Winterer, E.L., Sager, W.W., Firth, J.V., and Sinton, J.M. (Eds.), 1995 Proceedings of the Ocean Drilling Program, Scientific Results, Vol. 143 15. PETROLOGY AND GEOCHEMISTRY OF IGNEOUS ROCKS FROM ALLISON AND RESOLUTION GUYOTS, SITES 865 AND 866 1 P.E. Baker, 2 P.R. Castillo, 3 and E. Condliffe 2 ABSTRACT At Site 866 (Resolution Guyot), the volcanic basement (> 1620 mbsf and < 128 Ma) consists of a series of subaerial lava flows separated by scoriaceous, rubbly, and clay (iron-rich smectite) intervals. The lavas may be divided into three main groups on petrographic and geochemical grounds. Because of alteration, the geochemical evidence depends mainly on relatively immobile elements, such as Ti, Zr, Nb, and the rare earth elements. The lowest group (>1727 mbsf) is picritic and alkalic (high Nb/Ti and Nb/Zr). The middle group (1673-1727 mbsf), with megacrysts and phenocrysts of Plagioclase, olivine, and clinopyroxene, is more markedly alkalic (e.g., steep mid-ocean ridge basalt-normalized light rare earth element-enriched patterns). The upper group (above 1673 mbsf) includes basalts rich in Plagioclase megacrysts overlain by picrites, and is more tholeiitic (lower Nb/Ti, Zr/Ti, and flatter rare earth element patterns). At Site 865 (Allison Guyot), altered basaltic sills are intrusive into Albian clayey dolomitic limestones between 830 and 870 mbsf. Several lines of evidence indicate that the sediment was unconsolidated at the time of invasion by the basalt (<lll Ma). Salitic clinopyroxenes, abundances and ratios of the less mobile incompatible elements, and pronounced light rare earth element-enrichment all point to a decidedly more alkalic affinity than was seen at Site 866. Resolution and Allison guyots probably originated within the region of intense hotspot volcanism referred to as the South Pacific isotopic and thermal anomaly or SOPITA. The two guyots probably followed a similar tectonic pathway and may have passed over more than one hotspot. Geochemical evidence (e.g., Nb/Zr and Zr/Ti) suggests that they have more in common with the Society-Austral (Tubuai) Islands than with islands to the east (e.g., Marquesas, Easter) or the west (Cook, Samoa): this is also consistent with lineaments derived by backtracking. INTRODUCTION A vast area of the South Pacific Ocean, about 3000 km across, was the site of intense mid-plate volcanism during the Early Cretaceous. The location of this activity probably corresponds with the present-day South Pacific Superswell (McNutt and Fischer, 1987), which includes the Society, Cook, and Austral archipelagos. It also belongs to SOPITA, the area of the so-called South Pacific isotopic and thermal anomaly (Staudigel et al., 1991). The lavas of these islands also form part of the isotopically distinct Southern Hemisphere belt referred to as the Dupal Anomaly (Hart, 1984). In the western part of the Mid-Pacific Mountains (MPM), a series of broad plateaus is surmounted by flat- topped seamounts. The age of the oceanic crust beneath the MPM is estimated to be about 119 to 130 Ma near Allison Guyot and may be as old as 154 Ma farther west, around Resolution Guyot (Sager, Winterer, Firth, et al., 1993). The guyots are capped by shallow-water limestones of Barremian-Albian age (124-98 Ma). The Cretaceous seamounts of the MPM probably formed over the South Pacific Superswell and were translated northwestward on zig-zag pathways, determined by changes in plate motion, to their present location (Fig. 1). From the hotspot lineaments calculated by Duncan and Clague (1985), the seamounts of Sites 865 and 866 lie close to the Easter Island track, but would have passed near the Marquesas and Society hotspots during their transit. A period of rejuvenation and uplift may have accompanied their passage over these other hotspots. The ages calculated by Duncan and Clague (1985) are consistent with the location of the guyots over the Marque- sas or Society hotspots at 100 to 120 Ma. Site 866 is discussed first, as it represents a substantial section through volcanic basement and was the only instance where basement was reached during Leg 143. At Site 865, on Allison Guyot, igneous 1 Winterer, E.L., Sager, W.W., Firth, J.V., and Sinton, J.M. (Eds.), 1995. Proc. ODP, Sci. Results, 143: College Station, TX (Ocean Drilling Program). 2 Department of Earth Sciences, University of Leeds, Leeds LS2 9JT, United King- dom. Geological Research Division, Scripps Institution of Oceanography, University of California, San Diego, La Jolla, CA 92093-0220, U.S.A. rocks are confined to a group of basaltic sills intruded into the sedi- ments. Some comparisons are made with a few samples from Site 869, on a sediment apron extending southward from Wodejebato Guyot and Pikinni Atoll. The detailed petrology and geochemistry of the volcaniclastics from Site 869 are treated separately (see Janney et al., this volume). ANALYTICAL METHODS X-ray fluorescence (XRF) analyses for major and trace elements (excluding rare earth elements) were conducted at the Department of Earth Sciences, Leeds University, using a wavelength-dispersive auto- mated Philips PW 1400 spectrometer. Major elements were deter- mined on fused glass beads and trace elements on pressed powder pel- lets. Accuracy and precision for major elements are estimated at better than 3% for Si, Ti, Fe, Ca, and K and 7% for Mg, Na, Al, Mn, and P: for trace elements above 10 ppm they are estimated at better than 10%. Rare earth element (REE) determinations were performed on a VG Instruments PlasmaQuad 11+ inductively coupled plasma mass spectrometer (ICPMS) at the Scripps Institution of Oceanography. Multiplier voltage was set at 5 kV and nebulizing gas flow rate at 0.2 L/m. 115 In was used as an internal standard, and calibration was conducted using standard solutions of 5, 10, 50 and 100 ppb REE. Accuracy and precision of the analyses were monitored using the rock standards AGV-1 and BCR-1. Rock powders (0.014 g) were dis- solved in clean Teflon vessels using l mL 2:1 mixture of concen- trated HF and HNO 3 and then heated overnight over a hot plate at low temperature. The resulting solution was evaporated to dryness, resus- pended in a small amount of concentrated HNO 3 , and evaporated to dryness, and finally diluted to a factor of 1000 in a 1 % HNO 3 solution containing 100 ppb 115 In. Accuracy of the analyses, based on repeated measurements of BCR-1 and AGV-1 standards, is better than 5%, especially for the light elements. Microprobe analyses were performed using a CAMECA SX-50 instrument fitted with three wavelength dispersive spectrometers and a LINK 10/55S energy dispersive system. Analysis conditions were as follows: 245
Transcript
Winterer, E.L., Sager, W.W., Firth, J.V., and Sinton, J.M. (Eds.), 1995Proceedings of the Ocean Drilling Program, Scientific Results, Vol. 143
15. PETROLOGY AND GEOCHEMISTRY OF IGNEOUS ROCKS FROM ALLISONAND RESOLUTION GUYOTS, SITES 865 AND 8661
P.E. Baker,2 P.R. Castillo,3 and E. Condliffe2
ABSTRACT
At Site 866 (Resolution Guyot), the volcanic basement (> 1620 mbsf and < 128 Ma) consists of a series of subaerial lava flowsseparated by scoriaceous, rubbly, and clay (iron-rich smectite) intervals. The lavas may be divided into three main groups onpetrographic and geochemical grounds. Because of alteration, the geochemical evidence depends mainly on relatively immobileelements, such as Ti, Zr, Nb, and the rare earth elements. The lowest group (>1727 mbsf) is picritic and alkalic (high Nb/Ti andNb/Zr). The middle group (1673-1727 mbsf), with megacrysts and phenocrysts of Plagioclase, olivine, and clinopyroxene, is moremarkedly alkalic (e.g., steep mid-ocean ridge basalt-normalized light rare earth element-enriched patterns). The upper group(above 1673 mbsf) includes basalts rich in Plagioclase megacrysts overlain by picrites, and is more tholeiitic (lower Nb/Ti, Zr/Ti,and flatter rare earth element patterns). At Site 865 (Allison Guyot), altered basaltic sills are intrusive into Albian clayey dolomiticlimestones between 830 and 870 mbsf. Several lines of evidence indicate that the sediment was unconsolidated at the time ofinvasion by the basalt (<ll l Ma). Salitic clinopyroxenes, abundances and ratios of the less mobile incompatible elements, andpronounced light rare earth element-enrichment all point to a decidedly more alkalic affinity than was seen at Site 866. Resolutionand Allison guyots probably originated within the region of intense hotspot volcanism referred to as the South Pacific isotopic andthermal anomaly or SOPITA. The two guyots probably followed a similar tectonic pathway and may have passed over more thanone hotspot. Geochemical evidence (e.g., Nb/Zr and Zr/Ti) suggests that they have more in common with the Society-Austral(Tubuai) Islands than with islands to the east (e.g., Marquesas, Easter) or the west (Cook, Samoa): this is also consistent withlineaments derived by backtracking.
