470.17
99.33
90222
9418
6159
41123
549
611
21
25830.330.090.400.662.04
6560.31
0.02.8
21.427.8
0.026.610.9
3.82.03.30.4
581
58
49.71.68
15.310.280.146.91
11.072.860.410.18
98.56
79233
891917
16839
1245
495
161
< l
25810.310.130.401.011.73
2020.29
0.02.5
24.628.00.0
21.712.64.21.83.20.4
581
94
50.01.91
15.49.840.146.90
11.032.960.490.20
98.93
62227
802115
17544
1385
526
174
< l
2883
0.320.120.380.951.65
2700.30
0.02.9
25.327.7
0.021.411.6
4.21.73.70.5
582
99
50.31.83
15.59.420.107.52
10.693.090.220.18
98.83
73225
8422
3178
40126
750
717
52
1887
0.320.130.401.041.45
5950.28
0.01.3
26.528.2
0.019.714.0
3.91.73.50.4
583
29
46.11.36
11.113.850.17
14.017.481.730.190.16
96.13
405428102
124
972996
740
614
5<l
1485
0.300.150.421.191.15
3960.41
0.01.2
15.222.8
0.012.026.215.8
2.52.70.4
584
30
50.41.93
14.99.720.109.339.262.850.260.20
99.00
115246
8621
3154
40138
760
518
5< l
2084
0.290.130.431.111.21
7060.39
0.01.5
24.427.4
0.014.322.6
3.11.73.70.5
591
71
49.51.28
16.09.950.176.08
12.942.710.290.13
99.06
92286
6819
6173
2990
228
312
2< l
4585
0.320.130.311.021.90
3960.16
0.01.7
23.130.9
0.027.1
6.15.61.72.50.3
591
129
49.71.19
16.39.550.176.84
12.092.750.310.12
99.00
100285
6416
7165
2988
336
384
< l
2981
0.330.090.410.681.62
3740.22
0.01.9
23.531.5
0.023.1
8.46.51.72.30.3
592
15
49.71.25
16.29.060.137.84
11.582.680.100.12
98.66
104291
8718
1149
2791
323
411
51
30820.300.120.251.001.34
8630.15
0.00.6
23.032.4
0.020.413.5
5.11.62.40.3
592
91
50.01.24
16.48.940.137.34
12.072.720.060.12
99.02
108294
7819
< l156
2984
228
311
6< l
4288
0.350.130.330.931.41_
0.18
0.00.3
23.232.7
0.022.012.9
3.81.62.40.3
5934
49.71.22
16.39.130.137.95
11.622.710.100.11
98.99
113297
7917
1154
2887
331
484
< l
2984
0.320.090.360.701.33
8550.20
0.00.6
23.232.4
0.020.412.26.21.62.30.3
bly results from within-sill fractionation. The HYG-ele-ment ratios of Core 20-1, 73 cm, however, are indis-tinguishable from the rest of geochemical unit 444-1.
Geochemical unit 444-2 comprises basalts which arechemically similar to those recovered at Sites 442 and443. Thus, they contain neither normative nepheline nornormative quartz (Figure 28). They have low abun-dances of K2O, P2O5, Rb, Sr, and Ba, and have Ce/Zr,Ba/Zr, Ce/Y, and K/Rb ratios similar to those of theother Shikoku Basin basalts. Nevertheless, they haveslightly higher Nb (5-9 ppm), and consequently theyhave lower Zr/Nb ratios than the basalts from eitherSite 442 or Site 443, and they have significantly lowerTi/Zr ratios (Table 7; Figures 31 and 32). Cr and Ni areless than 100 ppm and 280 ppm, respectively, and themoderately high Fe/Mg ratios (greater than 1.5) testifyto the fractionated nature of these samples. The higherNi and Cr in Core 27-5, 68 cm, at the base of lithologic
unit 2 may be due to the presence of relict olivine pheno-crysts, possibly in part of cumulus origin. Within-unitfractionation of Ni is clearly indicated by Figure 34.
A plot of Ce against Y (Figure 35) shows that thebasalts of geochemical unit 444-1 are light-REE en-riched, CeN/YN ranging from 2.6 to 2.8, and CeN over30. Geochemical unit 444-2 is also enriched in lightREE, but to a much lesser degree (CeN/YN 1.12-1.46),and thus more closely resembles the basalts from Sites442 and 443 (Figure 35).
The available data suggest that it is unlikely that geo-chemical units 444-1 and 444-2 were derived from thesame parental melt by differing degrees of fractionalcrystallization. Ni and Cr are highest in 444-1 (apartfrom 20-1, 73 cm), which also contains basalts most en-riched in HYG elements. It is also unlikely that the mag-mas of the two units could have been derived from acompositionally identical source by varying the nature
824
GEOCHEMISTRY OF BASALTS
593
101
594
138
59565
60121
60180
602
106
TABLE 6
6040
— Continued
606
104
61193
612
107
61313
61329
613
101
61469
49.71.27
16.09.280.137.66
11.702.770.230.11
98.80
14530175193
15029914313105
<l
23840.320.110.340.851.40
6280.21
0.01.4
23.730.90.022.09.87.01.62.40.3
49.11.26
15.98.580.128.71
11.792.680.070.12
98.26
12429161201
20026901243112
<l
90840.290.120.271.041.14
5730.12
0.00.423.131.60.021.98.99.11.52.40.3
49.61.25
16.38.300.138.54
11.522.690.280.11
98.75
14031180194
15127872292104
<l
43860.310.110.330.911.13
5850.19
0.01.7
23.032.00.020.39.88.31.52.40.3
48.81.18
15.89.110.159.1610.902.610.060.12
97.94
1222935718<l142268512321321
85830.310.150.271.231.15-0.16
0.00.422.531.90.018.212.89.11.62.30.3
49.31.23
15.89.390.169.11
10.672.690.090.11
98.52
1243035718<l144299133128
<l<l
30810.320.090.340.681.20_0.22
0.00.523.131.30.017.613.88.61.72.40.3
49.11.24
15.89.640.168.5711.072.520.090.12
98.38
10329970181
137289213131043
92810.300.110.340.881.30
7060.23
0.00.521.732.20.018.515.66.21.72.40.3
49.31.27
15.89.510.178.5611.352.560.070.12
98.64
9630566192
1472990333366
<l
30850.320.070.370.511.29
3110.22
0.00.422.031.90.020.113.57.01.72.40.2
48.61.14
15.99.150.169.0411.162.570.170.11
97.95
14729974143
17226821292123
<l
82830.320.150.351.131.17
4650.17
0.01.0
22.232.10.019.59.411.01.62.20.1
48.71.11
15.99.390.159.5611.062.470.070.11
98.58
18332461172
1402580<l1951164
_
830.310.140.241.081.14
2780.14
0.00.4
21.232.60.018.112.610.21.72.10.3
48.91.20
16.09.500.158.24
11.862.530.100.11
98.52
12729970193
199278322931051
42870.330.120.350.911.34
2790.15
0.00.6
21.732.40.021.610.38.21.72.30.3
49.41.21
16.19.390.157.0713.062.590.210.11
99.32
10830274195
1672889<l342124
<l
_820.310.130.381.051.54
3450.20
0.01.2
22.131.90.026.66.07.21.62.30.3
49.71.23
16.19.350.137.1712.342.600.210.10
98.98
11730468195
16727884284935
22840.310.100.320.821.51
3440.17
0.01.2
22.232.00.023.810.94.81.62.40.3
49.41.23
16.19.360.138.11
11.882.520.050.11
98.85
105308102181
1512986224213<l<l
43860.340.150.281.101.34
4480.16
0.00.321.632.70.021.313.55.51.62.40.3
49.71.22
16.29.490.147.7812.072.620.050.11
99.39
11731573171
15227853213164
<l
28860.320.190.251.461.41
4070.14
0.00.322.332.60.021.911.95.91.72.30.3
and degree of partial melting, because the low bulk-dis-tribution coefficients of K, Rb, Zr, La, Ce, and Nb (see,for example Wood et al., 1979) make it difficult tochange ratios of the more-HYG elements, even if verysmall, but still realistic, degrees of partial melting are in-voked. Instead, we suggest that geochemical units 444-1and 444-2 were derived from compositionally distinctmantle sources, the source of 444-1 being enriched inCe, La, Nb, K, Rb, Ba, and Sr relative to the source of444-2.
Both units 444-1 and 444-2 were emplaced as sillslong after basement formation; therefore, it is not sur-prising that the basalts of unit 444-1 are compositionallysimilar to E-type ocean-ridge basalts, which are charac-teristic of off-axis volcanic activity (see, for exampleTarney et al., 1979, in press; Wood et al., 1979). Evenunit 444-2, which is not as enriched as 444-1, is more en-riched than N-type ocean-ridge basalts, having higher
Ce/Y and lower Ti/Zr and Zr/Nb ratios; the 444-2 ba-salts more closely resemble T-type ocean-ridge basalts.The Th, Hf, and La data presented by Wood et al. (thisvolume) also support this comparison.
NORTHEAST PHILIPPINE BASINTwo sites are in the northeast corner of the Philippine
Basin. Site 445 is in a small basin on the southern side ofthe Daito Ridge (Figure 1), and the only igneous rocksrecovered at this site are a few clasts and pebbles insome of the conglomerates. These have been studied indetail by Mills (this volume).
