deep sea drilling project initial reports volume 69 · 2007. 5. 4. · deep sea drilling project...
TRANSCRIPT
49. TRACE ELEMENT GEOCHEMISTRY OF BASALTS FROM HOLE 504B, PANAMA BASIN,DEEP SEA DRILLING PROJECT LEGS 69 AND 701
N. G. Marsh and J. Tarney, University of Leicester, Leicester, LEI 7RH, United Kingdomand
G. L. Hendry, University of Birmingham, Birmingham, B15 2TT, United Kingdom
ABSTRACT
Deep basement penetration during Legs 69 and 70 at Hole 504B in the Panama Basin allowed the recovery of a561.5-meter sequence of basaltic pillows, thin flows, and breccias interspersed with thick massive flows. The lavas,which are aphyric to moderately plagioclase-olivine-clinopyroxene phyric, are petrologically indistinguishable fromtypical mid-ocean-ridge basalts (MORB). Some units are distinctive in that they carry accessory chrome-spinel micro-phenocrysts or emerald green clinopyroxene phenocrysts.
Major and trace element analyses were carried out on 67 samples using X-ray fluorescence techniques. The basaltsresemble normal MORB in terms of major elements. However, the trace element analyses show that most of the basaltsare characterized by very strong depletion in the more incompatible elements compared with, for instance, normal (Ntype) MORB from the Atlantic at 22°N. Interdigitated with these units are one or two units that have distinctly higherincompatible element concentrations similar to those in basalts of the transitional (T) type from the Reykjanes Ridge(63°N in the Mid-Atlantic Ridge).
All the basalts appear to have undergone some high-level crystal fractionation, although this has not proceeded tothe extent of yielding ferrobasalts as it has at the adjacent Galapagos Spreading Center or along the East Pacific Rise.The magnetic anomalies are of lower amplitude than in the latter two regions, which suggests that the absence of ferro-basalts may be a general feature of the ocean crust generated at the Costa Rica Rift. The presence of two distinctmagma types, one strongly depleted and the other moderately enriched in incompatible elements, suggests that magmachambers at the spreading center are discontinuous rather than continuous and that there is some chemical heterogene-ity in the underlying mantle source. Observed variations in incompatible element ratios of basalts from the moredepleted group could, however, reflect mixing between these two magma types. In general it would appear that themantle feeding the Costa Rica Rift is significantly more depleted in incompatible trace elements than that feeding theMid-Atlantic Ridge.
INTRODUCTION
Legs 68, 69, and 70 of the Deep Sea Drilling Projectwere undertaken principally to study the problems ofhydrothermal circulation in young oceanic crust. Thedeep penetration and good recovery of basaltic base-ment, particularly at Site 504, for these hydrothermalstudies also provided the opportunity to study the igne-ous geochemistry and petrogenesis of the basalts in thePanama Basin. This part of the Panama Basin is flooredby young ocean crust produced by spreading at the CostaRica Rift (Fig. 1). The rift is a presently active oceanfloor spreading center some 200 km long; it lies midwaybetween the east-west-trending Carnegie Ridge and thenortheast-southwest-trending Cocos Ridge and is ap-proximately 500 km from the Colombian coast (Fig. 1).The Costa Rica Rift is the easternmost section of theGalapagos Rift System, which joins the East PacificRise at the Galapagos triple junction. The Costa RicaRift is bounded on the west and east by the Ecuador andPanama fracture zones, respectively, two of several ma-jor transforms that form offsets on the Galapagos RiftSystem (Lonsdale and Klitgord, 1978).
Previous studies in this area of the eastern PacificOcean have established that the basalts that have erupt-
1 Cann, J. R., Langseth, M. G., Honnorez, J., Von Herzen, R. P., White, S. M., et al.,Init. Repts. DSDP, 69: Washington (U.S. Govt. Printing office).
ed along the Cocos-Nazca Ridge, in the GalapagosSpreading Center, and along both the adjacent limbs ofthe East Pacific Rise are predominantly geochemicallyevolved basalts, in particular ferrobasalts (e.g., Clagueand Bunch, 1976; Schilling et al., 1976; Vogt, 1979; Ba-tiza and Johnson, 1980; Rosendahl et al., 1980). Theocean floor generated by spreading at these ridges is to-pographically smooth, with well developed high ampli-tude magnetic anomalies that are thought to reflect thepresence of titanomagnetite-rich ferrobasalts (Andersonet al., 1975; Vogt, 1979; Vogt and Johnson, 1973). ThePanama Basin is similarly floored by smooth ocean crust,with well developed magnetic anomalies up to 7.1 m.y.old and a moderate total spreading rate of 7.2 cm per yr.However, although the magnetic lineations in this basinare well developed, the magnetic anomalies are of loweramplitude than those associated with the presence offerrobasalts.
During DSDP Legs 68, 69, and 70, eight holes weredrilled at two different sites on the southern flank of theCosta Rica Rift. Sites 501 and 504 are essentially thesame site, Site 504 being offset 0.5 km to the east ofHole 501. The sites are at 1°13.64'N, 83°43-89'W, ap-proximately 250 km south of the Costa Rica Rift. Theyare on a negative magnetic anomaly between Anomalies3A and 4 in ocean crust about 5.9 m.y. old. Five holeswere drilled: Holes 501, 504, 504A, 504B, and 504C.Major basement penetration occurred in Hole 504B, in
747
N. G. MARSH, J. TARNEY, G. L. HENDRY
5*N
5°S
Siqueiros Fracture
C. R. R. Costa Rica RiftE. F.Z. = Ecuador Fracture ZoneP.F.Z. = Panama Fracture Zone
1000 w 77° w
Figure 1. Location of drill sites from Legs 68, 69, and 70 plus previous DSDP drill sites in the vicinity ofthe Costa Rica Rift. Adapted from van Andel et al. (1973).
which 561 meters of basalt were cored. The basalt con-sisted of highly fractured pillows, flows, and brecciasthat were interspersed with massive flows. The overlyingsediment was 264.5 meters deep and consisted mainly ofpelagic siliceous oozes, chalks, limestones, and cherts.Site 505 is at 1°55.2'N, 83°47.3'W, almost 100 kmnorth of Site 504 and between it and the Costa RicaRift. Basement penetration and recovery were low atSite 505, and little material became available for detailedwork.
This study concentrates on the characteristics of thematerial recovered from Hole 504B; initial shipboardpetrographic and chemical data suggest that basalts fromSites 501 and 504 share the same petrographic and geo-chemical features, and recovery and penetration werebetter there than at other holes. The trace and minor ele-ment geochemistry of selected samples from Hole 504Bis described, and within-hole chemical variations areevaluated with a view to developing an understanding ofthe genesis of the basalts.
IGNEOUS LITHOLOGY AND PETROGRAPHY OFHOLE 504B
The lithostratigraphy, igneous, and alteration petrog-raphies of the basement section cored by Hole 504B aredescribed in detail in the Site 504 chapter (this volume).The pertinent features are summarized below and inFigures 2 and 3. Three main petrographic groups of ba-salts are recognized:
1) Aphyric to moderately phyric plagioclase-olivinebasalts, some with glomerophyric clumps of Plagioclaseand augite. Plagioclase is the most commonly observedphenocryst phase. The phenocrysts are generally small(0.2-0.7 mm long); however, a rare second populationof slightly larger phenocrysts is present in some units.
Compositional zoning is poorly developed, although re-sorption affects a few small phenocrysts.
2) Variably plagioclase-olivine phyric basalts withaccessory chrome-spinel phenocrysts. The Plagioclasephenocrysts often have well developed normal or oscil-latory zoning; they often carry inclusions of glass andoccasionally carry anhedral Plagioclase laths. Olivine isfound as phenocrysts or as granular or skeletal forms inthe groundmass. Similarly, chrome-spinel phenocrystsare present as euhedral inclusions in Plagioclase or oliv-ine phenocrysts; skeletal varieties can also be found.Chrome-spinel is also present in the mesostasis.
3) Plagioclase-olivine-clinopyroxene phyric basalts.These basalts are notably more phenocryst rich (15-20%phenocrysts) than the other two petrographic types. Pla-gioclase phenocrysts with vaguely oscillatory zoned coresand more albitic rims are common. Rare glass inclusionsand small anhedral Plagioclase laths are found in Plagio-clase phenocrysts. Both Plagioclase and the characteris-tic emerald green clinopyroxene crystals of this petro-graphic type commonly show signs of resorption.
Most of the basalt units described are aphyric to mod-erately plagioclase-olivine phyric; the chrome-spinel-and emerald green clinopyroxene-bearing basalts areless common. Vesicles are rare except in Lithologic Units5 and 36 (Fig. 2); however, small, irregular miaroliticvoids occur in some units.