INTRODUCTION
A vast area of the South Pacific Ocean, about 3000 km across, wasthe site of intense mid-plate volcanism during the Early Cretaceous.The location of this activity probably corresponds with the present-daySouth Pacific Superswell (McNutt and Fischer, 1987), which includesthe Society, Cook, and Austral archipelagos. It also belongs toSOPITA, the area of the so-called South Pacific isotopic and thermalanomaly (Staudigel et al., 1991). The lavas of these islands also formpart of the isotopically distinct Southern Hemisphere belt referred to asthe Dupal Anomaly (Hart, 1984). In the western part of the Mid-PacificMountains (MPM), a series of broad plateaus is surmounted by flat-topped seamounts. The age of the oceanic crust beneath the MPM isestimated to be about 119 to 130 Ma near Allison Guyot and may be asold as 154 Ma farther west, around Resolution Guyot (Sager, Winterer,Firth, et al., 1993). The guyots are capped by shallow-water limestonesof Barremian-Albian age (124-98 Ma). The Cretaceous seamounts ofthe MPM probably formed over the South Pacific Superswell and weretranslated northwestward on zig-zag pathways, determined by changesin plate motion, to their present location (Fig. 1). From the hotspotlineaments calculated by Duncan and Clague (1985), the seamounts ofSites 865 and 866 lie close to the Easter Island track, but would havepassed near the Marquesas and Society hotspots during their transit. Aperiod of rejuvenation and uplift may have accompanied their passageover these other hotspots. The ages calculated by Duncan and Clague(1985) are consistent with the location of the guyots over the Marque-sas or Society hotspots at 100 to 120 Ma.
Site 866 is discussed first, as it represents a substantial sectionthrough volcanic basement and was the only instance where basementwas reached during Leg 143. At Site 865, on Allison Guyot, igneous
1 Winterer, E.L., Sager, W.W., Firth, J.V., and Sinton, J.M. (Eds.), 1995. Proc. ODP,Sci. Results, 143: College Station, TX (Ocean Drilling Program).
2 Department of Earth Sciences, University of Leeds, Leeds LS2 9JT, United King-dom.
Geological Research Division, Scripps Institution of Oceanography, University ofCalifornia, San Diego, La Jolla, CA 92093-0220, U.S.A.
rocks are confined to a group of basaltic sills intruded into the sedi-ments. Some comparisons are made with a few samples from Site869, on a sediment apron extending southward from WodejebatoGuyot and Pikinni Atoll. The detailed petrology and geochemistry ofthe volcaniclastics from Site 869 are treated separately (see Janney etal., this volume).
ANALYTICAL METHODS
X-ray fluorescence (XRF) analyses for major and trace elements(excluding rare earth elements) were conducted at the Department ofEarth Sciences, Leeds University, using a wavelength-dispersive auto-mated Philips PW 1400 spectrometer. Major elements were deter-mined on fused glass beads and trace elements on pressed powder pel-lets. Accuracy and precision for major elements are estimated at betterthan 3% for Si, Ti, Fe, Ca, and K and 7% for Mg, Na, Al, Mn, and P:for trace elements above 10 ppm they are estimated at better than10%. Rare earth element (REE) determinations were performed on aVG Instruments PlasmaQuad 11+ inductively coupled plasma massspectrometer (ICPMS) at the Scripps Institution of Oceanography.Multiplier voltage was set at 5 kV and nebulizing gas flow rate at 0.2L/m. 115In was used as an internal standard, and calibration wasconducted using standard solutions of 5, 10, 50 and 100 ppb REE.Accuracy and precision of the analyses were monitored using the rockstandards AGV-1 and BCR-1. Rock powders (0.014 g) were dis-solved in clean Teflon vessels using l mL 2:1 mixture of concen-trated HF and HNO3 and then heated overnight over a hot plate at lowtemperature. The resulting solution was evaporated to dryness, resus-pended in a small amount of concentrated HNO3, and evaporated todryness, and finally diluted to a factor of 1000 in a 1 % HNO3 solutioncontaining 100 ppb 115In. Accuracy of the analyses, based on repeatedmeasurements of BCR-1 and AGV-1 standards, is better than 5%,especially for the light elements.
Microprobe analyses were performed using a CAMECA SX-50instrument fitted with three wavelength dispersive spectrometers anda LINK 10/55S energy dispersive system. Analysis conditions wereas follows:
245
RE. BAKER, PR. CASTILLO, E. CONDLIFFE
Figure 1. Location of Allison (Site 865) and Resolution (Site 866) guyots inrelation to island groups and hotspot lineaments of the South Pacific (afterShipboard Scientific Party, 1981). Approximate location of SOPITA afterPalacz and Saunders (1986).
1. Silicates and opaque oxides, excluding feldspars: beam energy,15 kV; beam current, 15 nA; count times, Na, Mg, Al, Si, K, Ca, Ti,all 15 s on peak and 10 s on background; Cr, Mn, Fe, Ni, 30 s on peakand 20 s on background.
2. Feldspars: beam energy, 15 kV; beam current, 10 nA. Counttimes for all elements: 10 s on peak and 10 s on background. Wherenecessary, beam broadened to 2 or 3 µm to prevent excessive ele-ment loss.