Site 446 is in the Daito Basin, south of the DaitoRidge (Figure 1), at 24°42.04'N, 132°46.49'E, in 4952meters of water. Two holes were drilled, and basalts in-terbedded with sediments were encountered beneathabout 380 meters of sediment; the oldest sediment gavean age of early Eocene. From several lines of evidence,
825
N. G. MARSH, A. D. SAUNDERS, J. TARNEY, H. J. B. DICK
TABLE 6 - Continued
CoreSectionInterval (cm)
SiO2 (%)TiO2
AI2O3F e 2 O 3
MnOMgOCaONa2OK 2 O
P2O5
Total
Ni (ppm)CrZnGaRbSrYZrNbBaLaCePbTh
Zr/NbTi/ZrY/ZrCe/ZrBa/ZrCeN/YN
Fe tot/MgK/RbBa/Sr
QOrAbAnNeDi
Hy01Mt11
Ap
621
94
50.01.29
15.98.580.128.03
11.952.550.070.13
98.61
94290
6719
1158
2991< l37
311
3< l
_
85
0.320.120.410.931.24
6060.23
0.00.4
21.932.3
0.021.915.7
2.81.52.50.3
622
27
49.91.32
15.98.630.127.58
11.692.830.250.13
98.38
97293
6918
3161
3098
233
282
< l
4981
0.310.080.340.651.32
6950.20
0.01.5
24.330.5
0.022.410.6
5.51.52.50.3
623
56
49.91.25
15.99.280.157.66
12.322.560.250.11
99.41
127321
7519
4163
2890
228
31351
4583
0.310.140.311.141.40
5130.17
0.01.5
21.831.4
0.023.910.9
5.51.62.40.3
624
18
50.11.29
16.08.990.136.82
12.622.620.250.12
98.93
120310
9919
5165
3094
129
411
6< l
9482
0.320.120.310.901.53
4080.18
0.01.5
22.431.4
0.025.311.5
2.71.62.50.3
631
140
50.01.27
15.99.250.137.10
12.272.730.260.12
99.00
116306
7720
5165
2993
128
616
6< l
9382
0.310.170.301.361.51
4230.17
0.01.5
23.330.7
0.024.5
9.85.11.62.40.3
633
116
49.81.24
15.98.610.117.45
12.592.680.230.12
98.72
180307
6417
3166
2891
235
291
< l
45820.310.100.380.791.34
6420.21
0.01.4
23.031.1
0.025.5
8.16.01.52.40.3
6355
49.71.27
15.99.470.158.39
11.322.670.100.12
99.10
94308
8618
1146
3089
232
395
<l
4586
0.340.100.360.741.31
7970.22
0.00.6
22.831.4
0.019.813.7
6.41.72.40.3
636
12
49.51.29
15.89.450.168.52
11.372.650.090.13
99.02
97300
7719
1152
2891
233
210
7< l
4585
0.310.110.360.881.29
7640.22
0.00.5
22.631.4
0.020.113.0
7.11.72.50.3
6^8
54
50.01.30
16.09.440.158.11
11.532.810.090.14
99.56
96300
6321
2148
3090
326
414
22
3087
0.330.160.291.151.35
3690.18
0.00.5
23.931.0
0.020.811.5
7.01.62.50.3
641
78
48.71.15
15.89.590.159.47
10.672.380.110.12
98.17
139319
5520
2137
2782
230
412
61
4184
0.330.150.371.091.17
4650.22
0.00.7
20.532.7
0.016.6-16.4
8.11.72.20.3
642
97
48.41.11
15.69.470.15
10.0910.65
2.230.110.12
97.94
170323
6217
2137
2678
125
211
51
7885
0.330.140.321.041.09
4520.18
0.00.7
19.332.9
0.016.416.9
8.91.72.20.3
6430
48.71.14
15.79.450.159.82
10.712.310.130.12
98.20
144311
6317
3136
2684
133
511
5< l
8481
0.310.130.391.041.12
3570.24
0.00.8
19.932.6
0.016.816.5
8.41.72.20.3
6445
49.81.27
15.99.660.157.03
12.402.600.260.12
99.14
96310
7616
6173
3091
332
4114
< l
3084
0.330.120.350.901.59
3620.18
0.01.6
22.231.2
0.024.610.4
4.81.72.40.3
644
32
49.81.24
15.99.500.147.38
12.292.630.240.12
99.31
119304
7517
3193
2989
428
296
< l
2284
0.330.100.310.761.49
6720.15
0.01.4
22.431.2
0.024.010.0
5.81.72.40.3
aNormative values based on Fe2θ3/FeO ratio of 0.15. Fe 2 θ3 *s t o t a ^ ^on a s F e 2θ3 CeN/YN represents ratio of chondrite-normalized values.
but principally the abundance of thermally altered sedi-ments above and below most of the igneous units, all thebasalts at Site 446 are considered to have been emplacedas sills (Site 446 report, this volume) during post-earlyEocene time. Basement, which was not reached, isthought to have an early Eocene age at this site.
The basaltic rocks recovered from Hole 446A havebeen subdivided into 9 lithologic and 15 geochemicalunits (Figure 36), and 23 individual sills have been rec-ognized. These vary in thickness from 0.3 to 23 metersand can be delineated by aphanitic or occasionally evenglassy chill zones. Two main lithologic types of basaltand dolerite have been identified: a kaersutite, horn-blende-bearing variety (lithologic units 1, 4, 5), and ahornblende-free variety (lithologic units 2, 3, 6, 7, 8, 9).
Lithologic units 1, 4, and 5 contain 5 to 15 per cent ofprimary kaersutite, which occurs either as free-growinganhedral laths, as overgrowths on clinopyroxene, or as
fibrous clumps. The pyroxene in these kaersutite-bear-ing samples has a distinct purplish-pink coloration in-dicative of a high TiO2 content, and the rocks containabundant granular magnetite and skeletal ilmenite. Pla-gioclase and (in unit 5) olivine are the other main miner-al components; apatite crystals were observed inlithologic units 4 and 5.
Lithologic unit 1 comprises two equigranular, highlyvesicular (up to 30% vesicles), fine- to medium-grainedbasaltic sills. The sills have an intergranular texture, andappear to be fresh. Lithologic unit 4 also comprises twopetrographically similar sills, separated by a clay stoneinterbed. The thicker, upper sill has been divided intotwo sub-units by a break in recovery; the upper sub-unit(4A) consists of highly vesicular (25-40%), aphanitic tocoarse-grained basalt. Sub-unit 4B is a fine-grained,amygdaloidal basalt, and sub-unit 4C is a smaller-scaleversion of the upper sill. Unit 4 is cut by several clay-
826
GEOCHEMISTRY OF BASALTS
200
150
?αα- 100
50
-
-
-
-
-
Site.
#
A
X
0
D
443-1443-2443-3443-4443-5
1 'C \1\\\\1
\
\
°\o\\
<faon o
O -Q.
ooo\ 1
1
I\
\
\\
Site 442 —»-\
t i
1
\\ x\
\\\
\ \
AA
A
1
-
-
-
-
-
\ x
\ x "
\x \
x ^x x X
X\ -
• ×l• / --
50 100Zr (ppm)
150
Figure 18. TV/ versws ZΛ Site 443. The dashed line out-lines the field of Site 442 basalts.
2.0
0.5
Site 443
• 443-1A 443-2
- x 443-3O 443-4G 443-5
/
<S>y
Ax
A
D
i i , I i ,
'/ x
x 4$y^
X X ^ y / ^
-
:
100Zr (ppm)
Figure 19. TiO2 versus Zr, Site 443. Note the strongcolinearity of the data points.
and carbonate-lined veins or fractures, and the vesiclesin the lower sections of both sills are filled by light-olive-green clay and calcite. Lithologic unit 5 is a single sill,containing up to 35 per cent dark, chloride material,which is occasionally recognizable as pseudomorphsafter olivine (0.1-1.0 mm across).
The kaersutite-free units essentially comprise plagio-clase, augite, titanomagnetite, and chloritic or clayeypseudomorphs after olivine. Lithologic unit 2 is a singleaphyric-basalt sill, the upper part of which is vesicular;the basalt is fine- to medium-grained, with intersertaltexture. The unit is cut by several veins lined with clay,quartz, and pyrite. Lithologic unit 3 comprises three
3 0 -
20
Site 443
• 443-1A 443-2x 443-3O 443-4G 443-5
-
A A«EAcA *
OOCCEO X
αx>αo o/ D OCD O
O
AA X .
A
xx
X X
• XX
•β
X
X
X
/
X
X
-
1 1 1
50 100Zr (ppm)
Figure 20. Y versus Zr, Site 443.
Site 443
• 443-1A 443-2x 443-3O 443-4D 443-5
—
I > i i i I i i
MX
X
A A AOA AAAOO AAO
D DDOO3C33D
ooooooO
1 . . . . 1 . .
i i | i i ' '
•
X —
X
XXX X X X ^ — '
\1i * \. --—'^}i•—
-
-
1 1
100Zr (ppm)
Figure 21. P2O5 versus Zr, Site 443. The colinearity ofthe data points should be compared with the scatterobserved in the data from Site 442 (Figures 10 and15).
Site 443
• 443-1A 443-2x 443-3O 443-4G 443-5
^ ^
1
X ~
X X
x " x x×
x x J^~-~o x X
°rf&6 o x
°° A 10 -
D A
i
100Zr (ppm)
Figure 22. K2O versus Zr, Site 443.
827
N. G. MARSH, A. D. SAUNDERS, J. TARNEY, H. J. B. DICK
200
100
Site 443
• 443-1A 443-2x 443-3O 443-4D 443-5
' /
S i I I i
o o
o oo
o °
X
' l • ^ l 1 1
4/
A
/
/
x x × x
X X
** x* * × /*
-
I I
100
Zr (ppm)
Figure 23. Sr versus Zr, Site 443.
Site 443
• 443-1A 443-2x 443-3O 443-4D 443-5
100
Zr (ppm)
Figure 24. i?α versus Zr, Site 443.
20
1 0 r -
Site 443
• 443-1
A 443-2
x 443-3
O 443-4
H 443-5
•
• /
-
-
-
' A
O O
O Ax
CO OA • A
a Φ A J A
QO D CC3 O
O O 0 0 A
cno
CD O O
O
• 1 1 , 1
X X
X
X
• • • x
A •X XX
X
_
X
-
-
-
-
50 100
Zr (ppm)
150
20 30
Y (ppm)
Figure 25. Ce versus Zr, Site 443.
Figure 26. Ce VÉTSWS Y, Site 443. Field for Site 442shown for comparison.
sills: sub-unit 3A is a plagioclase (40%) and augite (1 °/o)microphyric basalt; 3B is a plagioclase (3-15%) andaugite (0.8%) phyric basalt; and 3C is a sparselyplagioclase-pyroxene phyric basalt. All sub-units con-tain a few clay pseudomorphs after olivine, and 3B and3C have vesicular upper portions.
Lithologic unit 6 comprises six fine-grained, sparselyphyric basalt sills. The basalt is essentially non-vesicu-lar, apart from the top of unit 6C, which contains 10 to15 per cent vesicles, and rare plagioclase phenocrysts.Lithologic unit 7, however, comprises two almost iden-tical plagioclase-pyroxene-phyric sills separated by aclaystone interbed. The upper sill, sub-unit 7A, has aglassy margin which contains a few red-brown pseudo-morphs, possibly after olivine. Lithologic units 8 and 9both comprise aphyric basalt sills, the former contain-ing rare plagioclase and augite microphenocrysts. Bothunits are cut by clay- and calcite-lined fractures.
Major- and trace-element data for the basalts recov-ered at Site 446 are listed in Tables 8 and 9. The ninelithologic units comprise at least 15 geochemical units(Figure 36), although lithologic units 6B and 6C werenot sampled. It is apparent that breaks between chemi-cal units usually coincide with breaks between lithologi-cal units or sub-units, but where a significant change inchemistry occurs without a change in petrography theunits are divided into geochemical sub-units.