The basalts are generally weakly to moderately al-tered. Clinopyroxene and Plagioclase crystals are usu-ally fresh in appearance, although clay minerals havebeen identified in the cleavage fractures of some Plagio-clase phenocrysts. The most noticeable alteration effectsare the replacement of olivine and mesostasis by clayminerals. A scheme of nonoxidative alteration is in-ferred from the shipboard observations; it is charac-
748
TRACE ELEMENT GEOCHEMISTRY OF BASALTS
Lithology Unit Petrography 1.0Fe/Mg
Plagioclase—ol•vine
Chrome—spinel-bearing
Plagioclase—olivine
microphyric
Chrome—spinel-bearing
Plagioclase—clinopyroxene
olivine
Plagioclase-olivine
microphyric
Chrome—spinel
Plagioclase-olivine
Aphyric
Plagioclase—clinopyroxene-
oliviπe
" T V
Cr (ppm)200 400
I i I
7
Zr (ppm)70
J I L
Magnetic Inclination
(deg)9 0 - 4 0 0 +40
I . I I I . I l _
M,
Figure 2. Downhole variation in lithology, magnetic inclinations, and selected geochemical parameters forthe basement section of Hole 504B cored during Leg 69.
terized by the replacement of olivine and filling of vesi-cles by saponite or (in regions of higher temperature) atalc-like mineral followed by Fe-Mg-rich smectites. Laterstages of alteration appear to have taken place under ox-idizing conditions, but the effects are more limited in ex-tent than those of the earlier nonoxidative phase of al-teration and result in the gradual conversion of smectiteto iron oxyhydroxides and titanomagnetite to maghem-ite.
Pyrite, carbonates, and clay minerals can be found,especially in the chrome-spinel-bearing units, wheresome of the pyrite may be of primary igneous origin.Calcite appears to be limited to the units above 600 me-ters sub-bottom (Unit 27); below this unit pyrite is moreabundant, particularly in vein fillings, and evidence foroxidative alteration diminishes.
The extent of alteration varies with lithology; themargins of pillow lavas and flows are more intensely al-tered than the more crystalline interiors, where observ-able alteration effects are generally limited to the areasimmediately adjacent to fractures. Most of the samplesanalyzed in this study were selected from the relativelyfresh pillow and flow interiors of the different units.
ANALYTICAL TECHNIQUES
Minicores or quarter-core sections were washed in distilled waterto remove surface contamination and dried at 110°C overnight priorto being ground to a fine powder (less than 60 µm) in an Agate TEMAswingmill. Powder briquettes 46 mm in diameter were made by bind-ing 15 g of rock powder with about 20 drops of a 7% aqueous solutionof polyvinyl alcohol under a pressure of 15 tons per square inch in asteel die with polished hardened surfaces.
X-ray fluorescence (XRF) analysis was performed on the Leg 69samples by using a Philips PW1450 automatic spectrometer with aPW1466 60 position sample changer. Count data from the spectrom-eter were processed by using a 32K Digital PDP 11-03 computer andthe University of Birmingham's ICL 1906A computer. The major, mi-nor, and trace elements were determined on the 46-mm briquettes us-ing rhodium (Si, Ti, Al, Fe, Mn, Mg, Ca, Na, K, and P on the majorelement program; Zr, Nb, Ni, and Cr on the trace element program),molybdenum (Y, Sr, Rb, Th, Pb, Ga, Zn, and Ba), and tungsten (Ce,La, and Nd) anode X-ray tubes. Instrumental repeatability was en-sured by measuring a reference sample for major and minor elementsafter every fourth sample or at the beginning of every loading of 60samples for trace elements.
Details of the analysis of major and minor elements using Rhanode excitation are given in Marsh et al. (1980). The analysis of traceelements (Y, Sr, Rb, Th, Pb, Ga, Zn, Ba) by Mo anode excitation andof Ce and La using W anode excitation is described in Tarney et al.(1979). For the analysis of the Leg 69 samples, Nd was added to the Wanode program. The program used the Lα peak, a LiF22o crystal,coarse collimation (0.55 mm), and a flow proportion detector (to mea-sure argon/methane). The mutual overlap of Nd, Lα, and Ce L ßlt 4was corrected. The remaining elements (Zr, Nb, Ni, and Cr) weredetermined by using Rh anode excitation, which provides improvedsensitivity when compared with the W anode previously employed. Zrand Nb have a statistical lower limit of detection (2σ above back-ground) of 0.15 ppm for a 100-s counting time; the previous limit ofdetection was 1 ppm. The practical limit of determination is still of theorder of 1 ppm, however. The Rh tube was operated at 70 kV and 30mA to excite the ‰ wavelengths. Fine collimation (0.15 mm) and aLiF22o crystal were used, as well as the scintillation detector for Nband Zr and the flow proportion detector for Cr and Ni.
Trace element calibrations were initially prepared from a widerange of rock types and pure silica spiked with an appropriate pureelement compound (Leake et al., 1969). Total mass absorption wasthen corrected by using the Mo, K,,, and Rh K,, Compton scatter linesfor elements with emission wavelengths shorter than the iron absorp-tion edge (Reynolds, 1967). For the analysis of Cr, iron counts and RhCompton intensity were used to estimate the absorption coefficents on
749
N. G. MARSH, J. TARNEY, G. L. HENDRY
Core Lithology Unit
38
63
23
Plagioclase—oli vine—clinopyroxene
_46_
Petrography 1.0Fe/Mg
1.8
Aphyric
Olivine—Plagioclase
Chrome—spinel
Plagioclase—olivine
Chrome-spinel
Plagioclase—olivine
Plagioclase—olivine
Plagioclase—olivine—
clinopyroxene
Plagioclase—olivine
Plagioclase—ol ivine
Vesicular,aphyric
Olivine
Plagioclase—olivine
Plagioclase—olivine
Plagioclase—olivine
Plagioclase—olivine
Aphyric
Olivine—clinopyroxene
Cr (ppm)200 400
J i |_
Zr (ppm)30 90
Figure 3. Downhole variation in lithology and selected geochemical parameters for the basement sectionof Hole 504B cored during Leg 70.
the long wavelength side of the iron absorption edge. For Zr, Nb, Ni,and Cr, the Rh Compton peak gave an improved estimate of the ab-sorption coefficient when compared to the WL^ Rayleigh peak previ-ously used. The peak is also clear of overlapping lines and is thereforesuperior to the Mo K̂ Compton scatter lines, which must be correctedfor Y interference. Ce, La, and Nd calibrations were obtained by us-ing basalts previously analyzed by isotope dilution, because no im-provement resulted from absorption corrections using tube scatterlines.
Trace element precision is high, as illustrated by Table 1, whichshows the average concentrations of the internal reference BOB-1measured on 23 consecutive days. However, a potential problem inprojects that extend over a period of a year or more is long term driftin instrumental accuracy. This was minimized by using some 40 sam-ples from Legs 49 and 58 together with other basaltic material that wasselected to cover as wide a concentration range as possible for all theelements measured. These samples were analyzed together with the
unknowns to ensure measurement consistency between different sam-ple batches.
Trace element analyses on the Leg 70 samples were also performedon the Philips PW1450 spectrometer at the University of Birminghamusing the techniques described above. However, major element analy-ses were performed on a Phillips PWI400 automatic X-ray spectrom-eter with a 72-position sample changer at the University of Leicester.Techniques similar to those described above were used to ensure con-sistency with the Leg 69 data. Count data from the PW1400 spectrom-eter were processed on line using a Phillips P851 minicomputer andX-14 software.
Values for the international reference samples BCR-1 and JB-1 aregiven in Table 1 to give some estimate of accuracy, although as yetbasic reference materials do not cover the range of concentrationsfound in ocean-floor basalts.
Unfortunately, several of the trace elements in the Leg 69 and Leg70 basalts have very low abundances that are close to, or lower than,
750
TRACE ELEMENT GEOCHEMISTRY OF BASALTS
Table 1. Analyses of standard reference samples.
a. Trace element analyses.
Cr c (ppm)NiRbSr
Y dZr e
NbBa f
LaCeSNd^PbThZnGa
BOB-la
251 ± 9103.4 ± 1.7
4.8 ± 0.5201.3 ± 0.926.8 ± 0.4
108.5 ± 1.14.8 ± 0.6
50.6 ± 3.46.7 ± 1.0
14.9 ± 1.310.4 ± 0.94.2 ± 1.30.2 ± 1.0h
62.2 ± 2.217.5 ± 0.9
Sample
BCR
181847
32936
19212
690255528166
12024
JB-1
35913442
44223
15035
50038632010118017
BOB-1
255114
619725
1074
533
121062
6814
a Mean concentrations determined on 23 consecutivedays, with lσ standard deviation during this study(Univ. Birmingham).
° Concentrations determined during this study (Univ.Leicester).
c Not corrected for V interference." Corrected for Rb interference.e Corrected for Sr interference.* Corrected for Ce interference.£ Corrected for mutual interference.