3. Raw counts were corrected for inter-elemental effects usingCAMECA proprietary software.
Resolution Guyot (21°19.9'N, 174°18.8'E) lies in 1373 m of waterand consists of a 1620-m cap of shallow-water carbonates and pelagicsediments that rests on a topographically subdued volcanic structurerising only about 0.5 km above the general level of the MPM plateau.K/Ar whole-rocks, whose ages inevitably are suspect with such highlyaltered states, yield ages in the range of 107 to 125 Ma (Pringle et al.,this volume). On the other hand, 40Ar/39Ar dates by the same authorsgive ages in the range of 120 to 129 Ma (i.e., mid-Barremian to mid-Aptian on Harland et al.'s [1990] time scale). Errors in the measure-ments are such that it is not possible to be certain whether this repre-sents a simple succession of lava flows or whether there are also someintercalated sills. However, the new petrographic and geochemicaldata presented here show substantial petrologic variations in the base-ment of Resolution Guyot. It is now thought likely that the 125-m-thickvolcanic sequence drilled at Site 866, originally interpreted as exclu-sively lava flows separated by rubbly intervals (Fig. 2), may containsome intrusive bodies. The average thickness of the flows and sills isestimated at about 10 m, but poor recovery of the contacts and thefriable nature of the interbasaltic beds probably means that they havebeen underestimated. The interbasaltic intervals range from clays toclast- or matrix-supported breccias in which the clasts are subangularto subrounded fragments of vesicular and amygdaloidal basalt, similarto the associated lava flows and sills. The matrix generally consists of
reddish sandy clay; distinct layers of red clay also are found within thebreccias (e.g., at interval 143-866A-180R-5, 54-57 cm). Fractures andcavities within the breccias also may be filled with red clay. X-raydiffraction (XRD) analysis of the clay indicates that it is an iron-richsmectite; some kaolinite also has been identified.
Most of the interbasaltic intervals are considered to be the prod-ucts of various processes operating on a tropical or subtropical vol-canic land surface. Each such interval probably represents decades orcenturies as opposed to the span of perhaps hours or days required foremplacement of a single lava flow. Some of the interbasaltic intervalsmay represent oxidized soils, boles, or lateritic horizons. Other inter-vals probably represent the oxidized rubbly surface of aa flows,where weathering products have washed down to fill the interveningspaces. Some breccias may have formed as screes or been redistrib-uted as mudflows or debris slides, but others may have been causedby the intrusion of sills. The fractured and porous nature of the Unit7 breccia provided access for hydrothermal fluids and their precipi-tates. Veins and cavities are commonly filled with calcite, zeolites,clay minerals, and analcime.
Petrography
In the Leg 143 Initial Reports volume (Sager, Winterer, Firth, etal., 1993), the volcanic succession was divided into 12 units on thebasis of the stratigraphic incidence of what originally were thought tobe exclusively flows and interbasaltic intervals, without implying thatthese necessarily coincided with any petrographic differences. As aresult of subsequent, and more detailed, investigations, petrographicand mineralogical compositional distinctions have now been drawnin respect of the lavas and sills and are shown in Figure 2. In summary,the different members are as follows:
Unit 1: A highly pyritized and calcified feldspathic basalt at thecontact of the volcanic sequence with the overlying oolitic grainstone.
Unit 2: A very highly altered basalt containing abundant ilmeniteand titanomagnetite in a matrix of alkali feldspar (e.g., Sample 143-866A-171R-3, 70-73 cm)
Unit 5: An olivine-phyric basalt with about 7% fresh olivinephenocrysts (Fo87_75, <3 mm) and smaller Plagioclase laths (An68^6,<0.5 mm) in a matrix of clinopyroxene (Ca46Mg3gFe16), magnetite,and feldspar. Cr-spinel crystals also occur (e.g., Sample 143-866A-177B-l,l-3cm).
Unit 6: This is distinguished by the abundance (15%) of Plagio-clase megacrysts and phenocrysts. Plagioclase megacrysts (<IO mm,An81_36) are accompanied by smaller Plagioclase laths (<2 mm,~An68) and olivines (<3 mm, Fo82_67) Examples of these plagioclase-phyric basalts are Samples 143-866A-180R-3, 116-20 cm, and-180R-4, 1-4 cm).
Units 8-11: Plagioclase megacrysts and/or phenocrysts are againcommon (-5%) in this group, but less abundant than in the overlyingbasalt described in Unit 6. The large plagioclases (An71_63) are locallyaccompanied by olivine (Fo86_74) and clinopyroxene megacrysts(Ca45Mg4!Fe13). Examples are Samples 143-866A-182R-1,102-105cm, and -185R-3, 88-92 cm.
Unit 12: Olivine-phyric basalt, similar to that of Unit 5. Olivines(<3 mm) make up about 8% of the volume and are accompanied byscattered feldspar megacrysts (<6 mm). Olivine compositions are inthe range Fo83_76, and the feldspars are about An70. As in Unit 5, someCr-spinels occur. Interstitial groundmass feldspar has a compositionofOr19Ab71An10.
Mineral Chemistry
A wide range of feldspar compositions is represented in the Reso-lution Guyot volcanic succession. Megacrysts and phenocrysts ofPlagioclase are common throughout the sequence of lavas. Zoned,mostly within the labradorite range, they attain their most calcic com-
PETROLOGY AND GEOCHEMISTRY OF IGNEOUS ROCKS
1640-
1660—
1680—
1700-
1720—
1740-
oC
170R
171RP
172R
173R
174R175R
10
11
12
T.D.=1743.6
PETROGRAPHIC SUMMARY
n/-/-A | Oolithic grainstone
VERY HIGHLY ALTERED BASALTAlk Fsp, Ti-Mt and Ilmenite
BRECCIA
BASALT DOMINATED BY PLAGIOCLASEMEGACRYSTS An81-36 + 01 Fo84-67
BRECCIA
PLAGIOCLASE-OLIVINE-PHYRIC BASALTSWITH MEGACRYSTS OF PLAG,OL AND CPX.OL. Fo 79-73PLAG An 73-51CPX. Ca45Mg41Fel3
LARGE REL. FRESH OL Fo 83-76PLAG. MEGACRYSTS An 67,ChromeSpinel.
Note: a represents red or brown soil/clay andrubbly, weathered surface of a lava flow
Figure 2. Stratigraphic succession in the volcanic base-ment of Resolution Guyot (Hole 866A) indicating majorpetrographic groupings distinguished within the lavas.
position (An8]) in the feldspar-phyric lavas of Unit 6. The wide rangeof feldspar compositions (Fig. 3 and Table 1) is partly attributable tomagmatic processes, but probably also is the consequence of sub-marine weathering and exchange. Some interstitial groundmass feld-spar is of albite composition and some is almost pure K-feldspar.
Olivine is invariably partially altered to, or wholly pseudomorphedby, serpentine minerals and iron oxides. The most magnesian compo-sitions occur in lavas where olivine attains its maximum modal abun-dance (i.e., Unit 5 [Fo87] and Unit 12 [Fo83]). Phenocrysts are normallyzoned to a minimum forsterite content of Fo65 at the rims (Table 2).