Perhaps the most noticeable feature of the DaitoBasin samples, apart from their high contents of TiO2,total iron, K2O, P2O5, Sr, Ba, La, Ce, Zr, and Nb com-pared with the basalts from the Shikoku Basin, is theirwide range in chemical composition, both within andamong the various units. This is illustrated by the down-hole plot (Figure 36), and also by the normative basaltdiagram of Figure 37, where the range of chemis-try—from nepheline-normative basalts to quartz-nor-mative tholeiites—becomes apparent. Two interdigi-tated groups have been recognized: an alkalic-basaltsuite comprising geochemical units 446-1, 4, 5, 7a, and10a; and a tholeiitic-basalt suite comprising units 446-2,3,6, 7b, 8,9, 10b, 11, 12, 13, 14, and 15. This division isnot based solely on the presence or absence of norma-tive nepheline; indeed, most of the samples in unit 446-5(and all those in 446-7a) contain normative hypersthene.Rather, the alkalic basalts are characterized by lownormative-hypersthene contents, or the presence of nor-mative nepheline, and also by high abundances of cer-
828
GEOCHEMISTRY OF BASALTS
Core Lithologic MagneticNo. Unit Inclination
-90° 0° +90° 100
Zr (ppm) Ba (ppm) Cr (ppm)
220 50 150 100 400
Fe/Mg Y/Zr Ti/Zr
1 2 3 0.1 0.4 60 70 80 90
E ~
15
1 1
240 33320
18 Mudstone and ClaystoneI " ~~~—I I I I I I I I I I 1 T -—i—i—i—i—r
Unit 1Equigranular dolerite ?
S 260-Q
oo 2 8 0 -
3 0 0 -
21
22
23"t2£|
25
26
27
Mudstoneand
claystone
Unit 2Plagioclase-pnyric
dolerite
Figure 27. Down-hole variations in lithology, magnetic inclination, and selected geochemical parameters, Site 444.
TABLE 7Analyses, Element Ratios, and CIPW Norms for Basalts from Hole 444Aa
CoreSectionInterval (cm)
SiO 2 (%)TiO 2
A1 2O 3
F e 2 O 3
MnOMgOCaONa2θK 2 OP2O5
Total
Ni (ppm)CrZnGaRbSrYZrNbBaLaCePbTh
Zr/NbTi/ZrY/ZrCe/ZrBa/ZrCeN/YNF etot/ M gK/RbBa/Sr
QOrAbAnNeDi
HyOlMt11Ap
201
22
47.71.55
16.08.650.158.678.073.051.780.31
95.95
137222
672014
33630
15824
1411433
53
759
0.190.210.892.701.16
10550.42
0.011.025.725.8
0.611.30.0
19.51.63.10.8
201
73
50.31.94
16.27.960.156.258.593.862.310.39
97.92
26127
791816
39138
20931
1771842
44
756
0.180.200.852.711.48
11980.45
0.013.927.420.5
3.216.7
0.011.5
1.43.80.9
2020
47.21.41
16.48.490.167.708.433.621.900.34
95.68
180246
912017
29527
14624
1461530
43
658
0.180.211.002.731.28
9280.49
0.011.722.123.9
5.414.30.0
16.71.52.80.8
203
39
47.11.43
16.28.450.157.898.773.561.800.29
95.71
172234
621914
28427
14923
1351431
4<l
658
0.180.210.912.821.24
10670.48
0.011.121.324.0
5.515.8
0.016.4
1.52.80.7
204
37
46.71.43
16.58.360.157.698.103.922.060.26
95.18
182245
571816
28025
14524
1511126
52
659
0.170.181.042.551.26
10690.54
0.012.819.722.3
8.214.7
0.016.5
1.52.90.7
232
84
49.61.50
15.99.690.317.01
12.412.720.240.20
99.63
71253
7619
3163
35118
557
61741
2476
0.300.140.481.191.60
6640.35
0.01.4
23.130.6
0.024.4
8.46.11.72.90.5
241
87
49.71.57
15.910.03
0.166.89
11.272.710.170.21
98.63
59238
7020
2160
37130
865
822
3<l
1672
0.280.170.501.461.69
6930.41
0.01.0
23.231.3
0.019.617.01.61.83.00.5
242
73
50.01.59
16.010.320.177.25
11.092.720.220.22
99.56
59245
7324
3153
37134
970
821
51
1571
0.280.160.521.391.65
6060.46
0.01.3
23.130.9
0.018.717.0
2.71.83.00.5
251
118
50.21.61
16.110.15
0.176.87
11.362.820.160.23
99.73
53245
6320
216040
1417
718
2032
2068
0.280.140.501.231.71
6640.44
0.00.9
23.930.9
0.019.716.3
1.91.83.10.5
253
50
49.81.46
16.79.500.166.87
11.612.640.170.21
99.14
57235
6220
2159
33120
857
617
61
1573
0.280.140.471.271.60
6930.36
0.01.0
22.533.6
0.018.916.4
1.81.72.80.5
261
49
50.01.47
16.49.620.176.72
11.862.670.240.21
99.40
62234
8821
3159
36123
854
1952
1572
0.290.150.441.301.66
6590.34
0.01.4
22.732.4
0.020.814.7
2.21.72.80.5
2630
50.01.37
17.18.960.156.54
12.222.750.190.20
99.49
69273
8319
2167
32123
652
517
31
2067
0.260.140.421.301.59
7970.31
0.01.1
23.433.8
0.021.112.1
3.01.62.60.5
272
62
48.91.15
16.48.910.198.09
10.962.690.340.18
97.87
97262
6222
2266
29109
654
715
8<l
1863
0.270.140.501.271.28
14030.20
0.02.0
23.332.5
0.017.810.6
8.81.62.20.4
273
21
49.71.33
16.69.600.177.35
11.732.610.230.19
99.48
95277
7217
2159
33117
663
615
<l<l
1968
0.280.130.541.121.51
9590.40
0.01.4
22.232.9
0.019.813.5
4.71.72.50.5
275
68
49.11.38
15.610.02
0.177.46
12.672.490.220.20
99.29
114325
8519
2380
34128
762
616
4<l
1865
0.270.130.481.161.56
8920.16
0.01.3
21.230.9
0.025.4
8.56.91.82.60.5
aNormative values based on Fβ2θ3/FeO ratio of 0.15. F e 2 θ 3 is total iron a s F e 2 θ 3 represents ratio of chondrite-normalized values.
829
N. G. MARSH, A. D. SAUNDERS, J. TARNEY, H. J. B. DICK
Figure 28. Normative nepheline (Ne), olivine (Ol), di-opside (Di), hypersthene (Hy) tetrahedron for Site444. Normative values based on Fe2O3/FeO ratio of0.15.
Px
10
2080
PI
70 60 50
Ol
30
Figure 29. Normative plagioclase-pyroxene-olivinediagram for basalts from geochemical unit 444-2(Site 444). The Fo-An cotectic (C) is from Shido etal. (1971).
tain hygromagmatophile elements (particularly K, Rb,Sr, Ba, Ti, Ce, and La), relative to the tholeiitic group.Note also that units 446-1, 4, and 5 contain kaersutite.
Intra-unit variation is well-illustrated by the range ofFe/Mg ratios (Figure 36) or, even better, by the Ni/Zrratio (Figure 38). With the exception of geochemicalunit 446-5, it is apparent from the high Fe/Mg ratios(generally in excess of 2.5), the low Ni abundances (less
2.5
2 . 0 -
1.5 -
1.0 -
0.5 -
1 ' . i | i . i
Site 444
•A
-
, ,
i 1
, ,
ii
1 ,
, .
. 1
i-
444-1444-2
Site 442 —>A
J, I i i •"~
i | > • .
,'2 \— I""* i i .
. | . . . . | . . . ,
•
•
••
|-<— Site 443
50 100 150Zr (ppm)
Figure 30. K2O versus Zr, Site 444. Field for Site 442from Figure 11; field for Site 443 from Figure 22.
30
I 20
Q
10
n
_
•
-
Site 444
• 444-1A 444-2
Site 442
. . . I
• . . . |
. . . . I
•
•
- Site 443
50
Zr (ppm)
Figure 31. Nb versus Zr, Site 444. Fields for Sites 442and 443 from Figure 9.
2.0
1.0
• Site 444
" 444-1" A 444-2
Site 4 4 2 — • ~ ^ ' '
• . . I . . .
1 . . . . 1
^ i/ /-*
/××/7 ± A >
i . . . . i
Site 443
•
-
-
100 150Zr (ppm)
Figure 32. TiO2 versus Zr, Site 444. Field for Site 442from Figure 7; field for Site 443 from Figure 19.
830
GEOCHEMISTRY OF BASALTS
1 1 1 1 1
Site 444
• 444-1A 444-2
Site 442
. . . . I
I '
J
I i
i i i | i i i i | i i i i
/ Site 443
"AV A *
r
Zr (ppm)
Figure 33. Y versus Zr, Site 444. Field for Site 442 fromTables 2 and 3; field for Site 443 from Figure 20.
150
3 100
50
Site 444
• 444-1A 444-2
100Zr (ppm)
150 200
Figure 34. Ni versus Zr, Site 444.
50
?30hQ.
<3 20
10t-
. 1 . 1 1 . . .
Site 444
• 444-1A 444-2
•
. i . i
••
A i
i 1 •
•
àA
1
A
[
\ -
-
20 30Y (ppm)
50
Figure 35. Ce versus Y, Site 444.
than 100 ppm, decreasing to less than 50 ppm in units446-1, 3, 4, 7a, 7b, 8, 9, 10a, 10b, 12, and 15), and thelow Cr abundances (generally less than 150 ppm) thatthe basalts of Site 446 have undergone extensive frac-tionation and that they do not represent primary liq-uids. Unit 446-5, however, has significantly lowerFe/Mg ratios (0.90-1.68) and concomitant high MgO(7.3-17.7%), high Ni (90-627 ppm), and high Cr (305-1206ppm) contents. The consanguineous nature of the 446-5
suite is clearly illustrated on the Ni/Zr plot (Figure 38),and on several of the hygromagmatophile-elementplots. High contents of altered olivine crystals (up to30%) in parts of unit 446-5 strongly suggest that cumu-lus olivine may have caused the strong variation withinthis unit. However, implicit in the data is that cumuluschrome spinel, or pyroxene, also played an importantrole, because Cr increases in proportion to Ni (Table 9);olivine contains very little Cr. Units 446-2, 3, and 7balso show within-unit variation on the Ni/Zr diagram,although in units 446-2 and 446-3 Ni (and Cr) is practi-cally invariant (Figure 38). This suggests that fractionalcrystallization involving olivine or pyroxene has notplayed an important role in determining the abundancevariation of Zr (and other HYG elements) in units 446-2and 446-3.
Although the high contents of many of the HYG ele-ments are in part due to the fractionated nature of the446 basalts, it is also apparent that many HYG-elementratios are very different between the tholeiitic and al-kalic groups, and between the Site 446 and Shikoku Ba-sin basalts (apart from unit 444-1, which resembles thealkaline groups of 446). This is illustrated by plottingvarious HYG elements against Zr (Figures 39 to 45).