Below statistical lower limit of detection.
b. Major element analyses for BOB-1
SiO 2 (°/o)TiO2
A12O3,F e 2 O 3
d
MnOMgOCaONa2OK 2 O
P2O5
Total
Medium of Analysis
PowderBriquettea
50.91.32
16.18.60.147.6
11.283.10.350.128
99.52
PowderBriquetteb
50.81.31
15.39.140.1477.2
10.933.180.360.141
98.51
Fusion Beadc
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
100.07
a Values determined on a 46-mm-diameter powder bri-quette during this study (Univ. Birmingham).
" Values determined on a 46-mm-diameter powder bri-quette during this study (Univ. Leicester).
c Mean anhydrous major element composition deter-mined on six fusion beads with lσ sigma standarddeviation.
" Fe 2 θ3 is total iron determined as Fe 2 θ3
the theoretical lower limit of detection. At these low abundance levels,precision falls dramatically, and abundances near or below the lowerlimit of determination for a particular element should be treated withcaution. The elements most affected are Nb (1 ppm), La ( 5 ppm),Ce ( 8 ppm), Nd ( 8 ppm), Pb ( 5 ppm), and Th ( 10 ppm).
We have not made a serious attempt to provide major element databy fusion techniques on the Leg 69 and Leg 70 samples. However, toprovide additional data for interpreting our trace and minor elementresults, major element analyses were carried out on the powder bri-
quettes. Fortunately, the samples have similar bulk chemistries, limit-ing potential variations to mineralogical effects. Al is expected to bethe most aberrant element, as well as (to a lesser extent) Fe and Mg.For comparison, the various major element analyses performed on theinternal reference sample BOB-1 are listed in Table 1.
TRACE AND MINOR ELEMENT TERMINOLOGY
Trace and minor elements may be either compatibleor incompatible with respect to rock-forming minerals.For convenience in petrogenetic discussions, incompati-ble elements can be grouped semiquantitatively intoless-hygromagmatophile (HYG) and more-hygromag-matophile types. Less-HYG elements are those with abulk crystal/liquid distribution coefficient (D) of ap-proximately 0.1. More-HYG elements have lower D val-ues, often significantly less than 0.01. In the relativelyanhydrous oceanic basalt systems, the more-HYG ele-ments include (in order of increasing D) Cs, Rb, Ba, U,Th, K, Ta, Nb, La, and Ce. The less-HYG elementsinclude Sr, Nd, P, Hf, Zr, heavy rare earth elements(HREE), Ti, and Y (Bougault et al., 1980; Bougault andTreuil, 1980; Saunders et al., 1980; Wood, Joron, et al.,1979). D values can vary according to the P-T-X condi-tions of the system considered; for example, during Pla-gioclase crystallization, D values for Sr and Eu sig-nificantly increase and can exceed unity.
Trace and minor elements may also be grouped ac-cording to their ionic character. Large ions with lowcharge and ions with the ability to form hydrationspheres (giving a low effective charge density) have beentermed large ion lithophile (LIL) (Schilling, 1973) or lowfield strength ions (LFS) (Saunders et al., 1980). Theyinclude Cs, Rb, K, Ba, U, Th, and Sr, and to a lesser ex-tent La and Ce. Several of these elements are easily mo-bilized during hydrothermal or ambient seawater alter-ation, particularly Cs, Rb, K, and U (Hart, 1971; Hartet al., 1974; Mitchell and Aumento, 1977). During ex-tensive halmyrolysis, even La, Ce (Luddon and Thomp-son, 1979) and possibly Th are mobilized. Small ionswith high charges are termed high field strength (HFS)ions and include Ta, Nb, P, Zr, Hf, the HREE, Ti, andY. They are generally less susceptible to alteration pro-cesses and hence more likely to reflect igneous abun-dances (Pearce and Norry, 1979). These classificationsare arbitrary, of course, because variations in both Dvalues and charge density are gradational (Saunders etal., 1980).
RESULTS
The results of the chemical analyses are listed in Ta-bles 2 and 3. All the basalts are similar in terms of majorelement chemistry, with 7.5 to 9.5% MgO, 9 to 11%total iron (as Fe2O3), and low Fe/Mg ratios. Figures 4and 5 illustrate some of the main features of the majorelement chemistry. Although Fe/Mg ratio varies to someextent with alteration, it is still useful as a parameter forillustrating major element variations, especially for fresh-er material.
MgO and total iron (Fe2O3) contents correlate wellwith Fe/Mg, and apart from the samples from Litholog-ic Unit 1 (solid circles, Fig. 4), scatter is limited. CaOand Na2O contents appear to be reasonably constant
751
N. G. MARSH, J. TARNEY, G. L. HENDRY
Table 2. Analyses and element ratios for basalts from Hole 504B recovered during Leg 69.
CoreSectionInterval (cm)Piece
31
131-134257
42
114-117307
52
68-71379
72
27-30454
74
84-87502
82
84-87530
91
48-51572
102
26-29629
112
134-138693
122
65-69736
132
109-112782
141
144-148858
153
108-112972
163
45-50992
164
14-191004
171
57-611039
181
63-661089
S:O2 (%)TiO2
A12O3
F e 2 θ 3 a
MnOMgOCaONa2OK2OP 2 θ 5
Total
Ni (ppm)CrZnGaRbSrYZrNbBaLaCeNdPbTh
Fe/MgZr/NbTi/ZrCe N /Y N
Lithologic Unit
48.81.15
14.612.20.188.8
11.842.210.1110.088
99.98
6619484154
752853
< l12
1572
< l
1.61
1300.4
1
49.60.89
15.49.80.199.2
12.552.050.0270.045
99.75
136405
64142
652244
< l6
< l< l
4< l< l
1.23
121—
2
50.10.89
15.89.80.178.2
12.782.080.1330.040
99.99
138424
66154
702546
< l7
< l< l
442
1.38
116—
2
49.50.92
15.89.90.178.5
12.712.210.0800.048
99.84
122403
73152
642347
< l9124
< l2
1.34
1170.2
2
49.40.90
15.59.20.218.5
12.492.080.0040.045
98.33
139383
67142
592246
< l5
< l1431
1.25
1170.1
2
48.10.88
15.88.50.158.6
12.132.150.0050.064
96.38
123378
6114
1612243
< l9
< l551
< l
1.14
1230.5
2
49.50.89
15.89.10.179.0
12.612.000.0070.004
99.12
151395
68142
662145
< l6114
< l2
1.16
1190.1
2
47.90.98
14.19.60.158.3
12.212.080.1360.046
95.50
107325
68144
742450
< l31362
< l
1.34
117—
3
49.00.99
14.810.20.178.1
12.132.130.3280.054
97.90
833096614
8752349
< l3
< l552
< l
1.47
121
—
3
50.61.01
15.29.00.189.0
12.412.230.1270.051
99.81
12232474154
762450
< l15
< l56
< l1
1.16
121
-
3
49.71.02
15.39.10.158.6
12.232.420.1740.063
98.78
93337
7113
5772450
< l8
< l56
< l< l
1.22
122
—
3
49.00.94
14.710.70.188.1
12.302.030.2330.048
98.23
97297
7314
5612443
< l81252
< l
1.52
131
—
3
49.50.97
15.19.80.178.4
12.312.040.2530.054
98.60
104279
72136
622445
< l5
< l4522
1.36
129
—
3
50.31.03
15.49.60.168.8
12.302.170.1280.062
99.95
100273
77144
652545
< l< l< l
66
< l< l
1.26
137
—
3
49.10.95
15.49.40.199.5
12.192.100.1400.065
99.03
175426
6613
3842353
< l2
< l761
< l
1.16
107
—
4
47.70.92
14.29.20.159.1
11.981.780.0740.057
95.16
173447
6714
1872354
< l916633
1.17
102—
4
49.91.28
15.69.60.178.9
11.832.390.3990.148
100.22
103326
64148
15626891051
6161023
1.258.9
861.4
5
Fe 2θ3 is total iron determined as Fe2θ3•
Table 3. Analyses and element ratios for basalts from Hole 504B.