Clinopyroxene occurs as megacrysts in the central part of thevolcanic sequence (Units 8-11), and occasionally appears as micro-phenocrysts; it is also a ubiquitous groundmass constituent. The cli-nopyroxene megacrysts in Sample 143-866A-182R-1, 102-105 cm,have a composition of Ca45Mg42Fe13. In Sample 143-866A-185R-3,88-92 cm, the megacrysts show sector zoning from Ca43Mg42Fe15 toCa41Mg34Fe25, the latter being the most iron-rich pyroxene composi-tion in the Resolution Guyot volcanics. Atypical groundmass compo-sition is Ca44Mg39Fe17 (Sample 143-866A-177B-1, 1-3 cm). Projec-tions of pyroxene compositions in terms of relative proportions ofCa-Mg-Fe (Fig. 4) show a concentration of points in the augite fieldwith a small scattering toward Ca depletion and Fe enrichment. Thereis remarkably little overlap with the clinopyroxenes of Sites 865 and
869. Schweitzer et al. (1979) demonstrated how differences in pyrox-ene composition reflect differences in the type of basalt in which theyoccur. They point, for example, to higher Cr2O3 and lower TiO2 intholeiitic as opposed to alkalic pyroxenes. Higher TiO2 and lowerCr2O3 concentrations occur in the pyroxenes from Hole 865 A as com-pared with Hole 866A (Fig. 5 and Tables 3 and 4), indicative of themore alkalic nature of the former. The Site 866 pyroxenes tend to beconcentrated in the augite field with a trend toward subcalcic ferro-augite. Fodor et al. (1975) demonstrated the different chemical char-acteristics of pyroxenes from the tholeiitic, alkalic, and nephelinicsuites of Hawaii. On the ternary diagram (Fig. 4), pyroxenes of Site865 plot largely in the salite field and are generally more calcic thanthose of Site 866. A high Wo component in clinopyroxenes has beenshown by both Le Bas (1962) and Fodor et al. (1975) to be a featureof highly alkalic or nephelinic lavas.
Oxide phases are mainly titanomagnetite, but include ilmenite (e.g.,Samples 143-866A-177B-1, 1-3 cm and -185R-2, 62-66 cm) andchrome spinel (e.g., Sample 143-866A-185-3, 88-92 cm) (Table 5).
Much of the matrix of the lavas has been altered to smectite, asconfirmed by XRD. Electron microprobe analysis (e.g., 21% FeOand 15% MgO in Sample 143-866A-181-3, 91-96 cm) indicates aniron- and magnesium-rich clay, approximate in composition to fer-roan saponite.
247
RE. BAKER, P.R. CASTILLO, E. CONDLIFFE
Table 1. Representative microprobe analyses (wt% oxide) and atomic proportions of feldspars in basaltic rocks from Holes 865A and 866A.
Core, section:
Interval (cm):
SiO2
A12O3Fe2O3MnOMgOCaONa2OK2OTotal
SiAlFe<MnMgCaNaKTotal
AbOrAn
865A-94R-4
78-80
(Lath)
50.2930.610.660.060.12
14.762.820.38
99.70
2.3061.6540.0230.0020.0080.7250.2500.0224.990
25.112.20
72.69
865A-93R-3
13-17
(Lath)
50.7130.48
0.650.000.06
14.523.020.24
99.68
2.3211.6440.0220.0000.0040.7120.2680.0144.985
26.931.43
71.64
865A-93R-3
13-17
(Gmass)
64.0718.770.130.000.000.030.41
15.9499.36
2.9791.0280.0050.0000.0000.0020.0370.9454.996
3.7796.06
0.17
865A-91R-1
128-130
(Lath)
51.4929.60
0.720.090.10
13.773.490.51
99.75
2.3571.5960.0250.0030.0070.6750.3100.0305.003
30.512.95
66.54
866A-189R-1
66-69
(Gmass)
51.5829.47
0.900.030.16
13.583.870.17
99.96
2.3591.5880.0310.0010.0110.6650.3430.0105.008
33.691.00
65.31
866A-186R-3
12-15
(Large K fsp)
65.1118.280.080.000.000.000.22
16.57100.27
3.0040.9940.0030.0000.0000.0000.0200.9754.996
2.0098.00
0.00
866A-185R-3
88-92
(Gmass)
54.1427.81
1.280.110.11
11.704.940.22
100.29
2.4531.4850.0430.0040.0070.5680.4340.0135.007
42.781.26
55.96
866A-182R-3
48-51
(Phen core)
49.3931.080.880.000.29
15.402.710.09
100.31
2.2671.6810.0300.0000.0200.7570.2410.0065.002
24.020.55
75.43
866A-182R-1
102-105
(Phen core)
51.3529.760.740.080.06
13.983.680.10
99.75
2.3491.6040.0260.0030.0040.6850.3260.0065.003
32.060.58
67.37
866A-180R-4
1 ^
(Mega core)
48.8431.750.720.000.24
16.232.530.03
100.34
2.2351.7120.0250.0000.0160.7950.2250.0025.010
21.990.19
77.82
866A-180R-3
116-120
(Phen core)
50.4031.030.540.080.08
15.403.080.07
100.67
2.2911.6620.0180.0030.0050.7500.2710.0045.004
26.460.40
73.15
Note: Lath = Plagioclase lath, Gmass = groundmass, Large K fsp = large K feldspar, Phen core = phenocryst core, and Mega core = megacryst core.
Figure 3. Compositional variations of feldspars from Sites 865 and 866 withSite 869 for comparison.
Geochemistry
Representative whole-rock X-ray fluorescence (XRF) analysesfor major and trace elements are presented in Table 6. Except for somepieces of slightly altered basalts from Core 143-866A-177B, all of thevolcanic rocks have been moderately to completely altered, mainly toclay minerals. Thus, little significance can be attached to the presentconcentrations of the more mobile elements (e.g., K2O, Na2O). How-ever, plots of major oxides vs. stratigraphic height in the volcanicsuccession reveal broad compositional changes that are probablyinsensitive to finer-scale variations caused by alteration. In a plot ofMgO vs. depth (Fig. 6A), the only clear distinction is between a groupof MgO-rich lavas below 1730 mbsf and generally less magnesianlavas higher up the succession. The more magnesian lavas coincide
Mg Fe
Figure 4. Pyroxene quadrilateral showing compositions of clinopyroxenesfrom Sites 865 and 866, with samples from Site 869 for comparison.
with those identified as the olivine-phyric group of Unit 12, and themost magnesian of the remainder is equivalent to the thin olivine-phyric Unit 5. A cluster of relatively low-MgO, more differentiatedlavas occurs in the middle part of the succession (1690-1710 mbsf).In terms of A12O3 vs. depth (Fig. 6B), this also distinguishes anA12O3-depleted group, corresponding to the Unit 12 olivine-phyriclavas, and a cluster of Al2O3-enriched lavas (17%—18% A12O3), cor-responding to the plagioclase-phyric suite of Unit 6. A plot of Fe2O3/MgO vs. depth (Fig. 6C) separates the lowermost group (Unit 12)from the rest as having lower Fe/Mg ratios. From 1710 mbsf upward,there is a general, though not sharply defined, shift from higher tolower Fe/Mg ratios.