The alkalic basalt group has high K/Zr, Ba/Zr,Rb/Zr, and Sr/Zr ratios compared with the tholeiiticgroup. The Ba/Zr ratio, for example (Figure 39), is gen-erally between 0.98 and 2.08 in the former, and between0.38 and 1.08 in the latter (Ba/Zr is less than 0.4 in theleast-altered samples from the Shikoku Basin, althoughthe upper sill, unit 444-1, has a Ba/Zr ratio of 0.85-1.04;see Table 7).
Ti/Zr ratios (Figure 40) are high in all Site 446 ba-salts, especially so in the alkalic group; however, this isunlike unit 444-1, which has low Ti/Zr ratios (56-59).The alkalic group has the highest Ti/Zr ratios (up to163) of any basalts recovered during Leg 58, a reflectionof the high ilmenite and TiO2 contents of these samples.The alkalic group also has the highest CeN/YN ratios(4.09-5.92; Figure 41), although the tholeiitic groupalso exhibit relative light-REE enrichment (CeN/YNover 3). However, it is not only the light-REE contentswhich are high; it is apparent from the REE patterns forthe Site 446 basalts (Wood et al., this volume) that theheavy REE, at least in the alkalic group, are significant-ly lower than in the Shikoku Basin basalts, even thoughthe Site 446 rocks have undergone extensive fractiona-tion. Thus, although the Y/Zr ratios of the alkalic andtholeiitic suites are similar (0.12-0.17; Figure 42), theyare significantly lower than those of most of the Shiko-ku Basin basalts (0.3-0.4), apart from unit 444-1 (0.18).
All Site 446 basalts have low Zr/Nb ratios (less than12); in particular, units 446-4, 5, and 7a have Zr/Nb ra-tios less than 7 (Figure 43). Low Zr/Nb ratios are char-acteristic of "enriched" or "plume"-type basalts pres-ently being erupted along certain sections of the Mid-Atlantic Ridge (e.g., 45°N; Erlank and Kable, 1976;Tarney et al., 1979; Wood et al., 1979), on ocean islands(e.g., Iceland; Wood et al., in press), and in intra-continental alkali-basalt provinces (e.g., East AfricanRift; Weaver et al., 1972). The Ce/Zr ratios of the Site
831
00
to
380-
4 0 0 -
4 2 0 -
4 4 0 -
4 6 0 -
500-
520 H
5 4 0 -
5 6 0 -
580 -
600
620 -
Lithologic MagneticUnit Inclination
-90° 0° +90° 100
.1Zr (ppm) Ba (ppm) Cr (ppm) Fe/Mg Y/Zr Ti/Zr |
o400 100 300 500 200 600 1200 1 2 3 4 0.1 0.2 80 100 120 140 160 <S
Unit 1 Hb-basalt
Unit 2
Aphyric basalt
Unit 3Plagioclase—pyroxene
phyric basalt
Unit 4Hb-bearing vesicular
doleriteUnit 5 Ol-phyric basalt
Unit 6Aphyric basalts
Unit 7Pl-Cp× basalt
Unit8Aphyric to sparsly
plagioclase-phyric basalt
Unit 9Aphyric basalt
t
, g _10a,
-12-
13
14
15
Figure 36. Down-hole variation in lithology, magnetic inclination, and selected geochemical parameters, Site 446. Shaded areas repre-sent sediment.
GEOCHEMISTRY OF BASALTS
CoreSectionInterval (cm)
SK)2 (%)TiO2
A12O3
Fe 2O 3
MnOMgOCaONa2OK 2 0
P2O5L.o.i.
Total
Ni (ppm)CrZnGaRbSrYZrNbBaLaCePbTh
Zr/NbTi/ZrY/ZrCe/ZrBa/ZrCeN/YN
Fetot/MgK/RbBa/Sr
QOrAbAn
NeDiHy01Mt11Ap
Analyses,TABLE 8
Element Ratios, and CIPW Normsfor Basalts from Hole 446a
41CC20
46.43.58
14.79.850.334.55
11.393.092.390.382.59
99.25
2839952337
69328
21029
33024
5131
7102
0.130.241.574.472.51
5350.48
0.014.214.119.3
6.628.9
0.03.91.76.90.9
42CC24
45.83.67
14.710.880.364.80
11.143.162.040.372.53
99.40
2918992236
68927
20627
3132245
51
1070.130.221.524.092.63
4700.45
0.0
12.114.919.9
6.527.5
0.05.81.97.00.9
434
77
47.54.18
13.613.950.256.349.792.680.810.480.74
100.31
77124132
289
39642
29634
2063168
63
985
0.140.230.703.982.55
7510.52
0.04.8
22.622.5
0.018.915.1
2.82.47.91.1
441
19
49.74.02
12.813.39
0.205.709.392.860.670.431.43
100.62
73122110
279
35937
27533
1932862
71
8880.130.230.704.122.72
6180.54
2.83.9
24.020.0
0.019.416.3
0.02.37.61.0
442
53
50.54.17
12.913.78
0.194.719.592.610.850.460.11
99.87
69113130
2515
36840
28335
2313058
41
888
0.140.200.823.563.39
4700.63
5.45.0
22.120.9
0.019.814.1
0.02.47.91.1
444
56
49.33.92
12.614.46
0.265.549.362.620.530.450.59
99.71
69112119
2613
37339
28032
1793164
37
9840.140.230.644.033.03
3360.48
3.63.1
22.321.2
0.018.718.2
0.02.57.51.1
461
49
47.14.12
14.113.810.175.529.832.630.510.451.69
99.91
4053
12928
6424
38290
34151
3270
24
9850.130.240.524.522.90
7070.36
0.63.0
22.325.2
0.017.217.5
0.02.47.81.1
aNormative values based on Fe2θ3/FeO ratio of 0.15. Fe 2 θ3 is total iron as Fe2θ3.Cej\[/Yfv[ represents ratio of chondrite-normalized values.
446 basalts are also characteristic of E-type ocean-ridgebasalts (Figure 44) (-0.3; Tarney et al., 1979), beingsignificantly higher than in the Shikoku Basin basalts(less than 0.2), or in T-type and N-type ocean-ridgebasalts.
Units 446-4, 5, and 7a have high P/Zr ratios (higherthan 10; Figure 45), unlike the other units at Site 446,and unlike the basalts recovered from the ShikokuBasin. The wide variation in P/Zr ratios within unit446-5 basalts may be due to the presence of cumulusapatite, although the high phorphorus content in thealkalic units must be a function of the melt (and proba-bly the source) composition. Note, however, that theother alkalic units—446-1 and 446-10a—have P/Zr ra-tios similar to the tholeiitic suites, although there areslight differences in the P/Zr ratios of the various units.
The processes giving rise to the inter-unit chemicalvariations at Site 446 are difficult to quantify withoutdetailed major- and trace-element modeling. A compre-
hensive analytical survey of the constituent minerals isnecessary before this can be attempted. It is apparent,however, that the two main compositional types—al-kalic basalts and tholeiitic basalts—cannot easily berelated by fractional crystallization, because of thesignificant differences in the Zr/Nb, Th/Hf (Wood etal., this volume), and alkali-elements/Zr ratios of thetwo suites. In addition, it may be difficult to relate thetwo suites by varying degrees of partial melting of acompositionally homogeneous source, although furthermodeling is necessary to ascertain this.
SUMMARY AND DISCUSSION
One of the objectives of Leg 58 was to obtain samplesof basaltic basement from the Shikoku and Daito Ba-sins, in order to compare the petrological and chemicalcharacteristics of this basement with those of other mar-ginal basins and of the major ocean basins. Unfortun-ately, although substantial amounts of basalt were re-covered at all sites except 445, only drilling at Site 442and possibly Site 443 penetrated basalt belonging une-quivocally to oceanic layer 2; drilling at all the re-maining sites recovered post-basement basaltic materialwhich may represent, in part, off-axis activity. It isunlikely, however, that the sill and flow emplacement atSites 442, 443, and 444 post-dates true basement bysignificantly more than 2 m.y. (80 km from the ridgeaxis at a spreading rate of 4 cm/yr), and for the pur-poses of the present discussion it is considered that this"off-axis" activity tapped a mantle source similar tothat which gave rise to the underlying basement.
The crustal structure of the Shikoku Basin, and thecomposition of the basalts recovered during Leg 58,support the suggestion of Karig (1971) that the majorityof the western Pacific marginal basins are extensionaland formed by sea-floor spreading broadly analogous tothat occurring in the major ocean basins. The ShikokuBasin basalts exhibit considerable chemical variation,which may be due to a combination of several processes:low-temperature and hydrothermal alteration by seawater; within-flow and within-sill differentiation; high-level fractionation in sub-ridge magma chambers; vary-ing degrees of partial melting under varying P, PH2o>and Pθ2 conditions, hence different stable-phase assem-blages during partial melting; and heterogeneity in themantle source beneath the ridge axis. It is important toassess these effects before making any regional com-parisons with basalts from different tectonic environ-ments, and before evaluating the significance of basaltchemistry on the mechanisms of back-arc spreading.
It is apparent from Figures 10 to 14 and 22 to 24 thatthe compositions of the basalts recovered from Sites 442and 443 have been modified by sea water alteration,resulting in variably increased concentrations of K, Rb,P2O5, Sr, and to a lesser extent Ba. This limits the use ofthese elements in any petrogenetic discussions, althoughthe low abundances of K and Rb relative to Zr in someof the fresher samples do suggest that the magmas prob-ably had low K/Zr and Rb/Zr ratios originally. In addi-tion, Ba/Zr ratios are consistently low, which corrobo-rates previous evidence (Hart et al., 1974) that Ba is less
833
N. G. MARSH, A. D. SAUNDERS, J. TARNEY, H. J. B. DICK
TABLE 9Analyses, Element Ratios, and CIPW Norms for Basalts from Hole 446Aa
CoreSectionInterval (cm)
SiO2 (%)TiO2
A12O3
Fe2O3
MnOMgOCaONa2OK2OP2O5L.o.i.