CoreSectionInterval (cm)Piece
Siθ2 (%)TiO2
A12O3
Fe 2 O 3
a
MnOMgOCaONa 2OK 2 O
P2O5
Total
Ni (ppm)CrZnGaRbSrYZrNbPbTh
Fe/MgZr/NbTi/Zr
Lithologic Unit
321
109-111150
49.00.92
14.810.70.187.9
12.481.8810.2000.078
98.14
112293
77153
532239
< l1
< l
1.57—
141
21
331
26-28184
49.20.95
15.210.20.168.8
12.422.080.1090.081
99.20
154418
6815
3802151
11
< l
1.3551
112
22
341
20-22221
49.00.97
14.710.30.178.5
12.372.200.1390.082
98.43
158404
69152
862253
1< l< l
1.4053
110
22
361
84-87229
48.30.92
14.510.30.158.0
11.892.070.0180.082
96.23
97303
7215
< l602444
< l< l
2
1.50—
125
24
372
142-148387
47.00.88
12.711.10.177.5
12.501.780.1070.062
93.80
92367
7315
1612442
< l1
< l
1.72
126
25
382
64-66438
47.20.90
13.511.10.177.3
12.831.%0.0480.056
95.06
112357
7316
< l602542
< l< l
2
1.77—
128
25
392
118-121494
47.90.83
15.69.60.158.5
12.232.010.0250.072
96.92
123370
6315
< l652039
12
< l
1.3139
128
27
404
127-129574
47.50.79
15.89.20.148.3
12.161.890.0230.069
95.87
136356
6316
< l611936
133
1.2936
132
27
414
82-85589
47.00.78
15.79.20.148.4
11.961.890.0220.068
95.16
151398
6215
1611937
111
1.27—
126
27
413
102-104624
47.50.96
14.910.30.156.9
12.052.170.0150.076
95.02
101314
6215
1592243
112
1.7443
134
27
422
83-86669
47.80.95
14.79 . %0.158.0
11.902.130.0230.078
95.69
105335
6717
1602542
111
1.4442
136
28
432
7-9694
48.10.87
14.810.20.158.5
11.822.080.0170.072
96.61
110344
7115
1572440
112
1.39—
130
28
442
80-83728
48.70.90
14.610.50.167.7
12.332.090.0130.062
97.05
1173717016
1592140
121
1.58—
135
29
is total iron determined as
752
TRACE ELEMENT GEOCHEMISTRY OF BASALTS
Table 2. (Continued).
191
90-941125
49.51.32
15.09.30.219.0
12.542.190.1340.140
99.33
12232766143
17626981364
6181233
1.307.5
811.6
5
192
2-51132
49.71.00
15.78.80.159.2
12.042.140.0940.057
98.88
155481
70152
942355
< l121461
< l
1.10_
1090.4
6
192
70-71143
50.21.36
14.99.70.198.6
12.472.240.1580.174
99.99
122343
75152
170329011n.d.n.d.n.d.n.d.12
1.318.2
90—
7
211
64-661187
48.71.04
14.710.50.158.5
12.112.180.1100.089
98.08
9637465153
962455
1n.d.n.d.n.d.n.d.
< l< l
1.4355
113—
9
212
27-301203
50.041.01
14.910.40.218.4
12.272.260.1710.055
99.72
108298
81145
672853
< l8
< l56
< l1
1.43—
1140.3
9
212
33-351203
49.11.02
14.810.40.188.2
12.252.3350.1130.095
98.49
9237669153
942556
1n.d.n.d.n.d.n.d.
< l< l
1.4856
109—
9
212
133-1361216
50.40.75
16.98.90.168.5
13.271.980.0680.037
100.96
12035865152
571932
< l2
< l561
< l
1.22_
1400.6
9
215
56-591265
49.20.81
14.110.50.188.9
12.971.780.0690.056
98.56
12335863132
542334
< ln.d.n.d.n.d.n.d.4
< l
1.37_
143—
15
251
93-971412
48.21.00
13.510.6
0.208.0
12.401.990.1000.042
96.03
9628371123
672550
< l41
n.d.621
1.54
120—
17
271
75-791468
49.41.02
14.910.50.167.8
12.122.260.4410.058
98.66
7727271139
672549
< l5
< l371
< l
1.56
1250.3
17
272
38-431480
49.21.07
15.610.10.168.6
11.842.410.1910.088
99.26
7937264155
952559
< l10
1262
< l
1.36—
1090.2
18
281
23-261489
49.91.05
15.210.20.188.5
12.802.1270.0950.073
100.12
18337469152
1082659
1n.d.n.d.n.d.n.d.11
1.3959
107—
18
282
22-251505
49.80.98
14.610.50.178.5
12.722.080.1360.072
99.56
145253
75153
522540
1n.d.n.d.n.d.n.d.—1
1.4340
147—
19
282
23-271505
49.00.98
14.810.10.177.9
12.092.060.2470.055
97.40
118228
75146
512342
< l515512
1.47—
1400.5
19
283
16-181522
49.40.98
15.110.20.197.8
12.792.010.1150.069
98.65
143267
67152
572644
< ln.d.n.d.n.d.n.d.2
< l
1.52—
134—
19
284
127-1281533
48.60.98
14.710.30.187.3
12.532.020.2620.076
96.95
10628170145
602542
< ln.d.n.d.n.d.n.d.
< l< l
1.64—
140—
19
291
88-911542
48.91.02
14.510.40.208.7
12.372.170.1020.066
98.43
9227170143
682549
< l814731
1.39—
1250.4
20
291
123-1271577
49.90.97
15.710.50.198.1
12.862.000.1470.079
100.45
85263
69155
612645
< ln.d.n.d.n.d.n.d.
< l2
1.50—
130—
20
Table 3. (Continued).
461
74-76775
47.40.92
13.010.90.187.3
12.571.900.0070.057
94.23
923057516
1632345
111
1.74—
123
30
472
123-126835
48.51.02
13.811.10.156.9
11.172.020.0140.088
94.76
821338417
1532645
111
1.8645
136
30
492
83-85946
49.40.89
14.610.80.188.3
12.741.950.0280.062
98.95
89254
71161
482338
132
1.50—
140
33
511
113-115985
49.10.83
15.79.70.159.0
12.861.950.0160.067
99.37
1163526417
1612138
112
1.25
131
34
522
89-921017
48.20.86
15.49.70.158.1
12.702.010.0190.077
97.22
113347
6416
1622239
112
1.39—
132
34
551
98-1011092
49.81.41
14.411.00.208.9
11.872.410.0340.138
100.16
912547518
< l1073298
341
1.433386
36
571
88-911140
49.91.02
14.410.40.169.3
12.462.100.0130.070
99.82
129328
6916
< l772351
123
1.3051
120
37
581
115-1161203
49.60.93
15.710.50.177.8
12.72.010.0130.076
99.50
93262
6816
< l562341
< l< l
1
1.57
136
38
601
37-401259
48.51.06
14.310.70.167.9
11.802.110.0220.094
96.67
97278
7315
1702453
1< l
3
1.5753
120
39
621
94-971325
48.41.03
13.711.80.177.0
13.031.680.0150.063
96.89
132313
8216
< l502952
< l11
1.96
119
40
632
118-1211371
47.80.98
13.611.20.178.3
12.081.750.0160.063
95.96
127303
7715
1512749
< l13
1.56
120
40
642
28-321424
48.41.18
13.811.80.186.4
11.352.040.0210.102
95.27
691438418
< l572956
1< l
2
2.1456
126
43
662
25-281509
48.01.0
13.610.90.198.0
12.381.80.0040.071
96.04
1073077816
1672347
< l42
1.59—
127
45
702
19-211564
48.30.90
14.810.20.19.2
12.491.800.0110.070
97.94
14244368152
602041
< l33
1.29—
132
49
753
N. G. MARSH, J. TARNEY, G. L. HENDRY
150 -
100 -
%
*o
θ
C.4.
u
V
o
• A
θ
V
β
β
A " T
•
β
•
1.4
S? 1.2
s2 1.0
0.8
ΔΔ
# • • •
•
X
•
Fe/Mg Fe/Mg
Figure 4. Fe/Mg variation diagram of selected major and trace elements for samples from Leg 69. Symbols for litho-logic units as in Figure 2.
with changing Fe/Mg except for Lithologic Unit 2 (opencircles, Fig. 4), where CaO content seems to rise with in-creasing Fe/Mg. The decrease in Ni content with in-creasing Fe/Mg, with considerable overlap between thevarious units, suggests a general compositional consan-guinity. Differences in chemistry between the lithologicunits are revealed, however, by variations in trace andminor element abundances. For similar Fe/Mg ratios,TiO2 contents cover a significant range. Samples fromLithologic Units 5, 7, and 36 (triangles and pluses, Fig.4; designated datum, Fig. 5) have the highest TiO2 con-tents, and those from Lithologic Unit 27 (Fig. 5) gener-ally have the lowest, although one sample from Litho-logic Unit 9 (solid triangle, Fig. 4) also has a very lowTiO2 content. That sample also has high A12O3 and CaOcontents suggestive of Plagioclase accumulation.