Studies of the alteration of submarine lavas (e.g., Cann, 1970; Hartet al., 1974) have shown that elements such as Ti, Zr, Nb, Y, and mostof the REE are not strongly affected during seawater alteration ofbasalt. In a plot of TiO2 vs. depth (Fig. 7A) three major geochemicalgroups are apparent: the Unit 12 lavas with low TiO2, the Units 8-11group with high TiO2, and a later group, above 1673 mbsf, whichagain has lower TiO2 but indicates a trend toward higher concentra-tions with decreasing depth. However, it is plots of Zr, Nb, and REEthat are the most illuminating. A plot of Nb/Zr vs. depth (Fig. 7B)distinguishes the uppermost plagioclase-phyric lavas from Unit 6from the rest of the lavas, because these samples are displaced tolower (0.13-0.15) Nb/Zr ratios. The olivine-phyric lavas from Units5 and 10 and the lowermost portion of Unit 12 have intermediate
PETROLOGY AND GEOCHEMISTRY OF IGNEOUS ROCKS
Table 1 (continued).
866A-179R-5 866A-177B-1 866A-177B-1
51-59 1-3 1-3
(Phen core) (Phen core) (Lath)
48.5430.95
0.660.100.11
15.952.820.06
99.18
2.2491.6910.0230.0040.0070.7920.2540.0035.023
24.180.32
75.50
51.0629.38
0.920.140.19
14.393.650.13
99.85
2.3401.5870.0320.0060.0130.7070.3240.0075.016
31.230.71
68.06
56.7226.13
0.620.030.119.606.120.26
99.59
2.5651.3930.0210.0010.0080.4650.5360.0155.004
52.761.47
45.77
0.20
1.7 1.8 1.9
Si (atomic prop.)
2.0
Figure 5. Ti vs. Si (cation proportions on basis of 6 oxygens) in clinopyroxenesfrom Holes 865A and 866A.
(0.15-0.18) Nb/Zr ratios, whereas the remainder of the volcanic suc-cession, consisting of Units 7,8,9, and 11 as well as the upper portionof Unit 12, has high (0.19-0.22) Nb/Zr ratios.
Chondrite-normalized patterns and total concentrations of REE(Fig. 8 A) reinforce the three-fold subdivision of the lavas drawn fromthe ratio of immobile incompatible elements Nb and Zr. The REE con-tents of the low Nb/Zr basalts from the uppermost part of the succes-sion (i.e., above 1673 mbsf) are only 8 to 60 times chondritic values.More importantly, they have flattened light REE patterns. The relativeenrichment of the light to middle REE, represented by La/Sm ratio,for these uppermost basalts is low (-1.4-1.8; Fig. IC). Conversely,the REE contents of the high Nb/Zr basalts from the middle and lowerportions of the succession are 10 to 130 times chondritic values.These samples also are enriched in light REE, having high (-2.4-2.9)La/Sm ratios. Lavas with intermediate Nb/Zr ratios from the middleto lower portions of the Hole 866A basement have 10 to 90 timeschondritic values in their REE contents; these lavas are only moder-ately enriched in light REE (La/Sm -0.15-0.18). It is important tonote, however, that the intermediate group generally overlaps withboth the low and high Nb/Zr groups, but more so with the former in
£ 1710Q.α>
Q 1730
1750
1650
1670
JS 1690
x: 1710 r+-•Q.α>Q 1730
6 8 10 12MgO (wt%)
1750
_
-
•
-
- Φ1 Φ
Φ
i
V
<
Φ
Φ
ΦΦ
i ,
Φ
Φ
#ΦΦ
#
i
ΦΦ
0.8 1.2 1.6Fe O /MgO
2 3 β
2.0 2.4
Figure 6. A. Concentration of MgO with depth in the volcanic rocks of Hole866A. Note the more magnesian lavas below 1730 mbsf and the cluster of moredifferentiated, low-MgO flows, around 1700 mbsf. B. Concentration of A12O3
with depth in the volcanic rocks of Hole 866A. C. Plot of Fe2O3/MgO (atomic)vs. depth in the volcanic rocks of Hole 866A.
terms of concentrations and light-REE enrichment. Specifically, theUnit 5 lavas belong to the less light REE-enriched group despite theirintermediate Nb/Zr ratios. The three-fold subdivision of Site 866samples into low Nb/Zr, high Nb/Zr, and intermediate Nb/Zr geo-chemical lava groups, as well as the conflicting behavior of the lattergroup, are clearly shown in a plot of La/Sm vs. Nb/Zr ratios (Fig. 9).
Plots of Nb and Zr vs. TiO2 (Figs. 10A and 10B) also clearly sep-arate the low Nb/Zr and high Nb/Zr Site 866 lavas, which are also evi-dent in Figure 11. The Plagioclase phyric lavas from above 1673 mbsfcontain lower Nb and Zr for given TiO2 than do the plagioclase-olivine-phyric basalts from below 1673 mbsf. The intermediate
RE. BAKER, P.R. CASTILLO, E. CONDLIFFE
Table 2. Representative microprobe analyses (wt% oxide) and atomic proportions of olivines from basaltic rocks in Hole 866A.
Core, section:
Interval (cm):
SiO2TiO2A12O3FeOMnOMgOCaONiOTotal
SiTiAlFe2
MnMgCaNi
TotalFo
189R-1
66-69
(Phen core)
39.700.010.06
15.360.22
43.590.290.24
99.47
1.0050.0000.0020.3250.0051.6450.0080.005
2.99583.3
189R-1
66-69
(Large unzoned)
38.460.040.04
21.390.28
39.030.260.11
99.61
1.0010.0010.0010.4660.0061.5140.0070.002
2.99876.2
182R-3
48-51
(Phen relict)
39.430.070.04
18.480.19
41.110.270.13
99.72
1.0090.0010.0010.3950.0041.5680.0080.003
2.98979.7
182R-3
48-51
(Phen rim)
38.080.020.04
23.800.41
36.120.330.12
98.91
1.0100.0000.0010.5280.0091.4280.0090.003
2.98872.7
182R-3
48-51
(Mega core)
39.180.050.03
19.850.35
40.350.270.19
100.26
1.0050.0010.0010.4260.0081.5420.0080.004
2.99578.1
182R-1
102-105
(Phen core)
40.920.010.09
13.420.09
45.220.240.34
100.33
1.0150.0000.0030.2780.0021.6720.0060.007
2.98385.7
182R-1
102-105
(Phen rim)
38.730.050.06
23.540.32
37.270.260.22
100.46
1.0090.0010.0020.5130.0071.4470.0070.005
2.99173.6
180R-4
(Phen core)
39.840.060.08
17.040.28
42.260.280.28
100.12
1.0090.0010.0020.3610.0061.5960.0080.006
2.98981.3
180R-4
1-4
(Phen rim)
37.230.050.04
28.560.41
32.580.410.10
99.37
1.0070.0010.0010.6460.0091.3130.0120.002
2.99166.7
180R-3
116-120
(Phen core)
40.450.050.05
13.320.22
45.550.250.33
100.22
1.0060.0010.0020.2770.0051.6890.0070.007
2.99485.7
180R-3
116-120
(Phen rim)
37.350.090.03
30.760.46
32.180.350.17
101.38
1.0000.0020.0010.6890.0101.2840.0100.004
3.00064.7
Note: Phen core = phenocryst core, Phen relict = phenocryst relict, Phen rim = phenocryst rim, and Mega core = megacryst core.