Total
NiCrZnGaRbSrYZrNbBaLaCePbTh
Zr/NbTi/ZrY/ZrCe/ZrBa/ZrCeN/YN
Fetot/MgK/RbBa/Sr
0OrAbAnNeDiHy01Mt11Ap
33
35
48.34.62
14.812.100.274.489.823.051.310.530.60
99.90
80133147
2919
42344
32838
35434744
<l
9840.130.231.084.133.13
5730.84
0.07.8
25.822.90.0
18.510.60.52.18.81.3
34
33
50.73.88
12.514.230.195.409.342.420.840.440.06
99.98
72115129
2420
35939
27132
2302863
43
8860.140.230.853.973.06
3480.64
5.45.0
20.520.70.0
18.917.30.02.57.41.0
41
39
50.43.89
12.514.180.185.119.392.490.830.440.24
99.59
67105127
2515
36540
27535
2223162
42
8850.150.230.813.813.22
4580.61
5.34.9
21.220.40.0
19.516.20.02.57.41.1
41
139
50.73.90
12.614.030.225.229.452.530.810.450.24
100.06
62111123
2515
36140
27934
2192967
73
8840.140.240.784.113.12
4470.61
5.24.8
21.420.5
0.019.516.30.02.47.41.1
52
26
49.53.92
12.514.460.225.769.682.480.560.450.38
99.98
65114128
249
37540
26936
2263167
87
7870.150.250.844.112.91
5150.60
3.83.3
21.021.40.0
19.718.10.02.57.41.1
62
34
50.93.91
12.614.220.185.099.252.550.910.450.04
100.10
69115122
2421
35741
27734
2272869
21
8850.150.250.824.133.24
3600.64
5.35.4
21.520.20.0
18.916.40.02.57.41.1
71
48
51.13.82
13.113.830.155.098.342.580.900.401.11
100.43
3446
1272515
37737
25832
2152762
43
8890.140.240.834.123.15
4980.57
6.25.3
21.721.50.0
14.318.20.02.47.20.9
73
97
50.54.07
13.113.390.164.889.342.570.790.420.72
99.93
4154
1402712
39836
27031
2062866
74
990
0.130.240.764.503.18
5490.52
5.74.7
21.821.80.0
18.214.90.02.37.71.0
81
66
50.63.92
13.013.380.165.009.372.800.610.410.71
99.95
4681
10827
9388
36258
301962757
23
9910.140.220.763.893.10
5590.51
5.03.6
23.721.10.0
18.915.10.02.37.51.0
91
79
49.84.02
12.713.530.244.449.452.700.920.451.60
99.81
3339
1192613
39040
28632
2132964
33
9840.140.220.743.933.54
5840.55
4.75.4
22.919.80.0
20.213.20.02.47.71.1
92
120
50.84.23
12.813.480.205.168.282.690.920.521.88
100.96
3840
1352813
40543
30535
2323274
61
9830.140.240.764.233.03
5890.57
6.05.4
22.520.0
0.014.417.10.02.38.01.2
93
97
47.54.16
12.915.190.215.559.712.490.550.470.69
99.39
3940
12924
6409
41281
301842968
54
9890.150.240.654.073.17
7560.45
1.53.2
21.222.60.0
19.018.70.02.78.01.1
105
107
44.84.90
13.311.850.196.25
12.442.811.550.990.48
99.53
4569982422
108826
18832
3622558
4<l
6156
0.140.311.935.482.20
5860.33
0.09.2
14.819.24.9
29.80.06.92.19.32.4
mobile than K, Rb, or Sr during low-temperature altera-tion by sea water. Halmyrolytic and hydrothermal activ-ity appear to have been less important in affecting thechemistry of the Site 444 (especially unit 444-1) and Site446 basalts, where K, Rb, Ba, and Sr behave in a coher-ent fashion. This may be a function of the generallymuch higher abundances of these elements at these sites.As a consequence of the alteration at Sites 442 and 443,we shall restrict our discussion primarily to those ele-ments which are considered immobile during hydrother-mal and halmyrolitic activity—particularly Ti, Zr, Nb,Y, Ni, and Cr (e.g., Pearce and Cann, 1971, 1973;Pearce, 1975), and the REE, although noting the possi-bility of Ce mobility illustrated by Ludden and Thomp-son(1979).
Mineralogically, the basalts from Site 442 and 443and unit 444-2 closely resemble abyssal basalts recov-ered from the major ocean basins. However, one impor-tant difference is the exceptionally high vesicularity of
all the Site 442 basalts, and some of the Site 443 and 444basalts. It is not unusual to find vesicular basalts onmid-ocean ridges, but vesicles are rare at the depthsfrom which the basalts at Sites 442 (4600 meters), 443(4400 meters), and 444 (4800 meters) were recovered,because of the inverse relationship between depth of ex-trusion and degree of vesicularity. It could be suggestedthat the basalts were emplaced in shallow water, beforeundergoing subsidence to their present depth, but this isconsidered unlikely. First, the basalts would have tohave been emplaced in very shallow water (less than 500meters) to produce the observed degree of vesicularity,but the sedimentary evidence indicates that the depth ofemplacement was in fact close to the carbonate-compen-sation depth (~ 4000 meters, near the present depth).Second, at Site 443 at least, not all the basalts are vesicu-lar.
It is more likely that the high vesicularity is a result ofa high volatile content in the magmas. Two suites of ba-
834
GEOCHEMISTRY OF BASALTS
TABLE 9 - Continued
11227
48.5
2.9714.310.37
0.194.899.183.741.951.381.49
98.94
00
1122830
15532922946476348065
5780.130.352.086.78
2.46
5400.31
0.011.6
31.116.70.517.1
0.09.91.85.73.3
121
59
42.4
4.0710.312.77
0.2212.4311.251.490.740.732.88
99.33
2931137
801916
4801915024212224275
61630.130.281.41
5.431.19
3830.44
0.04.412.7
19.40.025.8
2.020.02.27.81.7
12272
41.43.60
9.413.690.27
15.439.091.130.670.694.11
99.43
4961206941916
41817
13825210234182
61560.120.301.525.92
1.03
3480.50
0.04.09.6
18.70.017.9
9.024.62.46.91.6
12436
45.6
3.7612.210.59
0.167.30
12.492.421.190.633.53
99.85
903058023217242216327304244431
6138
0.130.271.874.91
1.68469
0.42
0.07.0
17.7
19.01.6
31.7
0.08.21.87.21.5
12469
42.8
3.109.813.460.25
15.039.281.200.570.504.18
100.16
5631169861514230171292412720381
<l
5144
0.130.290.985.491.04
3410.55
0.03.4
10.119.60.018.7
11.821.7
2.35.91.2
131
109
41.4
2.269.113.680.35
17.698.120.870.25
0.365.61
99.71
6271070
90167
23316
13422172203842
6101
0.120.281.285.830.90
3020.74
0.01.57.4
20.30.0
14.315.027.22.44.30.9
14192
47.23.77
13.613.41
0.196.6210.302.650.550.351.13
99.83
77144122248
3792819823137224452
91140.140.22
0.693.862.35
5750.36
0.03.3
22.523.7
0.020.713.14.12.37.20.8
14364
49.63.61
13.013.010.17
6.339.882.690.270.341.07
100.02
72141115231
367311941912118423
<l
101120.160.220.62
3.332.38
22170.33
3.01.6
22.822.7
0.019.918.00.02.36.90.8
15170
49.63.58
13.013.100.176.329.87
2.600.450.341.22
100.30
72139105243
368301942013920392
<l
101110.150.200.72
3.192.40
1257
0.38
2.82.7
21.922.50.019.918.00.02.36.80.8
15375
49.33.60
13.213.07
0.166.359.992.500.420.341.08
100.03
72135113242
37332
20224133204152
8107
0.160.200.663.152.39
17240.36
2.82.5
21.2
23.60.019.5
18.30.02.36.80.8
16272
49.7
3.6013.212.300.195.8010.392.650.410.341.46
99.95
74131104235
374341981712820431
<l
12109
0.170.220.653.112.46
6870.34
3.42.4
22.422.80.0
21.914.80.02.16.80.8
18485
46.94.17
13.413.370.175.829.933.011.330.611.18
99.98
3848882414
67326191273342961<l2
71310.140.321.75
5.762.66
7910.50
0.07.9
25.519.20.021.71.99.92.37.91.4
18532
46.14.0813.712.840.185.9310.342.941.050.631.68
99.50
49521142412
720272003228431644
<l
6122
0.140.321.425.822.51
7280.39
0.06.2
25.021.20.0
21.71.5
10.02.27.81.5
19254
47.43.45
13.912.560.196.1810.083.010.690.421.63
99.53
761471202510
3853624322161235063
11850.150.210.663.41
2.36
5720.42
0.04.1
25.622.4
0.020.78.46.32.26.61.0
1947
49.5
3.9812.913.64
0.205.31
9.532.660.350.431.17
99.67
3113114268
4464130231228286562
10790.140.220.753.892.98
3590.51
4.82.1
22.622.20.018.616.40.02.47.61.0
salts from the East Scotia Sea back-arc basin exhibit ex-ceptionally high vesicularity (Saunders and Tarney,1979), not unlike the Site 442 basalts. It has been sug-gested (Saunders and Tarney, 1979) that this vesicularityis due to high volatile contents in the basaltic magmaand in the mantle source, possibly resulting from thepenetration of hydrous fluids from the adjacent subduc-tion zone. Subsequent studies of the East Scotia Sea ba-salt glasses have confirmed the hydrated nature of thevesicular basalts (Meunow et al., in press).
The suggestion that fluids arising from the subductedslab locally have contaminated the mantle source of theEast Scotia Sea basalts is supported by the fact that K,Rb, Sr, and Ba contents and 87Sr/86Sr ratios are cor-respondingly higher in the fresh vesicular basalts than inthe non-vesicular basalts, these elements presumably be-ing carried with fluids expelled from the slab. Unfor-tunately, because the Shikoku Basin basalts have under-gone excessive alteration by sea water, their K, Rb, Ba,
and Sr contents cannot be used as supporting evidence.Nonetheless, the high vesicularity of many of the Shiko-ku Basin basalts does indicate a high volatile content inthe source region beneath the basin. Basalts dredgedfrom the Mariana Trough, while not exhibiting the samedegree of vesicularity as the Shikoku Basin basalts, docontain significantly more H2O+ and CO2 than mid-ocean-ridge basalts (Garcia et al., 1979). Available datasuggest therefore that back-arc-basin basalts are derivedfrom a mantle source with a locally increased volatilecontent, and it may be presumed that the source forthese volatiles was an adjacent subduction zone (whichsince migrated eastward beneath what is now the BoninIslands).
It is apparent that none of the Leg 58 basalts can beconsidered as primary mantle melts: their Fe/Mg ratiosare too high, and their Cr and Ni contents too low, forthe melts to have been in equilibrium with predictedmantle compositions. Those basalts with high MgO, Ni,
835
N. G. MARSH, A. D. SAUNDERS, J. TARNEY, H. J. B. DICK
TABLE 9 - Continued
CoreSectionInterval (cm)
SiO2 (%)TK)2A12O3
F e 2 θ 3
MnOMgOCaON 3 2 OK 2OP2O5L.o.i.