Trace and minor element variations are illustrated byusing Zr as an index of fractionation. Zr is particularlyuseful in this respect. It has a low bulk distribution coef-ficient in ocean-ridge basalt systems and appears to beimmobile during the hydrous alteration of basalts. Inaddition, it is easily determined with high precision byXRF techniques (Tarney et al., 1979) (Figs. 6-11). TiO2,Y, and particularly Sr exhibit strong colinearity with
Zr, even when the small number of samples and limitedconcentration ranges within most units are considered.Samples from Lithologic Units 5 and 7 (open trianglesand pluses) have higher concentrations than most unitsof Zr, Ti, and Sr, as well as a higher Sr/Zr ratio, butlower Ti/Zr and Y/Zr ratios and only a small increase inY concentration. Although the sample from LithologicUnit 43 also has higher TiO2, Zr, and Y contents andlower Ti/Zr and Y/Zr ratios than samples from most ofthe other units, no increase in Sr concentration or Sr/Zrratio is apparent. Less marked differences in these traceelement ratios are also seen between the other units.These differences are more clearly apparent in the Leg69 data, where on average a larger number of samplesper unit were analyzed. Unit 6 (asterisks) and thechrome-spinel-bearing units (Units 2, 4, and 18; opencircles, open squares, and inverted open triangles) haveslightly lower Ti/Zr and Y/Zr ratios than the aphyricto sparsely phyric units, whereas their Sr/Zr ratios aremarginally higher on average.
The concentrations of La, Ce, and Nd are, apartfrom those of Unit 5 (and, we suspect, Lithologic Unit7), very low. A plot of Ce versus Y (Fig. 12) shows con-siderable scatter, most of which is due to poor analytical
754
TRACE ELEMENT GEOCHEMISTRY OF BASALTS
|oo
13
12
11
10
27
-
I
•.22
^ * .* .3*6 *
i
•
.•40•
•••*27
40
•
_
•43 _
12 -
11 -
10 -
9 -
8 -
150
100
1••• *36 *
i
•27
•43
1
-
-
i
1
36 <S^•
•22 *
/̂^27
i
2*7
i
40
i
4*3
-
1.4
5 1.2
I3*1.0
0.8
1 1
3*6
-
• 22 *--• " —
• .
1
-4.3
^3)40
-
1.0 1.5Fe/Mg
2.0
1 -
Figure 5. Fe/Mg variation diagram of selected major and trace elements for samples from Leg 70. For clarity, only lithologic unitsmentioned in the text are labelled.
precision for these low abundances. Nevertheless, theconsistently low Ce concentrations and low chondrite-normalized Ce/Y ratios suggest strong depletion in lightrare earth elements (LREE). Lithologic Unit 5 samples(open triangles), however, appear to have LREE-en-riched chondrite-normalized rare earth element (REE)patterns. Similarly, Nb concentrations in samples fromall the units except Lithologic Units 5 and 7 are verylow, generally less than 1 ppm, and the Zr/Nb ratios arecorrespondingly high ( 50). However, basalts fromLithologic Units 5 and 7 are distinctive in having 10 to13 ppm Nb and Zr/Nb ratios of approximately 8.
Chemical differences between the lithologic units arenot restricted solely to incompatible trace elements.Whereas Cr contents decrease with increasing Fe/Mgratio (Fig. 13), the higher Cr contents of the chrome-spinel-bearing units of Lithologic Unit 6 are clearly ap-parent in Figure 13. Additional analyses of samplesfrom Sites 504 and 505 are listed in Table 4.
ALTERATION EFFECTS
Before evaluating the igneous variations between thesamples from Hole 504B, the possible effects of altera-
tion on the basalt chemistry need to be assessed. TheLIL elements Rb, K, and to a lesser extent Ba and Sr arethe most susceptible to alteration. A plot of K2O versusan alteration-resistant element such as Zr (Pearce andCann, 1971) (Fig. 14) exhibits considerable scatter, sug-gesting that K2O contents in many samples have beenmodified by secondary processes. Rb abundances havebeen similarly affected, but Rb and K2O contents stillappear to co-vary with a K/Rb ratio of approximately400 (Fig. 15), a feature possibly more characteristic ofhydrothermal than of low-temperature alteration (hal-myrolysis). Some of the scatter at very low values ofK2O (<0.05%) and Rb (< 1 ppm) (as well as in the Badata for < 10 ppm, not shown) reflects low analyticalprecision. Sr appears to have been unaffected by altera-tion and still represents original igneous abundances, asindicated by the good positive correlation with Zr withinmost units (Figs. 10 and 11).
Generally, HFS ions are alteration resistant (Pearceand Cann, 1971; Pearce and Norry, 1979), and the goodcoherence between Zr, Ti, and Y concentrations (Figs.6-9) supports this observation. Similarly, Cr and Ni (notshown) appear to have been unaffected, the Ni contents
755
N. G. MARSH, J. TARNEY, G. L. HENDRY
1.50
100
Figure 6. Ti versus Zr for samples from Leg 69. Symbols for lithologicunits as in Figure 2.
1.50
100Zr (ppm)
0.50
Figure 7. Ti versus Zr for samples from Leg 70.
Figure 8. Y versus Zr for samples from Leg 69. Symbols for units as inFigure 2.
of Unit 1 being an exception. P2O5 (Fig. 16) is alsocommonly assumed to be alteration resistant, althoughMarsh et al. (1980) found that P2O5 content was modi-fied by alteration processes in the crystalline interiors ofpillow lavas. This may well explain the scatter of theP/Zr ratios within individual units, as in LithologicUnits 1 and 2 (solid and open circles), for example. TheP2O5 and Sr data probably reflect mineralogical con-trol; the Sr is held in Plagioclase, whereas in less crystal-line samples P2O5 is held in the mesostasis and is morereadily mobilized.
Changes in major element chemistry are less easilyassessed. Most of the samples analyzed are from theleast altered sections of flow or pillow interiors. Petro-graphically, the major alteration features are the partialto total replacement of olivine and the mesostasis, plussome filling of vesicles by clays. These features indicatethat Fe, Mg, and Si have been mobile (Floyd and Tar-ney, 1979; Pritchard et al., 1979; Site 504 chapter, thisvolume). The good correlations between Fe/Mg andtotal iron or MgO contents; the general rise in Fe/Mgratio with increasing TiO2; and the falling Ni and Crcontents (Figs. 4, 5, and 13) suggest that although Feand Mg may have been mobile, the major element datacan still provide qualitative estimates of the original ig-neous compositions.
756
TRACE ELEMENT GEOCHEMISTRY OF BASALTS
200
75
Zr (ppm)
Figure 9. Y versus Zr for samples from Leg 70.
200
150 -
100 -
Zr (ppm)
100
100
Figure 10. Sr versus Zr for samples from Leg 69. Symbols for units asin Figure 2.
150 -
100 -
50 75 100
Zr (ppm)
Figure 11. Sr versus Zr for samples from Leg 70.
15
Figure 12. Ce versus Y for samples from Leg 69. CeN/YN indicateschondrite-normalized Ce/Y ratios (approximately equivalent toCeN/HθN) Symbols for units as in Figure 2.
DISCUSSION
Comprehensive studies of basalts from the North At-lantic Ocean have illustrated the chemical variability ofthe magmas being erupted at the mid-ocean ridges (e.g.,Schilling, 1973; Hart et al., 1973; White et al., 1976;Rhodes et al., 1979; Bougault et al., 1979; Bougault andTreuil, 1980; Tarney et al., 1979; and Wood, Tarney, etal., 1979). On the basis of these studies three main ba-
757
N. G. MARSH, J. TARNEY, G. L. HENDRY
3 U U
400
300
200
-
-
A
i
*
D
D
A
•> Δ Δ
•
I
1
VVA AX
••
A •
β
β
500
400 -
0.5 1.0 1.5
Fe/Mg
3 300 -
ò
200 -
2.0
10075 100
Zr (ppm)
Figure 13. A. Cr versus Fe/Mg for samples from Leg 69. B. Cr versus Zr for samples from Leg 69 showing the higher Crcontents of the chrome-spinel-bearing units and Lithologic Unit 6 plus the limited concentration ranges within indi-vidual units. Symbols for units as in Figure 2.
salt types characterized by different HYG element andisotope ratios have been defined (Sun et al., 1979; andTarney et al., 1980). They are:
1) Normal or N type MORB (deleted tholeiitic ocean-ridge basalt), with low initial 87Sr/86Sr ratios (0.7023-0.7027, Hart et al., 1973); high 143Nd/144Nd ratios(0.5131-0.5133, DePaolo and Wasserburg, 1976, andO'Nions et al., 1977); and low abundances of HYG ele-ments, particularly the more-HYG elements, giving lowratios of more- to less-HYG elements compared to chon-dritic values.
2) Enriched or E type MORB (tholeiitic and alkalicocean-ridge basalts), with moderate to high initial 87Sr/86Sr ratios (0.7027-0.7035, White et al., 1976); 143Nd/144Nd ratios of less than 0.5131 (O'Nions et al., 1977);high abundances of more-HYG elements; and high ra-tios of more- to less-HYG elements.
3) Transitional or T type MORB (tholeiitic oceanridge basalt), with isotopic and HYG element character-istics intermediate between the N and E type MORB.
Representative analyses and selected ratios for eachtype of MORB are listed in Table 5 and illustrated inFigure 17.