Table 3. Representative microprobe analyses (wt% oxide) and atomic proportions of clinopyroxenes from basaltic rocks in Hole 866A.
Note: Gmass = groundmass, Mega core = megacryst core, Mega rim = megacryst rim, Phen core = phenocryst core, and Microphen = microphenocryst.
Nb/Zr lavas surprisingly contain approximately constant Nb and Zrfor any given TiO2. An interesting feature shown by the high Nb/Zrlava group is that it displays a subtle, but nevertheless consistent,chemostratigraphy. The general trend is from lower to higher and thenback to lower Nb/TiO2 ratios upward through the succession, with themaximum value occurring at -1700 mbsf. This trend also is clearlyshown by the Nb/Zr and La/Sm ratios (Figs. 7B and IC). HigherNb/TiO2, Nb/Zr, and La/Sm ratios are associated with more alkalicrocks; thus, the progression is from the mildly alkalic compositionsat the bottom of Hole 866A to more alkalic in the middle, and finallyto more tholeiitic lavas higher up. The REE-concentration patterns ofSite 866 lavas (Fig. 8A) reinforce this notion.
SITE 865
A series of basaltic sills is intrusive into Albian clayey dolomiticlimestones between 830 and 870 mbsf in Hole 865A. Two whole-rock
K/Ar dates were determined on the sill rocks. The upper (Sample143-865A-93R, 13-17 cm) gave an age of 87 ± 3 Ma, and the second(Sample 143-865A-93R-4,78-80 cm), 10 m lower, an age of 102 ± 6Ma (Pringle et al., this volume). 40Ar/39Ar dates by the same authorsyield an age of 111.1 ± 2.6 Ma.
Several lines of evidence suggest that the sills were injected intounconsolidated sediments. For example, at the top of Unit 1 (Fig. 12)the overlying soft and muddy sediment becomes increasingly wellindurated toward the contact. The actual contact between the basaltand the sediment is irregular and, in some places, is almost vertical,rather than horizontal (Sager, Winterer, Firth, et al., 1993, p. 149, fig.33). Detached pieces of the clayey bioclastic limestone are includedwithin the basalt, indicating that the latter is younger. Also, there isusually a reaction halo in the basalt where it comes in contact with thesediment, presumably the result of hydrothermal alteration duringdevolatilization of the wet host material. In some places (e.g., the topof Unit 4), the basalt is clearly chilled against the clayey limestone. In
the interior of the basalt, dark reddish brown sedimentary xenolithsare completely decarbonated.
Petrography and Mineralogy
The basaltic sills are highly altered, and few of the original mineralconstituents remain unaffected. There are almost certainly oxidizedpseudomorphs after olivine in some specimens (e.g., Sample 143-865A-93R-3, 13-17 cm), but there is no fresh relict olivine. Clino-pyroxene occurs as a common groundmass constituent and is fairlywell preserved. The clinopyroxenes have relatively high Ca in com-parison with those from Hole 866A (Fig. 4 and Tables 3 and 4): theyrange in composition from Ca48Mg43Fe9 to Ca50Mg28Fe22 and arecommonly zoned. Much of the feldspar is probably primary, and lathsshow a compositional spread from An72 to An51. However, some ofthe interstitial feldspar is secondary (e.g., the alkali feldspar[Or96Ab4] in Sample 143-865A-93R-3, 13-17 cm). Opaque oxidesshow a variety of textural forms, some suggestive of rapid cooling.
1650 -
1670 -
(mbs
(ept
h
1690
1710
1730
ß
1750
2.0
1650 -
2.6 3.2
TiO (wt%)3.8
C* 1670</>
^ 1690
Q. 1710<DQ
1730 r
•9-
-
CD
oÅ
0.12 0.16 0.2 0.24
1650
1670
(0•| 1690
£ 1710Q.Φ
Q 1730
1750
°8V
I high Nb / Zr lavas
I intermediate Nb / Zr
TO low Nb / Zr lavas
1.5 2La / Sm
2.5
(N)
Figure 7. A. Plot of TiO2 vs. depth in Hole 866A. B. Plot of Nb/Zr vs. depth
in Hole 866A. C. Plot of La/Sm vs. depth in Hole 866A. In B and C, solid
circles represent a high Nb/Zr group, solid squares an intermediate Nb/Zr
group, and open circles a low Nb/Zr group.
Compositionally, they range from titanomagnetite to ilmenite (e.g.,Sample 143-865A-94R-4, 78-80 cm), and there are also somechrome spinels (e.g., Sample 143-865A-93R-3, 13-17 cm).
Geochemistry
The sills are of basaltic composition, but are so highly altered that,again, little reliance can be placed on concentrations of the moremobile elements. XRF whole-rock analyses of a selection of samplesfrom the sills are given in Table 7. The plot of total alkalis (Na2O +
251
RE. BAKER, PR. CASTILLO, E. CONDLIFFE
A 100it
es
Ch
on
dri
—».V)Φ
Sa
mp
l
200
100
100
10200
B100
10
Site 866 Samples :
Low Nb / Zr basalts
100
10
10
10200
100
A
Nb
(p
pm
)
B
"ET
(PP
'
N
90
80
70
60
50
40
30
20
1 0
400
350
300
250
200
150
LaCePrNd Sm Eu10
Ho Er Yb Lu
Figure 8. Chondrite-normalized REE patterns for groups of lavas from Hole
866A (A) and the basaltic sills of Hole 865A (B).
100
: Low Nb / Zr: • HighNb/Zr
' O Int. Nb/Zr
' D Site 865 basalts
Oo oo
o
•m
D
C
D
•
•
Q>
D
2.5 3 3.5
TiO (wt%)
Figure 10. A. Plot of Nb vs. TiO2 for lavas of Hole 866A and the Hole 865A
sills. B. Plot of Zr vs. TiO2 for Hole 865A sills and the lavas of Hole 866A.
z
N
13Z
0.26
0.24
0.22
0.20
0.18
0.16
0.14
0.12
:. •: o
D
-
Low Nb / Zr
High Nb / Zr
Int. Nb/Zr
Site 865 samples
C
•
CO
Unit 1 |Cj
D g• Units 3 &4
1.5 2 2.5La/Sm(N)
3.5
Figure 9. Plot of Nb/Zr vs. La/Sm for the volcanic rocks of Hole 866A.