Total
Ni
CrZnGaRbSrYZrNbBaLaCePbTh
Zr/NbTi/ZrY/ZrCe/ZrBa/ZrCeN/YN
Fetot/MgK/RbBa/Sr
QOrAbAnNeDiHy01Mt11Ap
201
97
47.84.16
13.114.49
0.215.82
10.162.500.200.430.94
99.86
3323
12227
2447
37302
28143
3260<l
1
1183
0.120.200.473.982.89
8220.32
2.41.2
21.224.0
0.019.718.0
0.02.57.91.0
211
19
47.13.72
13.414.27
0.186.15
10.162.390.180.391.49
99.47
3628
11727
1439
35274
26105
2560
51
1181
0.130.220.384.212.69
14720.24
1.41.1
20.425.4
0.018.819.7
0.02.57.10.9
214
79
50.93.97
13.213.19
0.165.538.922.720.860.420.54
100.36
3527
1172410
43037
29027
1742765
51
1182
0.130.220.604.312.77
7100.40
4.65.0
22.921.2
0.016.717.1
0.02.37.51.0
221
29
46.34.95
13.714.97
0.235.568.303.071.030.511.11
99.70
203
13728
9453
41378
32225
3470
54
1279
0.110.190.604.193.12
9490.50
0.06.1
26.020.6
0.014.511.2
6.02.69.41.2
222
82
47.13.72
14.99.760.244.97
11.532.971.950.352.03
99.45
2830932230
75628
21125
3322448
5<l
8106
0.130.231.574.212.28
5390.44
0.011.618.021.6
3.927.6
0.04.81.77.10.8
224
26
46.14.91
13.115.52
0.215.73
11.082.440.270.480.86
100.67
2312
13426<l
44340
29930
1312969<l
3
1098
0.130.230.444.243.14_0.30
0.01.6
20.523.8
0.023.015.6
0.32.79.31.1
232
52
50.53.47
13.411.92
0.176.449.242.740.740.401.14
100.15
93176127
2712
41337
26828
1692554
82
1078
0.140.200.633.582.15
5130.41
3.14.4
23.122.0
0.017.418.2
0.02.16.60.9
235
73
50.53.61
13.611.65
0.146.099.432.830.860.411.17
100.26
93174
9527
9409
36273
27151
2358
4<l
1079
0.130.210.553.962.22
7930.37
2.85.1
23.921.8
0.018.216.3
0.02.06.81.0
241
19
46.44.46
13.713.80
0.455.16
11.102.340.420.540.66
98.95
5064
13523
3452
38341
42252
3670
21
878
0.110.210.744.523.10
11540.56
1.22.5
20.025.8
0.021.814.5
0.02.48.61.3
244
73
48.43.20
13.513.77
0.256.28
10.422.410.340.380.56
99.50
5383
10924
3415
35251
27129
2758
71
976
0.140.230.514.072.54
9360.31
1.42.0
20.525.1
0.020.219.6
0.02.46.10.9
252
51
51.03.42
13.113.28
0.155.389.022.610.770.400.50
99.60
5150
1162612
39337
25527
1792760
41
980
0.150.240.703.982.86
5300.46
5.34.5
22.121.8
0.017.117.7
0.02.36.50.9
261
33
48.13.36
14.412.04
0.255.88
10.802.770.730.391.19
99.81
88169111
289
42335
24327
1652454
62
983
0.140.220.683.792.37
6740.39
0.04.3
23.524.6
0.021.911.1
3.02.16.40.9
263
46
49.63.17
13.312.53
0.236.19
10.862.560.280.371.09
100.08
94175102
2111
42633
22924
1022650
2<l
1083
0.140.220.453.722.35
2090.24
2.51.6
21.723.8
0.022.816.4
0.02.26.00.9
271
112
45.94.45
13.215.01
0.266.259.492.510.800.471.40
99.72
3532
12124
8431
42312
33193
3068
84
985
0.130.220.623.982.78
8290.45
0.04.7
21.322.5
0.018.014.0
4.62.68.51.1
272
94
46.14.50
13.514.44
0.226.029.572.480.860.461.57
99.74
4236
12323
8442
41307
34206
3168
6<l
988
0.130.220.674.072.78
8890.47
0.05.1
21.023.3
0.017.615.2
2.82.58.61.1
aNormative values based on Fe2θ3/FeO ratio of 0.15. Fβ2θ3 is total iron as Fβ2θ3. Cejsf/YjM represents ratio of chondrite-normalized values.
and Cr invariably contain cumulus olivine, pyroxene, orchrome spinel, probably resulting from within-unit(within-flow or within-sill) crystal settling. This is espec-ially noticeable in geochemical units 442-1, 443-3, 444-2,and 446-5. Within-sill or within-flow differentiationmay play an important role in determining the observedwithin-unit chemical variation, but it is often difficult toascertain whether pre-eruptive differentiation, followedby contemporaneous emplacement of basalts of slightlydifferent compositions as single flow or sill units, is animportant factor. Within-sill or within-flow differentia-tion may be caused by gravitational settling of mineralphases with a high specific gravity, or crystal sortingduring magma flow through the sill (or flow) body. De-tailed studies of individual sills and flows (Marsh, inprep.) suggests that in situ crystal cumulation plays animportant role.
Compositional variation between individual litholog-ic and geochemical units is probably caused by pre-erup-tive differentiation, i.e., fractional crystallization and(or) partial melting. However, proving consanguinity ofmagmas at the same site is difficult, because of uncer-tainty surrounding the time between successive magmabatches. Thus, at Site 442 for example, geochemicalunits 442-2, 3, and 4 cannot be directly related to unit 1because of the 2-m.y. interval separating the twogroups. However, the four units at Site 442 appear io beessentially comagmatic, indicating derivation from acompositionally identical mantle source, althoughdistinct partial-melting episodes rather than magmabatches from a single fractionating magma chamber areinvolved. Similarly, the basalts from Site 443 appear tohave been derived from a single mantle source, although,unlike the units of Site 442, the less-HYG/more-HYG
836
GEOCHEMISTRY OF BASALTS
446-1/J)
AlOa
• 446-1,2,4,5,9, 15
Δ 446-3O 446-6* 446-7X 446-8A 446-10T 446-11O 446-12• 446-13D 446-14
Hy
Figure 37. Normative nepheline (Ne), olivine (Ol), di-opside (Di), Hypersthene (Hy), quartz (Q) diagramfor Site 446. Normative values based on Fe2O3/FeOratio of 0.15.
100
z 50
Site 446
O 446-12
• 9
150 250Zr (ppm)
350 400
Figure 38. M versus Zr, Site 446. Separate geochemicalunits are labelled. Symbols as in Figure 37.
element ratios (e.g., Y/Zr, Th/Hf) show slight variationamong the various units (e.g., Figure 20; Wood et al.,this volume). It is considered unlikely that closed-systemfractional crystallization (Neumann et al., 1954) of abasaltic magma could produce significant changes inless-HYG/more-HYG ratios, unless the degree of frac-tionation becomes unreasonably high. Batch partialmelting may, however, change these ratios (e.g., data onLeg 49 basalts in Tarney et al., 1978), so it is reasonableto assume that the various units of Site 443 are relatedthrough different degrees of partial melting of the samesource. In addition, the compositional similarity of theSite 442 and 443 basalts suggests that the sources fromwhich the basalts were derived are very similar. This isnot unexpected in view of the fact that both sites are
250
E
α 200
100
150 200
Zr (ppm)
350
Figure 39. Ba versus Zr, Site 446. Field for Site 442from Figure 13; for Site 443 from Figure 24; and forSite 444 from Table 5.
5.0
4.0
3.0
1.0
Site 44610b A 446-9«
<0>12
50 150 2 0 0Zr (ppm)
Figure 40. TiO2 versus Zr, Site 446. Field for Site 442from Figure 7; for Site 443 from Figure 19; for Site444 from Figure 32.
located on the same mantle flow line on opposite sidesof the spreading axis.
Inter-unit variation at Sites 444 and 446 is less easilyexplained. Off-ridge basaltic material was penetrated atboth sites, and the basalts at both sites are composi-tionally heterogenous and distinct from the Site 442 and443 basalts. The within-unit variations at Site 444 maybe adequately explained by fractional crystallization insub-axis magma chambers, and perhaps also in individ-ual flow units. However, the inter-unit variation cannotbe explained by these processes, because of the widedisparity in less-HYG/more-HYG element ratios (e.g.,
837
N. G. MARSH, A. D. SAUNDERS, J. TARNEY, H. J. B. DICK
40 50
Figure 41. Ce versus Y, Site 446. Field for Site 442 fromFigure 16; for Site 443 from Figure 25; for Site 444from Figure 35.
30
10200
Zr (ppm)
Figure 42. Y versus Zr, Site 446. Fields for Sites 442,443, and 444 from Figure 33.
Y/Zr, TiO2/Zr) and more-HYG element ratios (e.g.,Ce/Zr, Zr/Nb). More important, however, is that themore-HYG-element ratios cannot be changed during thepartial melting processes which may be expected duringthe formation of basaltic magmas (Tarney et al., 1979;Wood et al., 1979; Bougault et al., 1979). Detailedstudies of basalts from the North Atlantic Ocean haveshown that variations in more-HYG-element ratios(Bougault et al., 1979; Tarney et al., 1979; Wood et al.,1979) and variations in Sr-, Pb-, and Nd-isotope ratios
Figure 43. Nb versus Zr, Site 446. Fields for Sites 442,443, and 444 from Figures 9 and 31.
Figure 44. Ce versus Zr, Site 446. Fields for Sites 442,443, and 444 from Figures 8 and 26 and Table 5,respectively.
Site 446
-
1
1442/
• / '
/ / Vs
V/
/446-5/
\
^443
/ 444-1
• / f~~& 1 ^ -
• /
446-4
^ 7 a
446-1
~~g * a*"^P A 1 0 ε446-6
/ ^
10b\ 446-2
^ X X 446-15
200
Zr (ppm)
Figure 45. P2O5 versus Zr, Site 446. Fields for Sites 442,443, and 444 from Figures 10 and 21 and Table 5,respectively.
838
GEOCHEMISTRY OF BASALTS
(e.g., O'Nions et al., 1977; Sun et al., 1979) are morelikely to have been caused by chemical heterogeneities inthe mantle source (e.g., Tarney et al., 1979). The natureand causes of these heterogeneities are still in dispute(Schilling, 1975a,b; Hart et al., 1973; White et al., 1976;Tarney et al., 1979, in press; Wood et al., 1979, inpress), but they appear to have arisen during successiveepisodes of crustal extraction and mantle metasomatism(e.g., Lloyd and Bailey, 1975; Frey et al., 1978; Pearceand Norry, 1979; Tarney et al., in press). In the ShikokuBasin, and possibly also the Daito Basin, there is the ad-ded possibility and complexity of a chemical componentbeing derived from an adjacent subduction zone.
Detailed modeling is required before partial meltingcan be ruled out as the dominant process governing theHYG-element distributions in the Site 444 and 446 ba-salts, and the variations attributed solely to mantleheterogeneity. The available data do suggest, however,that the various units at each of the two sites were de-rived from compositionally varied sources. The basaltsof unit 444-1 and Site 446 are chemically similar toevolved E-type basalts commonly found on oceanicislands (e.g., Iceland; Wood et al., in press), where suchbasalts occur in conjunction with T-type tholeiites: themantle source of the E-type basalts characteristicallyyields a mixture of enriched and depleted basalts in anyone area.