A comparison of the data from Hole 504B with thedifferent types of MORB illustrates some of the maingeochemical features of these basalts. Although in termsof major element composition the basalts from Hole504B are similar to N type MORB from the Mid-Atlan-tic Ridge (MAR 22°N), their HYG element abundancesare in general lower (e.g., Zr ~ 50 ppm, Nb < 1 ppm,TiO2 < l°7o; see Table 2). High values for Zr/Nb ( 50),Ti/Zr ( 120, Figs. 6 and 7), Y/Zr ( 0.5, Figs. 8 and9), and low values of Ce/Y ratios (< 0.5 when chondrite
normalized, Fig. 12) suggest that not only are the abun-dances of the HYG elements low but that the more-HYG elements are more severely depleted relative to theless-HYG elements than in N type MORB from MAR22°N. This feature is clearly illustrated in Figure 17,where selected elements in representative samples havebeen normalized to the composition of the N typeMORB from MAR 22°N. The elements are plotted inorder of increasing estimated D values for typical man-tle mineral assemblages (Mattey et al., 1980).
Samples from most of the units exhibit the progres-sive depletion of the alteration-resistant more-HYG ele-ments (Nb, La, Ce, and Sr) in comparison to the less-HYG elements (Ti and Y). The depletion of the more-HYG elements is less marked in Lithologic Unit 4,whereas Lithologic Unit 5 appears to be identical to theT type MORB example, exhibiting progressively greaterenrichment of Ce, La, Nb, and to a lesser extent Sr.Even if the reduced analytical precision for Nb, La, andCe due to their low abundances in the more depletedunits is taken into account, the element distribution pat-terns suggest that Lithologic Units 5 and 7 have signifi-cantly different more-HYG element ratios than the otherunits—Nb/La ratios of 2 and 1, respectively, forexample, and La/Ce ratios of 0.4 and 0.2. None-theless, a general feature of all the samples from Hole504B is their low Ti and Y contents compared to N typeMORB, which has a similar Fe/Mg ratio and similar Ni,Cr, total iron, and MgO contents.
Detailed quantitative modeling of the possible chemi-cal relationships between the units will have to await theacquisition of high precision data for other alteration-resistant more-HYG elements. However, the present
758
TRACE ELEMENT GEOCHEMISTRY OF BASALTS
Table 4. Analyses and element ratios for basalts from Leg 69.
HoleCoreSectionInterval (cm)Piece
Siθ2 (%)Tiθ2A1 2O 3
F e 2 O 3
a
MnOMgOCaONa2θK 2 O
P2O5
Total
Ni (ppm)CrZnGaRbSrYZrNbBaLaCeNdPbTh
Fe/MgZr/NbTi/ZrC e N / Y N
504A61
90-9319
48.11.10
14.410.60.207.6
12.172.09—0.066
96.32
761817915
1692854
18
< l34
< l< l
1.6254
1220.2
504A71
63-6566
49.31.15
14.411.60.198.1
11.892.230.2450.074
99.18
14418977154
702853
< l1126641
1.66—
1300.5
505243
53-5591
49.50.67
17.07.50.148.7
12.591.920.1690.031
98.22
199473
50134
821637
< l18
12312
0.99—
1090.3
505B32
109-112118
48.80.93
15.69.40.159.2
12.352.130.0580.037
98.66
164381
66142
762452
< l6
< l47
< l1
1.19—
1070.4
5O5B61
46-49210
49.40.93
15.69.10.159.4
12.382.220.0490.037
99.27
169380
64132
762553
110
1662
< l
1.1253
1050.5
is total iron determined as
100Zr (ppm)
Figure 14. K2O versus Zr for samples from Legs 69 and 70. Symbolsfor lithologic units for the Leg 69 samples as in Figure 2. Leg 70samples are represented by small dots.
•
0.2 0.3 0.4 0.5
K2O (wt.
Figure 15. Rb versus K2O for samples from Legs 69 and 70. Sampleidentification as in Figure 14.
0.15
_ 0.10
0.05 -
75 100
Zr (ppm)
Figure 16. P2O5 versus Zr for samples from Leg 69. Symbols for unitsas in Figure 2.
body of data still places a number of constraints on themechanisms involved in producing these basalts.
Closed system fractional crystallization acting on asingle magma or several magmas of similar compositioncannot produce the observed differences in HYG ele-ment ratios between Units 5 and 7 and the other units.Units 5 and 7 have the highest more-HYG element abun-dances but still have lower Fe/Mg ratios and higher Crand Ni contents than many other units. Closed systemfractionation is also unlikely to produce the differencesin Ti/Zr ratio that appear in the different depleted
759
N. G. MARSH, J. TARNEY, G. L. HENDRY
Table 5. Representative analyses of dif-ferent types of MORB from the Mid-Atlantic Ridge.
Siθ2 (%)Tiθ2A12O3
F e 2 O 3
d
MnOMgOCaONa2θK 2 O
P2O5
Ni (ppm)CrRbSrYZrNbBaLaCe
Fe/MgZr/NbTi/ZrCe N /Y N
N a
50.01.60
16.2010.20
—7.70
11.302.800.221.25
138290
11363588
2.512
310
1.5335
1090.70
Basalt Type
rpb
49.31.74
14.0112.130.198.09
11.652.300.180.22
106317
410335
1101370
619
1.748
951.33
E c
45.12.78
12.7112.970.16
12.578.763.481.800.81
342310
36830
24281
62460
45.593
1.204.5
599.5
a N type MORB from MAR 22°N (val-ues compiled by Tarney et al., 1981).
b T type MORB from Reykjanes Ridge(values from Unit 409-3, Wood, Tar-ney, et al., 1979).
c E type MORB from Iceland, SampleISL 79 (values from Wood, Joron, etal., 1979).
is total iron as
units. Specifically, Lithologic Units 2 and 4 have higherCr and Ni contents and lower Ti/Zr and Fe/Mg ratiosthan Lithologic Units 1 and 3. Crystal fractionation ofthe observed phenocryst phases (chrome spinel, Plagio-clase, olivine, and, more rarely, clinopyroxene) wouldbe expected to leave the Ti/Zr (and Y/Zr) ratios un-changed or slightly reduced (Tarney et al., 1979).
Open system fractional crystallization as proposed byO'Hara (1977) provides a mechanism for preserving ma-jor element abundances while increasing both HYG ele-ment abundances and more-HYG/less-HYG element ra-tios, particularly those involving the least-HYG ele-ments (i.e., Ti, Y, and the HREE). Saunders (in press)suggests that this process is the most likely explanationfor the observed increases in the La/Yb, Zr/Y, andZr/Ti ratios with increasing Zr content in the basalts re-covered from the East Pacific Rise at the mouth of theGulf of California. The changes observed there in Zr/Tiand Zr/Y ratios are comparable to the differences ob-served between Unit 5 and the depleted units, but theywere accompanied by larger increases in the abundanceof Zr ( 50 to 150 ppm), Ti ( 1.0 to 2.5%) and Y («25to 50 ppm) than exist in the Hole 504B samples (Figs. 6and 8). In addition, REE patterns remained LREE de-
pleted and the Zr/Nb ratios were unchanged. Similarly,Mattey and Muir (1980) ascribe chemical variations inGalapagos Rift basalts and ferrobasalts to open systemfractionation, where again the REE patterns remainLREE depleted and Ta/Hf (analogous to Zr/Nb) ratiosremain constant. Theoretical modeling of high levelopen system crystal fractionation also produces similarresults (Tarney et al., 1980; and Pankhurst, 1977), so itappears unlikely that the entire range of compositionsshown by the Hole 504B basalts could be produced bycrystal fractionation from a magma, or several batchesof magmas of the same composition, at crustal levels.
Nevertheless, the Fe/Mg ratios, the Ni, Cr, and MgOcontents, and the petrographic evidence all suggest thatall the Hole 504B basalts have undergone some crystalfractionation and did not solidify as unmodified mantlemelts. Variations in HYG element abundances and ra-tios could well reflect low-pressure open system crystalfractionation involving more than one parental magma.
One possible mixing scheme would involve a depletedparent magma with HYG element ratios similar to thoseof Lithologic Unit 1 diluting a magma with HYG ele-ment ratios similar to those of Lithologic Unit 5, the re-sultant mix then undergoing varying amounts of crystalfractionation to produce the basalts of Lithologic Units2, 3, and 4. Batiza and Johnson (1980) explained theproduction of T type MORB at the East Pacific Riseclose to the Siqueiros Fracture Zone by binary magmamixing during low pressure open system crystal frac-tionation. Additional trace element and petrographicdata (Natland and Melson, 1980; Thompson and Hum-phris, 1980; and Humphris et al., 1980), however, sug-gest that their precise mixing scheme is not tenable andthat independent magmas with a T type chemistry weremore likely to have been involved.