K2O) vs. SiO2 (Fig. 13) suggests that Site 865 basalts are generallymore alkalic than the Site 866 basalts, but little significance canbe attached to this in view of the alteration. However, chondrite-normalized REE patterns of Site 865 basalts (Fig. 8B) are steep andmore enriched in light REEs than those of the Site 866 lavas. In fact,chondrite-normalized trace element patterns (Fig. 14) point to greaterenrichment in incompatible elements in Site 865 basalts comparedwith the both the Site 866 and Site 869 basalts. Although this must betreated with reservation because of the alteration, comparisons of
N
0.3
0.25
0.2
0.15
0.1
0.05
Mixing trends
• Low Nb / Zr
• HighNb/Zr
O Int. Nb/Zr
20 30 40Nb (ppm)
5 0 6 0
Figure 11. Plot of Nb/Zr vs. Nb for the volcanic rocks of Hole 866A.
such elements suggest that Site 865 basalts are more alkalic in com-position than any of the other igneous rocks drilled during Leg 143.Of more significance are the consistently higher concentrations andratios of the alteration-resistant incompatible elements (Nb, Zr, Ti, La,and Sm) in the Site 865 rocks. Plots of Nb/Zr vs. La/Sm (Fig. 9), Nbvs. TiO2 (Fig. 10A), and Zr vs. TiO2 (Fig. 10B) illustrate the lack ofoverlap between basalts from the two sites and show that the Site 865basalts have generally higher Nb/Zr, Nb/TiO2, and Zr/TiO2 ratios. Thetrace element signature of the Site 865 basalts is typical of highlyalkalic or nephelinitic rock suites.
Figure 12. Simplified stratigraphic section through the basaltic sills and theirhost rocks, 830-870 mbsf, Allison Guyot (Hole 865A).
Site 865 samples show a slight increase in La/Sm ratios with depth(Fig. 15A,) except for the two samples just below the sedimentaryhorizon located at the bottom of Section 143-865A-91R-3 to the topof Section 143-865A-92R-3 (i.e., samples from the top of Unit 3);these two samples have the lowest La/Sm ratios of all. Basalts fromthe thin (-15 cm.) Unit 2, which are apparently interbedded with thesediments, were not analyzed because these samples are extremelyaltered. It was noted before, however, that Unit 2 could have beencomposed of cobbles from Unit 1, because they lie on top of Section143-865A-92R (Sager, Winterer, Firth, et al., 1993). Interestingly, allUnits 3 and 4 samples have similar and low Nb/Zr ratios (0.21-0.22)that are clearly distinct from the Nb/Zr ratios (0.24-0.25) of Unit 1samples above the sediments (Figs. 9 and 15B). Unit 1 basalts alsohave generally higher Nb/TiO2 ratios than Units 3 and 4 basalts (Fig.10A). In summary, therefore, the general trend of Site 865 basalts istoward less alkalic composition with depth.
PETROGENESIS
The similar ratios of alteration-resistant trace elements of the geo-chemically enriched (i.e., high Nb/Zr ratio) Site 866 lavas clearlysuggest that samples belonging to the group are petrogenetically re-lated through a simple magmatic differentiation mechanism. The sameis true for the geochemically depleted (low Nb/Zr ratio) group. Therelationship among the mildly enriched (intermediate Nb/Zr) lavas isprobably more complex because their geochemical characteristicsoverlap with both the depleted and the enriched groups. Because of thealtered state of the majority of the lavas, it is difficult to evaluatequantitatively the main differentiation mechanism or mechanismsthat relate samples within and between the geochemical groups. How-ever, a plot of a ratio of two highly incompatible trace elements vs.the concentration of one of those elements can identify semiquantita-
Figure 13. Total alkalis (Na2O + K2O) vs. silica plotted for basaltic rocks of
Sites 865, 866, and 869.
400 n
fc•>
I 100o
58|
10
-\—i—i—i—r Site 865 basalts
Site 866 high &int. Nb/Zr
η Site 866 lowNb/Zr
Site 869 basalt
Ba Nb La Ce Pr Sr Nd Sm Zr Eu Tb Y Ho Er Yb Lu
Figure 14. Comparative chondrite-normalized trace-element patterns for ba-
salts from Sites 865, 866, and 869.
tively such mechanisms (e.g., Minster and Allegré, 1978). Figure 11,a plot of Nb/Zr ratios vs. Nb, shows that the enriched and depletedSite 866 lavas define crystal fractionation trends. Figure 11 alsoshows it is highly unlikely that the mildly enriched lavas have theirown crystal fractionation trend; rather, the lavas are probably theproducts of mixing between the first and second groups.
To test the crystal fractionation hypothesis for Site 866 lavas fur-ther, the major element contents of the least-altered samples, based onpetrography and low LOI contents, were modeled through the least-squares mixing program (Le Maitre, 1980) and using mineral analysesactually present in these rocks (Tables 1-5) as the fractionating phases.Two methods were employed: (1) the incremental step method, inwhich each differentiated rock was modeled by subtracting an appro-priate combination of minerals from a slightly more mafic rock and soon until the most differentiated rock was produced and (2) the cumu-lative step method, in which all differentiated rocks were modeledfrom a single most mafic magma. Representative results of the mod-eling are shown in Table 8. Basically, models involving samples thathave similar Nb/Zr ratios (i.e., models between samples from withinthe depleted and the enriched group) produce acceptable results. Theresidual sum of squares of the difference in the major element contentsof the actual parental rock and the parental magma predicted by themodel (Σ^,2) are generally <<O.l, except in models involving moredifferentiated samples, and the kind and proportion of minerals re-
253
RE. BAKER, P.R. CASTILLO, E. CONDLIFFE
Table 4. Representative microprobe analyses (wt% oxide) and atomic proportions of clinopyroxenes from basaltic rocks in Hole 865A. All are smallgrains in range 0.02-0.2 mm.
Note: All analyses are small grains ranging from 0.02 to 0.2 mm.
A
CO
*-<Q.α>Q
835
845
855
865
R7*;
-
-
--
•Unit 1 •Units 3 and 4 •
3.1 3.15 3.2 3.25 3.3 3.35 0.21
La/Sm ( N )
0.22 0.23
Nb/Zr0.24 0.25
(NJ
Figure 15. Plots of La/Sm (A) and Nb/Zr (B) vs. depth in Hole 865A.
moved from the parental magma are those actually present in theparental magma. Model results for samples belonging to the mildlyenriched group also produce acceptable results, but only for the incre-mental step method because the more mafic lavas of this group havehigh LOI and K2O contents (e.g., Samples 143-866A-189R-3,2-4 cm,and -189R-4,11-14 cm). Surprisingly, cumulative step results betweenmafic depleted basalts and differentiated mildly alkalic lavas are alsoacceptable. Other intragroup models produce unreasonable results.
The Site 865 basalts are more difficult to evaluate for crystal frac-tionation because, in addition to having fewer samples, they are gen-erally more altered than Site 866 basalts. Specifically, the more maficbasalts have high K2O (and LOI) values so that results of cumulativemodels are unreasonable. Incremental model results for the differen-tiated basalts from Units 3 and 4 are reasonable.