Representative analyses of the Shikoku Basin basalts,and of the basalts from the East Scotia Sea and Brans-field Strait back-arc basins, are given in Table 5; typicalanalyses of N-, T-, and E-type basalts from the AtlanticOcean are included in Table 4. Because of the effects ofpartial melting, fractional crystallization, and within-flow or within-sill differentiation, it is more valid toconsider HYG-element ratios, rather than absolute ele-ment abundances, when comparing basalt compositionsfrom different tectonic environments. In particular,ratios involving the more-HYG elements which inocean-ridge basalts are generally considered to includeLa, Ce, Nb, Ba, K, Rb, and Zr, are less responsive tofractionation processes, as mentioned before (see Woodet al., 1979).
The basalts from Sites 442 and 443, and unit 444-2,are chemically similar to basalts recovered by dredges 20and 23 (e.g., D20.43B, Table 5) in the East Scotia Sea(Saunders and Tarney, 1979). Although they tend incomposition toward N-type (depleted) ocean-ridge ba-salts, having high Zr/Nb ratios and low Ce/Y, Ce/Zr,and P/Zr ratios, they nonetheless have higher Ce/Y,Ba/Zr, Ba/Sr ratios, and probably higher K/Zr andRb/Zr ratios, than N-type ocean-ridge basalts. Equally,they have higher Zr/Nb and lower Ce/Y and Ce/Zrratios than transitional (T-type) basalts, for examplethose from 63°N on the Reykjanes Ridge (Tarney et al.,1979; Wood et al., 1979).
The transitional island-arc-tholeiite characteristicsexhibited by the basalts from dredge 24 in the EastScotia Sea (e.g., D24.14; Table 5) (high SiO2, high K/Zrand Ba/Zr ratios, and low Ni and Cr contents) are notobserved in any of the Leg 58 basalts. Island-arc tholei-ites and calc-alkaline basalts typically have low absolute
abundances of the high-field-strength (HFS) elements(Ti, P, Zr, Hf, Nb, Ta), and higher absolute abun-dances of the large-ion lithophile (LIL) elements (K, Rb,Ba, Sr, Th, U, Ce, and La) (Saunders et al., in press).Although some of the Site 444 basalts and all the Site446 basalts exhibit enrichment in the LIL elements, theyalso show significant enrichment in the HFS elements, acorrelation more characteristic of alkaline basalts thanorogenic basalts.
Lavas from Bransfield Strait, a small, young, intra-continental, back-arc basin at the north end of the An-tarctic Peninsula, exhibit chemistries which may be con-sidered transitional between ocean-ridge and calc-alka-line basalts (Weaver et al., 1979). Some of the lavashave significantly higher levels of K and Rb (BridgemanIsland) and Ba and Sr, and higher CeN/YN ratios thanthe Shikoku Basin basalts (with the exception of geo-chemical unit 444-1). Apart from their vesicularity,however, basalts recovered from the Shikoku Basin can-not be distinguished from basalts being erupted alongsome sections of ocean ridges in the major ocean basins.They do not show the extreme depletion in certain HYGelements (K, Rb, Ba, light REE, Nb) typical of N-typebasalts, and may be best considered as intermediate be-tween N- and T-type ocean-ridge basalts. Whereas theydo not show any chemical characteristics transitionaltoward orogenic basalts, their high vesicularity doessuggest that their mantle sources may have been en-riched in volatiles, possibly derived from the adjacentsubduction zone.
The HYG-element-enriched alkaline and tholeiiticbasalts from Site 446 and from unit 444-1 more closelyresemble enriched E-type basalts recovered from oceanislands and so-called "hot-spot" regions (e.g., 45°N) inthe major ocean basins (Table 4). However, there aresome significant differences, both between the 444-1and 446 basalts and between these and E-type basalts.For example, the basalts from Site 446 have muchhigher levels of TiO2 and total iron than basaltsrecovered from 45 °N on the Mid-Atlantic Ridge; theymore closely resemble HYG-element-enriched ferroba-salts commonly found on the oceanic islands (Table 4).In addition, the basalts of unit 444-1 have very highK/Rb ratios, a feature atypical of E-type basalts andtheir differentiates; nonetheless they have low Zr/Nbratios and high Ba/Zr, Ce/Y, and Ce/Zr ratios, similarto the Site 446 basalts.
It appears that the compositional range of basalts re-covered from the Shikoku and Daito Basins is as greatas for those being erupted at major oceanic spreadingcenters. Because the trace-element characteristics of ba-salts reflect, to a first approximation, those of theirmantle source regions, it may be assumed that an equaldiversity of mantle sources is available in the marginal-basin environment. This attains significance if "en-riched" mantle sources are attributed to the influence ofa deep mantle plume (e.g., Schilling, 1973), because,with the close association between marginal basins andsubduction zones, there is an inherent tectonic problemin circumventing the subducted slab. This problem maybe overcome if, for example, the "enriched" mantle
839
N. G. MARSH, A. D. SAUNDERS, J. TARNEY, H. J. B. DICK
overlies the N-type basalt mantle source, or if the differ-ent mantle types are intimately associated by a veiningprocess, as suggested by Tarney et al. (in press). In addi-tion, the marginal-basin mantle source may contain acomponent from the subducted slab (Saunders and Tar-ney, 1979). On the basis of the available chemical data,this does not appear to be the case for the Shikoku Basinsamples. However, their high vesicularity may reflect ahigh volatile content in the mantle source, possibly dueto a hydrous component arising from the subductedslab.
ACKNOWLEDGMENTSWe would like to thank Dr. G. L. Hendry for providing
analytical facilities, and for assistance in drafting the sectionon analytical techniques. We also thank Nigel C. B. Donnellanfor drawing many of the text figures, and the DSDP techni-cians for their assistance in sample and thin-section prepara-tion. N. G. M. and A. D. S. would like to acknowledge theNatural Environment Research Council, U. K., for providingfinancial support during the tenure of this project.
REFERENCESBarker, P. F., 1972. A spreading centre in the East Scotia Sea.
Earth Planet. Sci. Lett., 15, 123-132.Bass, M. N., Moberly, R., Rhodes, J. M.,Shih, C. Y., and
Church, S. E., 1973. Volcanic rocks cored in the CentralPacific, Leg 17, Deep Sea Drilling Project. In Winterer, E.L., Ewing, J. I., et al., Init. Repts. DSDP, 34: Washington(U. S. Govt. Printing Office), pp. 429-446.
Bennett, H., and Oliver, G. J., 1976. Development of fluxesfor the analysis of ceramic materials by X-ray fluorescencespectrometry. Analyst, 101, 803-807.
Bougault, H., Cambon, P., Corre, O., Treuil, M., and Joron,J.-L., 1979. Evidence for variability of magmatic processesand upper mantle heterogeneity in the axial region of theMid-Atlantic Ridge near 22°N and 36°N. Tectonophysics,55, 11-34.
Cann, J. R., 1970. Rb, Sr, Y, Zr and Nb in some ocean floorbasaltic rocks. Earth Planet. Sci. Lett., 10, 7-11.
Dewey, J. F., 1976. Ophiolite obduction. Tectonophysics. 31,93-120.
Engel, A. E., Engel, C. G., and Havens, R. G., 1965. Chemi-cal characteristics of oceanic basalts and the upper mantle.Geol. Soc. Am. Bull., 76, 719-734.
Erlank, A. J., and Kable, E. J. D., 1976. The significance ofincompatible elements in Mid-Atlantic Ridge basalts from45°N with particular reference to Zr/Nb. Contr. Mineral.Petrol., 54, 281-291.
Frey, F. A., Bryan, W. B., and Thompson, G., 1974. AtlanticOcean floor: geochemistry and petrology of basalts fromLegs 2 and 3 of the Deep Sea Drilling Project. /. Geophys.Res., 79, 5507-5527.
Frey, F. A., Green, D. H., and Roy, S. D., 1978. Integratedmodels of basalt petrogenesis. /. Petrol., 19, 463-513.
Garcia, M. O., Liu, N. W. K., and Meunow, D. W., 1979.Volatiles in submarine volcanic rocks from the Mariana Is-land arc and trough. Geochim. Cosmochim. Ada, 43,305-312.
Green, D. H., 1970. The origin of basaltic and nepheleniticmagmas. Trans. Leicester Phil. Soc, 64, 26-54.
Hart, S. R., 1971. K, Rb, Cs, Sr and Ba contents and Sr-iso-tope ratios of ocean floor basalts. Phil. Trans. Roy. Soc.London, A268, 573-588.
Hart, S. R., Glassley, W. E., and Karig, D. E., 1972. Basaltsand seafloor spreading behind the Mariana Island Arc.Earth Planet. Sci. Lett., 15, 12-18.
Hart, S. R., Erlank, A. J., and Kable, E. J. D., 1974. Seafloor basalt alteration: some chemical and strontium iso-topic effects. Contr. Mineral. Petrol., 44, 219-230.
Hart, S. R., Schilling, J.-G., and Powell, J. L., 1973. Basaltsfrom Iceland and along the Reykjanes Ridge: Sr-isotopegeochemistry. Nature, 246, 104-107.
Harvey, P. K., Taylor, D. M., Hendry, G. L., and Bancroft,F., 1973. An accurate fusion method for the analysis ofrocks and chemically related materials by X-ray fluores-cence spectrometry. X-Ray Spectrom., 2, 33-44.
Hawkins, J. W., 1974. Geology of the Lau Basin, a marginalbasin behind the Tonga arc. In Burke, C. A. and Drake, C.L. (Eds.) The Geology of Continental Margins: New York(Springer-Verlag), 505-520.
, 1976. Petrology and geochemistry of basaltic rocksof the Lau Basin. Earth Planet. Sci. Lett., 28, 283-297.
., 1977. Petrological and geochemical characteristicsof marginal basin basalts. In Talwani, M., and Pitman, W.C. Ill,(Eds.), Island Arcs, Deep Sea Trenches and Back-Arc Basins: Washington (Amer. Geophys. Union), 355-365.
Hendry, G. L., 1975. X-ray fluorescence. In Nicol, A. W.(Ed.), Physicochemical methods of mineral analysis: NewYork (Plenum Press), pp. 87-152.
Jones, J. G., 1969. Pillow lavas as depth indicators. Am. J.Sci., 267, 181-195.
Karig, D. E., 1971. Origin and development of marginal basinsin the western Pacific. J. Geophys. Res., 76, 2542-2561.
, 1974. Evolution of arc systems in the western Pacif-ic. In Annual Review of Earth and Planetary Sciences, 2,51-75.
., 1975. Basin genesis in the Philippine Sea. In Karig,D. E. and Ingle, J. E., et al., Init. Repts. DSDP, 31: Wash-ington (U. S. Govt. Printing Office), pp. 857-879.
Karig, D. E., Ingle, J. E., et al., 1975. Init. Repts. DSDP, 31:Washington (U. S. Govt. Printing Office).