Even allowing for the possibility of magma mixingduring low pressure open system crystal fractionation,magmas of at least two distinct compositions are neces-sary to produce the HYG element variations within theHole 504B basalts, one being similar perhaps to T typeMORB and the other to N type MORB. Studies onbasalts from the Mid-Atlantic Ridge in the FAMOUSarea and on those from Iceland have shown that while itis possible to generate basalts with REE patterns rang-ing from moderately LREE enriched (CeN/YbN * 2) toLREE depleted (CeN/YbN 0.5) from an initially ho-mogeneous mantle source by continuous batch partialmelting (Langmuir et al., 1977) or dynamic partial melt-ing (Wood, 1979b), ratios between the more-HYG ele-ments remained essentially unaffected. Moreover, thedifferences in isotopic and more-HYG element ratiosbetween N, T, and E type MORB appear to be adequate-ly explained only by some degree of mantle heterogene-ity (e.g., Tarney et al., 1980; and O'Nions et al., 1980).Evidence from the Atlantic suggests the veining of ahost mantle depleted in more-HYG elements with re-spect to the less-HYG elements by material with highmore-HYG/less-HYG element ratios and high abun-dances of more-HYG elements. The host mantle is thesource of N type MORB, and by increasing the quantityof vein material T or E type magmas can be produced
760
100
50 -
100
5.0 -
1.0 -
0.5 -
0.2 -
0.1
'. NMORB 'O TMORB -
. .-̂ EMORB _' - ' " " N ^ D Basanite
1 I 1 I 1 I I I i i I 1 i 1 T
; Hole 504B ILithologic Unit
• 1O 2D 3Δ 5
\ 1I /\ /\ 1
i i In i i i i i i i i i i i i
TRACE ELEMENT GEOCHEMISTRY OF BASALTS
Hole 504BLithologic Unit
D 4* 6V 17
Rb Ba K Nb La Ce Sr P Zr Ti Y Fe Mg Cr Ni Rb Ba K Nb La Ce Sr P Zr Ti Y Fe Mg Cr Ni Rb Ba K Nb La Ce Sr P Zr Ti Y Fe Mg Cr Ni
Figure 17. Comparative element abundances in selected samples from Leg 69 normalized to N type MORB. Normalizing values are based on N typeMORB from the Mid-Atlantic Ridge, 22°N, as compiled by Tarney et al. (1981). The example of E type MORB is from Iceland (ISL 79; Wood,Joron, et al., 1979), and that of T type MORB is from the Reykjanes Ridge (409-3; Wood, Tarney, et al., 1979). The basanite sample from south-eastern Australia (2679; Frey et al., 1978) illustrates the possible composition of the vein material involved in generating T and E type MORB.Dashes connect estimated element abundances.
during partial melting. The favoured composition of thevein material would be similar to the compositions ob-served in many strongly undersaturated alkalic basalts(see Fig. 17).
The difference in Zr/Nb ratios and inferred differ-ences in Nb/La and La/Ce ratios within the Hole 504Bbasalts, by analogy with basalts from the MAR, suggestthat the mantle source feeding the Costa Rica Rift washeterogeneous with respect to more-HYG element con-tents. The depleted basalts at this site would have beenproduced by partial melting of the host mantle, possiblywith a small and variable contribution from the veinmaterial; the enriched magma that eventually gave riseto Unit 5 would have been produced by partial meltingof a mantle with a higher vein content. However, thelow contents of Ti, Y, and Zr plus the progressively in-creasing depletion of the more-HYG elements observedin most units in Hole 504B relative to 22 °N MAR Ntype MORB suggest that the host mantle underlying thispart of the Panama Basin is more severely depleted inHYG elements than the host mantle underlying the At-lantic.
The question remains as to how two apparently dis-tinct magma compositions could exist at a ridge spread-ing at a moderate rate. Normal models of the spreadingprocess envisage a continuous magma chamber. On slow-ly spreading ridges such as a Mid-Atlantic Ridge, basaltsrepresenting two or more separate magma types are notuncommon in cored sections, but this presents no diffi-culty if the magma chambers are only semicontinuousor discontinuous; small separate magma chambers cancoexist. On a faster-spreading ridge there would be no
serious problem either if the enriched lavas were at thetop of the cored section and could be regarded as theproduct of off-axis volcanism. However, in Hole 504Bthe enriched units occur at too deep a level to be con-sidered off-axis lavas. In theory a continuous magmachamber should smooth out the trace element variationsthat result from different degrees of batch partial melt-ing or small-scale mantle source heterogeneities. Onepossible explanation is that magma chambers along theCosta Rica Rift were long lived but not necessarily con-tinuous, in which case separate magma types could beextruded, albeit to a very limited extent. Some of thetrace element variations within the remainder of the lavasequence may have arisen through the mixing of the twomagma types within semi-continuous magma chambers,although clearly the depleted magma type is dominantat Site 504.
In conclusion, the basalts sampled so far from thefloor of the Panama Basin appear to be predominantlyhighly HYG-element-depleted mid-ocean-ridge basalts.Occasionally interdigitated with these depleted basaltsare units with T type MORB characteristics. All the ba-salts have undergone some high-level crystal fractiona-tion, although this has had far less effect on the com-position of the erupted basalts than observed in the ad-jacent Galapagos Spreading Center or along the EastPacific Rise, where the magmas are commonly fraction-ated to ferrobasalts. Variations in HYG element ratiosbetween different units could reflect some mixing be-tween the two magma types at crustal levels, and a het-erogeneous mantle source is necessary to explain thetotal range of HYG element ratios observed.
761
N. G. MARSH, J. TARNEY, G. L. HENDRY
REFERENCES
Anderson, R. N., Clague, D. A., Klitgord, K. D., Marshall, M., andNishimori, R. K., 1975. Magnetic and petrologic variations alongthe Galapagos Spreading Center and their relation to the Galapa-gos melting anomaly. Geol. Soc. Am. Bull., 86:683-694.
Batiza, R., and Johnson, J. R., 1980. Trace element and isotopic evi-dence for magma mixing in alkalic and transitional basalts near theEast Pacific Rise at 8°N. In Rosendahl, B. R., Hekinian, R., etal., Init. Repts. DSDP, 54: Washington (U.S. Govt. Printing Of-fice), 63-70.
Bougault, H., Joron, J. L., and Treuil, M., 1980. The primordialchondritic nature and large-scale heterogeneities in the mantle:evidence from high and low partition coefficient elements in oce-anic basalts. Philos. Trans. R. Soc. London, Ser. A, 297:203-213.
Bougault, H., and Treuil, M., 1980. Mid-Atlantic Ridge: zero-agegeochemical variations between Azores and 22°N. Nature, 286:209-212.
Bougault, H., Treuil, M., and Joron, J. L., 1979. Trace elements inbasalts from 23 °N and 36 °N in the Atlantic Ocean: fractionalcrystallization, partial melting, and heterogeneity of the UpperMantle. In Melson, W. G., Rabinowitz, P. D., et al., Init. Repts.DSDP, 45: Washington (U.S. Govt. Printing Office), 493-506.
Clague, D. A., and Bunch, T. E., 1976. Formation of ferrobasalt atEast Pacific midocean spreading centers. /. Geophys. Res., 81:4247-4256.
DePaolo, D. J., and Wasserburg, G. J., 1976. Nd isotopic variationsand petrogenetic models. Geophys. Res. Lett., T> .lA‰TSl.
Floyd, P. A., and Tarney, J., 1979. First-order alteration chemistry ofLeg 49 basement rocks. In Luyendyk, B. P., Cann, J. R., et al., In-it. Repts. DSDP, 49: Washington (U.S. Govt. Printing Office),693-708.
Frey, F. A., Green, D. H., and Ray, S. D., 1978. Integrated models ofbasalt petrogenesis. /. Petrol., 19:463-513.
Hart, S. R., 1971. K, Rb, Cs, Sr and Ba contents and Sr isotope ratiosof ocean floor basalts. Philos. Trans. R. Soc. London, Ser. A,268:573-587.
Hart S. R., Erlank, A. J., and Kable, E. J. D., 1974. Sea floor basaltalteration: some chemical and Sr isotopic effects. Contrib. Miner-al. Petrol., 44:219-230.
Hart, S. R., Schilling, J.-G., and Powell, J. L., 1973. Basalts fromIceland and along the Reykjanes Ridge: Sr isotope geochemistry.Nature Phys. Sci., 246:104-107.
Humphris, S. E., Thompson, R. N., Gibson, I. L., and Marriner, G.F., 1980. Comparison of geochemistry of basalts from the EastPacific Rise, OCP Ridge and Siqueiros Fracture Zone, Deep SeaDrilling Project Leg 54. In Rosendahl, B. R., Hekinian, R., et al.,Init. Repts. DSDP, 54: Washington (U.S. Govt. Printing Office),635-650.
Langmuir, C. H., Bender J. F., Bence, A. E., Hanson, G. N., andTaylor, S. R., 1977. Petrogenesis of basalts from the FAMOUSarea: Mid-Atlantic Ridge. Earth Planet. Sci. Lett., 36:133-156.
Leake, B. E., Hendry, G. L., Kemp, A., Plant, A. G., Harvey, P. K.,Wilson, J. R., Coats, J. S., Aucott, J. W, Lünel, T., and How-arth, R. J., 1969. The chemical analysis of rock powders by auto-matic X-ray fluorescence. Chem. Geol., 5:7-86.
Lonsdale, P., and Klitgord, K. D., 1978. Structure and tectonic his-tory of the eastern Panama Basin. Geol. Soc. Am. Bull., 89:981-999.
Luddon, J. N., and Thompson, G., 1979. An evaluation of the behav-ior of the rare earth elements during the weathering of sea-floorbasalt. Earth Planet. Sci. Lett., 43:85-92.
Marsh, N. G., Saunders, A. D., Tarney, J., and Dick, H. J. B., 1980.Geochemistry of basalts from the Shikoku and Daito Basins, DeepSea Drilling Project Leg 58. In Klein, G. deV., Kobayashi, K., etal., Init. Repts. DSDP, 58: Washington (U.S. Govt. Printing Of-fice), 805-842.
Mattey, D. P., Marsh, N. G., and Tarney, J., 1980. The geochemistry,mineralogy, and petrology of basalts from the West Philippine andParece Vela Basins and from the Palau-Kyushu and West MarianaRidges, Deep Sea Drilling Project Leg 59. In Kroenke, L., Scott,R., et al., Init. Repts. DSDP, 59: Washington (U.S. Govt. PrintingOffice), 753-800.
Mattey, D. P., and Muir, I. D., 1980. Geochemistry and mineralogyof basalts from the Galapagos Spreading Center, Deep Sea Drill-ing Project Leg 54. In Rosendahl, B. R., Hekinian, R., et al., In-it. Repts. DSDP, 54: Washington (U.S. Govt. Printing Office),755-772.
Mitchell, W. S., and Aumento, R, 1977. Uranium in oceanic rocks:DSDP Leg 37. Can. J. Earth Sci., 14:794-808.
Natland, J. H. and Melson, W. G., 1980. Compositions of basalticglasses from the East Pacific Rise and Siqueiros Fracture Zone,Near 9CN. In Rosendahl, B. R., Hekinian, R., et al., Init. Repts.DSDP, 54: Washington (U.S. Govt. Printing Office), 705-724.
O'Hara, M. J., 1977. Geochemical evolution during fractional crystal-lization of a periodically refilled magma chamber. Nature, 266:503-507.
O'Nions, R. K., Evensen, N. M., and Hamilton, P. J., 1980. Differen-tiation and evolution of the mantle. Philos. Trans. R. Soc. Lon-don, Ser. A, 297:479-493.
O'Nions, R. K., Hamilton, P. J., and Evensen, N. M., 1977. Varia-tions in 143Nd/144Nd and 87Sr/86Sr ratios in oceanic basalts. EarthPlanet. Sci. Lett., 34:13-22.
Pankhurst, R. J., 1977. Open system crystal fractionation and incom-patible element variation in basalts. Nature, 268:36-38.
Pearce, J. A., and Cann, J. R., 1971. Ophiolite origin investigated bydiscriminent analysis using Ti, Zr and Y. Earth Planet. Sci. Lett.,12:339-349.
Pearce, J. A., and Norry, M. J., 1979. Petrogenetic implications ofTi, Zr, Y, and Nb variations in volcanic rocks. Contrib. Mineral.Petrol., 69:33-47.
Pritchard, R. G., Cann, J. R., and Wood, D. A., 1979. Low-tempera-ture alteration of oceanic basalts, DSDP Leg 49. In Luyendyk, B.P., Cann, J. R., et al., Init. Repts. DSDP, 49: Washington (U.S.Govt. Printing Office), 709-714.
Reynolds, R. C , Jr., 1967. Estimation of mass absorption coefficientsby Compton scattering: improvements and extensions of the meth-od. Am. Mineral., 52:1493-1502.
Rhodes, J. M., Blanchard, D. P., Dungan, M. A., Rodgers, K. V.,and Brannon, J. C , 1979. Chemistry of Leg 45 basalts. In Melson,W. G., Rabinowitz, P. D., et al., Init. Repts. DSDP, 45:Washington (U.S. Govt. Printing Office), 447-460.
Rosendahl, B. R., Hekinian, R., et al., 1980. Init. Repts. DSDP, 54:Washington (U.S. Govt. Printing Office).
Saunders, A. D., in press. The Gulf of California: geochemistry of ba-salts recovered during Leg 65 of the Deep Sea Drilling Project. InLewis, B. T. R., Robinson, P., et al., Init. Repts. DSDP, 65:Washington (U.S. Govt. Printing Office).
Saunders, A. D., Fornari, D. J., Joron, J. L., Tarney, J., and Treuil,M. in press. Geochemistry of basic igneous rocks recovered fromthe Gulf of California: Deep Sea Drilling Project Leg 64. In Cur-ray, J. R., Moore, D. G., et al., Init. Repts. DSDP, 64, Pt. 2:Washington (U.S. Govt. Printing Office).
Saunders, A. D., Tarney, J., Marsh, N. G., and Wood, D. A., 1980.Ophiolites as ocean crust or marginal basin crust: a geochemi-cal approach. Proc. Int. Ophiolite Conf., Nicosia, Cyprus, pp.193-204.
Schilling, J.-G., 1973. Iceland mantle plume: geochemical evidencealong Rekjanes Ridge. Nature, 242:565-571.
Schilling, J.-G., Anderson, R. N., and Vogt, P., 1976. Rare earth, Feand Ti variations along the Galapagos spreading centre, and theirrelationship to the Galapagos mantle plume. Nature, 261:108-113.
Sun, S.-S., Nesbitt, R. W., and Sharaskin, A. Ya., 1979. Geochemicalcharacteristics of mid-ocean ridge basalts. Earth Planet. Sci. Lett.,44:119-138.
Tarney, J., Saunders, A. D., Mattey, D. P., Wood, D. A., and Marsh,N. G., 1981. Geochemical aspects of back-arc spreading in theScotia Sea and western Pacific. Philos. Trans. R. Soc. London,Ser. A, 300:263-285.
Tarney, J., Saunders, A. D., Weaver, S. D., Donnellan, N. C. B., andHendry, G. L., 1979. Minor-element geochemistry of basalts fromLeg 49, North Atlantic Ocean. In Luyendyk, B. P, Cann, J. R., etal., Init. Repts. DSDP, 49: Washington (U.S. Govt. Printing Of-fice), 657-691.
Tarney, J., Wood, D. A., Saunders, A. D., Cann, J. R., and Varet, J.,1980. Nature of mantle heterogeneity in the North Atlantic: evi-
762
TRACE ELEMENT GEOCHEMISTRY OF BASALTS
dence from deep sea drilling. Philos. Trans. R. Soc. London, Ser.A, 297:179-202.
Thompson, R. N., and Humphris, S. E., 1980. Silicate mineralogy ofbasalts from the East Pacific Rise, OCP Ridge, and SiqueirosFracture Zone: Deep Sea Drilling Project, Leg 54. In Rosendahl,B. R., Hekinian, R., et al., Init. Repts. DSDP, 54: Washington(U.S. Govt. Printing Office), 651-670.
van Andel, Tj. H., Heath, G. R., et al., 1973. Init. Repts. DSDP, 16:Washington (U.S. Govt. Printing Office).
Vogt, P. R. 1979. Amplitudes of oceanic magnetic anomalies and thechemistry of oceanic crust: synthesis and review of 'magnetic tele-chemistry.' Can J. Earth. Sci., 16:2236-2262.
Vogt, P. R., and Johnson, G. L., 1973. Magnetic telechemistry of oce-anic crust? Nature, 245:373-375.
White, W. M., Schilling, J.-G., and Hart, S. R., 1976. Evidence of theAzores mantle plume from strontium isotope geochemistry of theCentral North Atlantic. Nature, 263:659-663.
Wood, D. A., 1979. Dynamic partial melting: its application to thepetrogeneses of basalts erupted in Iceland, the Faeroe Islands, theIsle of Skye (Scotland) and the Troodos Massif (Cyprus). Geo-chim. Cosmochim. Acta, 43:1031-1046.
Wood, D. A., Joron, J.-L., Treuil, M., Norry, M. J., and Tarney, J.,1979. Elemental and Sr isotope variations in basic lavas from Ice-land and the surrounding ocean floor: The nature of mantle sourcein homogeneities. Contrib. Mineral. Petrol., 70:319-339.
Wood, D. A. Tarney, J., Varet, J., Saunders, A. D., Bougault, H.,Joron, J. L., Treuil, M., and Cann, J. R., 1979. Geochemistry ofbasalts drilled in the North Atlantic by IPOD Leg 49: implicationsfor mantle heterogeneity. Earth Planet. Sci. Lett., 42:77-97.
763