Katsumata, M., and Sykes, L. R., 1969. Seismicity and tecton-ics of the western Pacific: Izu-Mariana-Caroline and Ryu-kyu-Taiwan region. /. Geophys. Res., 74, 5923-5948.
Kay, R. W., Hubbard, J. J., and Gast, P. W., 1970. Chemicalcharacteristics and origin of oceanic ridge volcanic rocks. /.Geophys. Res., 75, 1585-1613.
Kobayashi, K., and Isezaki, N., 1976. Magnetic Anomalies inthe Sea of Japan and the Shikoku Basin: Possible tectonicimplications. In Sutton, G. H., Manghani, M. H., Mober-ly, R. (Eds.), The Geophysics of the Pacific Ocean Basinand Its Margin. Am. Geophys. Union Monogr., 19.
Kobayashi, K., and Nakata, M., 1977. Local magnetic anom-aly profiles, Shikoku Basin, Northwestern Pacific Ocean.Contr. Geodynamics Project Japan, 77-2.
LaBrecque, J. L., Kent, D. V., and Cande, S. C , 1977. Re-vised magnetic polarity time scale for Late Cretaceous andCenozoic time. Geology, 5, 330-335.
Leake, B. E., Hendry, G. L., Kemp, A., Plant, A. G., Harvey,P. K., Wilson, J. R., Coats, J. S., Aucott, J. W., Lunel,T., and Howarth, R. J., 1969. The chemical analysis ofrock powders by automatic X-ray fluorescence. Chem.Geol, 5, 7-86.
Lloyd, F. E., and Bailey, P. K., 1975. Light element metaso-matism of the continental mantle: the evidence and the con-sequences. In Ahrens, L. H., Press, F., Runcorn, S. K. andUrey, H. C. (Eds.), Physics and Chemistry of the Earth,(Vol. 9): New York (Pergamon Press), pp. 389-416.
840
GEOCHEMISTRY OF BASALTS
Ludden, J. N., and Thompson, G., 1979. An evaluation of thebehaviour of the rare earth elements during the weatheringof sea floor basalt. Earth. Planet. Sci. Lett., 43, 85-92.
Meunow, D. W., Lui, N. W. K., Garcia, M. O., and Saunders,A. D., in press. Volatiles in submarine volcanic rocks fromthe spreading axis of the East Scotia Sea back-arc basin.Earth Planet., Sci. Lett.
Mitchell, W. S., and Aumento, F., 1977. Uranium in oceanicrocks: Deep Sea Drilling Project Leg 37. Can. J. Earth Sci.,14, 794-808.
Mizuno, A., Okuda, Y., Tamaki, K., Kinoshita, Y., Nohara,M., Yuasa, M., Nakajima, M., Murakami, F., Terashima,S., and Ishibashi, K., 1975. Marine geology and geologicalhistory of the Daito Ridge area, northwestern PhilippineSea. Marine Sci., 7, 484-491, 543-548.
Mizuno, A., Okuda, Y., Nagumo, S., Kagami, H., and Nasu,N., in press. Subsidence of the Daito Ridge and associatedbasins, North Philippine Sea.
Moore, J. G., 1965. Petrology of deep-sea basalt near Hawaii.Am. J. Sci., 263, 40-52.
Murauchi, S., Asanuma, T. and Saki, K., 1976. Seafloorspreading in the Shikoku Basin, South of Japan. Abstr.Progr., Int. Ewing Symp., Harriman, New York.
Murauchi, S., Den, N., Asano, S., Hotta, H., Yoshii, T., Asa-numa, T., Hagiwara, K., Ichikawa, K., Sata, R., Ludwig,W. J., Ewing, J. I., Edgar, N. T., and Houtz, R. E., 1968.Crustal structure of the Philippine Sea. J. Geophys. Res.,73, 3143-3171.
Neumann, H., Mead, J., and Vitaliano, C. J., 1954. Trace ele-ment variation during fractional crystallization as calcu-lated from the distribution law. Geochim. Cosmochim. Ac-ta, 6, 90-100.
O'Hara, M. J., 1973. Non-primary magmas and the dubiousmantle plume beneath Iceland. Nature, 243, 507-508.
, 1977. Geochemical evolution during fractional crys-tallization of a periodically refilled magma chamber. Na-ture, 226, 503-507.
O'Nions, R. K., Hamilton, P. J., and Evensen, N. M., 1977.Variations in 143Nd/144Nd and 87Sr/86Sr ratios in oceanicbasalts. Earth Planet. Sci. Lett., 34, 13-22.
CTNions, R. K., Pankurst, R. J., and Gronvold, K., 1976. Na-ture and development of basalt magma sources beneath Ice-land and the Reykjanes Ridge. /. Petrol., 17, 3-30.
Pearce, J. A., 1975. Basalt geocemistry used to investigate pasttectonic environments on Cyprus. Tectonophysics, 25, 41-67.
Pearce, J. A., and Cann, J. R., 1971. Ophiolite origin investi-gated by discriminant analysis using Ti, Zr and Y. EarthPlanet. Sci. Lett., 12, 339-349.
, 1973. Tectonic setting of basic volcanic rocks usingtrace element analyses. Earth Planet. Sci. Lett., 19, 290-300.
Pearce, J. A., and Norry, M. J., 1979. Petrologic implicationsof Ti, Zr, Y and Nb variations in volcanic rocks. Contrib.Mineral. Petrol., 57.
Reynolds, R. C , 1967. Estimation of mass absorption coeffi-cients by Compton scattering: improvements and exten-sions of the method. Am. Mineral., 52, 1493-1502.
Rhodes, J. M., Blanchard, D. P., Rodgers, K. V., Jacobs, J.W., and Brannon, J. C , 1976. Petrology and chemistry ofbasalts from the Nazca Plate: Part 2—Major and trace ele-ment chemistry. In Hart, S. R., Yeats, R. S., et al., Init.Repts. DSDP, 34: Washington (U. S. Govt. Printing Of-fice), pp. 239-244.
Saunders, A. D., and Tarney, J., 1979. The geochemistry ofbasalts from a back-arc spreading centre in the East ScotiaSea. Geochim. Cosmochim. Ada, 43, 555-572.
Saunders, A. D., Tarney, J., Stern, C , and.Dalziel, I. W..D.,1979. Geochemistry of Mesozoic marginal basin floor ig-neous rocks from southern Chile. Geol. Soc. Amer. Bull.,90, 237-258.
Saunders, A. D., Tarney, J., and Weaver, S. D., in press.Transverse variations in chemistry across the Antartic Pen-insula: implications for calc-alkaline magma genesis. EarthPlanet. Sci. Lett.
Schilling, J.-G., 1971. Seafloor evolution: rare earth evidence.Phil. Trans. Roy. Soc. Lond., A268, 663-706.
, 1973. Iceland mantle plume: geochemical study ofReykjanes Ridge. Nature, 242, 565-575.
, 1975a. Azores mantle blob: rare earth evidence.Earth Planet. Sci. Lett., 25, 103-115.
, 1975b. Rare earth variations across "normal seg-ments" of the Reykjanes Ridge, 60 N-53 N, Mid-AtlanticRidge, 29 S, and East Pacific Rise, 2 S-19 S, and evidenceon the composition of the Low-velocity layer. J. Geophys.Res., 80, 1459-1473.
Shido, F., Miyashiro, A., and Ewing, M., 1971. Crystalliza-tion of abyssal tholeiites. Contr. Mineral. Petrol, 31, 251-266.
Sun, S.-S., Nesbitt, R. W., and Sharaskin, A. Ya., 1979.Chemical characteristics of mid-ocean ridge basalts. EarthPlanet. Sci. Lett., 29.
Tarney, J., Saunders, A. D., Weaver, S. D., Donnellan, N. C.B., and Hendry, G. L., 1978. Minor element geochemistryof basalts from Leg 49, North Atlantic Ocean. In Luyen-dyk, B. P., Cann, J. R., et al., Init. Repts. DSDP, 49:Washington (U. S. Govt. Printing Office), pp. 657-691.
Tarney, J., Wood, D. A., Saunders, A. D., Varet, J., andCann, J. R., 1979. Nature of mantle heterogeneity in theNorth Atlantic: evidence from Leg 49. In Talwani, M.(Ed.), Results of Deep Sea Drilling in the Atlantic. MauriceEwing Series 2 (Amer. Geophys. Union).
Tarney, J., Wood, D. A., Saunders, A. D., and Cann, J. R.,in press. Nature of mantle heterogeneity in the North At-lantic: evidence from deep sea drilling. Phil. Trans. Roy.Soc. London.
Tomada, Y., Kobayashi, K., Segawa, J., Nomura, M., Kimu-ra, K., and Saki, T., 1975. Linear magnetic anomalies inthe Shikoku Basin, northeastern Philippine Sea. /. Geo-magnet. Geoelec, 28, 47-56.
Watts, A. B., and Weissel, J. K., 1975. Tectonic history of theShikoku marginal basin. Earth Planet. Sci. Lett., 25,239-250.
Watanabe, T., Epp, D., Uyeda, S., Langseth, M., and Yasui,M., 1970. Heat flow in the Philippine Sea. Tectonophysics,10, 205-224.
Weaver, S. D., Saunders, A. D., Pankhurst, R. J. and Tarney,J., 1979. A geochemical study of magmatism associatedwith the initial stages of back-arc spreading. The quater-nary volcanics of Bransfield Strait, from South Shetland Is-lands. Contr. Mineral. Petrol., 68, 151-169.
Weaver, S. D., Sceal, J. S. C , and Gibson, I. L., 1972. Trace-element data relevant to the origin of trachytic and pantel-leritic lavas in the East African Rift system. Contr. Min-eral. Petrol., 36, 181-194.
White, W. M., and Schilling, J.-G., 1978. The nature and ori-gin of geochemical variations in Mid Atlantic Ridge basaltsfrom the central North Atlantic. Geochim. Cosmochim.Acta, 42, 1501-1516.
White, W. M., Schilling, J.-G. and Hart, S. R., 1976. Evi-dences for the Azores mantle plume from strontium isotopegeochemistry of the Central North Atlantic. Nature, 263,659-663.
841
N. G. MARSH, A. D. SAUNDERS, J. TARNEY, H. J. B. DICK
Wood, D. A., Tarney, J., Varet, J., Saunders, A. D., Bou- Wood, D. A., Joron, J.-L., Treuil, M., and Tarney, J., ingault, H., Joron, J.-L., Treuil, M., and Cann, J. R., 1979. press. Elemental and Sr isotope variations in basic lavasGeochemistry of basalts drilled in the North Atlantic by from Iceland and the surrounding ocean floor: the natureIPOD Leg 49: implications for mantle heterogeneity. Earth of mantle source inhomogeneities. Contr. Mineral. Petrol.Planet. Sci. Lett., 42, 77-97.
842