In summary, major-element modeling results are consistent withthe crystal fractionation origin of the intergroup Nb variations, shownby Site 866 depleted and enriched lavas. Modeling results for themildly enriched lavas suggest that these rocks are possibly related
through crystal fractionation, but most probably are also affected to alarge extent by mixing between geochemically depleted and enrichedmagmas. The intermediate behavior of the trace-element contents ofthe mildly enriched group, between the enriched and the depletedgroup (e.g., Figs. 7-10), is also consistent with this mixing scenario.The geochemical variations of Site 865 basalts are subject to greateruncertainty because of alteration, but some may be attributable tosimple crystal fractionation.
The above interpretations are consistent with the Sr, Nd, and Pbisotopic ratios of the samples (P.R. Castillo et al., unpubl. data), par-ticularly those for the Site 866 lavas. The Pb and Sr isotopic ratios ofthe enriched basalts from Site 866 are generally higher than those ofthe depleted basalts; the Sr and Pb isotopic ratios of the mildly enrichedgroup lie between the first two groups, supporting the suggestion frommajor- and trace-element data that they are likely to be mixing prod-ucts. Site 865 basalts exhibit a range of isotopic values, but these donot show any systematic difference between the less alkalic Units 3 and4 and the more alkalic Unit 1.
Figure 16. Plot of Nb vs. Zr for basaltic rocks of the Leg 143 sites, comparedwith lavas from other South Pacific seamounts and archipelagos. Sources ofdata as follows: Austral (Palacz and Saunders, 1986); Easter (Baker et. al.,1974); Marquesas (Woodhead, 1992); Marshall (Davis et al., 1989); MORB(Sun and McDonough, 1989); Pitcairn seamounts (Woodhead and Devey,1993); Samoan seamounts (Johnson et al., 1986).
DISCUSSION
The basement of Site 866 was originally considered to be a se-quence of subaerial lava flows with interbasaltic intervals (Sager,Winterer, Firth, et al., 1993). The latter were of various types, inter-preted as rubbly aa, scoriaceous talus, debris flows, or boles. How-ever, neither the radiometric dating (Pringle et al., this volume) northe geochemical variations support this contention. Deviations froma systematic progression may be explained by the occurrence ofintrusives within the lava pile, although evidence for this remainsindirect as no intrusive contacts were recognized in the core.
Fractional crystallization is the dominant magmatic differentiationmechanism responsible for the variation of major- and trace-elementcontents of samples within different rock groups. These groups aremost clearly distinguished on the basis of REE, Nb/Zr, La/Sm, andother trace-element criteria. However, the different groups must havecome from different mantle sources, as implied by the geochemical
evidence presented here and confirmed by their Sr, Nd, and Pb isotopiccompositions (P.R. Castillo et al., unpubl. data).
Allison and Resolution seamounts originated within the regionof the South Pacific Isotopic and Thermal Anomaly or SOPITA(Staudigel et al., 1991). SOPITA embraces the area previously desig-nated the South Pacific Superswell (McNutt and Fischer, 1987) andincludes a part of the Dupal Anomaly (Hart, 1984). It incorporates aseries of island hotspots whose lavas bear the imprint of both HIMU(high U/Pb) and EMU (enriched mantle) source regions, as defined,for example, in Zindler and Hart (1986). The proportions of thesecomponents vary across the region, with EMU prevailing in the north-ern islands (Samoa-Marquesas) and HIMU in the more southerlyhotspots, such as the Cook-Austral groups (Staudigel et al., 1991).From their work on Cretaceous seamounts, the same authors estab-lished the longevity of the anomaly by demonstrating that isotopicallydistinctive lavas have been generated at SOPITA for at least 120 m.y.A more recent study by Castillo et al. (1992) pushed the isotopicanomaly even farther back, to -160 Ma. However, evidence suggeststhat the degree of enrichment or proportions of mantle componentsmay not have remained constant over this long period of time. For thisreason, trace element and isotopic characteristics of Cretaceous sea-mounts may not necessarily be used to relate them to specific present-day hotspots or island groups within the SOPITA area.
Lineaments derived by backtracking seamounts in the Pacific hot-spot reference frame (Duncan and Clague, 1985; Smith et al., 1989)point to an origin of Resolution and Allison seamounts in the vicinityof the Tubuai, Society, or Tuamotu islands. Despite reservations aboutcompositional changes with time, this link is supported by compara-tive plots of the immobile incompatible elements, such as Nb and Zr(Fig. 16 ). The Site 865 and 866 sills and lavas are more closelycoincident with the Austral Islands lavas than with those of any othergroup. There is a certain amount of scatter, and the Samoan seamounts(Johnson et al., 1986) separate into trends with both high and lowNb/Zr ratios. However, in general, high Nb/Zr ratios are characteristicof the more westerly islands and lower ratios prevail in the east(Marquesas, Pitcairn seamounts, and Easter Island). Resolution andAllison guyots, together with the Austral Islands, lie both geographi-cally and compositionally in the middle. In a broad sense, the inci-dence of more extreme alkaline compositions seems to increasetoward the west across the SOPITA hotspots and their derivativeseamounts. Similar east-west compositional gradations were reportedby Palacz and Saunders (1986), who found that westward the islandsbecome more enriched in large ion lithophile elements (LILE) andthat they are isotopically enhanced in the Dupal components (high87Sr/86Sr, low 143Nd/144Nd, and low 207Pb/206Pb).
Judging by their inferred tracks (Duncan and Clague, 1985), theResolution and Allison seamounts may have followed essentially thesame path and, therefore, may have been fed at different times by thesame hotspot. The lavas of Resolution Guyot show an upward shiftfrom more alkaline to more tholeiitic compositions, which may re-flect an increase in partial melting as the volcano approached andpassed over a hotspot. On this basis, the alkaline sills of Allison Guyotmay represent residual activity after the seamount had passed over ahotspot. But this scenario seems unlikely because light REE and otherincompatible elements are more enriched as well, and the ratios tendto be slightly higher in the Site 865 sills than in the Site 866 lavas. Thehigher La/Sm and La/Yb ratios in the Site 865 basalts may be indica-tive of a greater depth of melt segregation. Moreover, the Sr, Nd, andPb isotopic ratios of Resolution Guyot have more HIMU componentsthan those of Allison (P.R. Castillo et al., unpubl. data). Thus, the com-positional variations might be an expression of influx from two man-tle components (Palacz and Saunders, 1986) whose relative contribu-tions varied at different hotspots. It is also possible that the guyotseach passed over at least two distinct hotspots. In the case of AllisonGuyot, where the volcanic basement was not reached during drilling,the sills may represent a later phase of activity unrelated to the mainphase of volcanism.
255
RE. BAKER, P.R. CASTILLO, E. CONDLIFFE
Table 5. Representative microprobe analyses (wt%) and atomic proportions of Fe-Ti oxides and spinels from Hole 866A.
We thank members of the technical staff of the Department ofEarth Sciences, Leeds University, and Scripps Institution of Ocean-ography for the preparation of thin sections and rock crushing. Inparticular, we thank Alan Gray for the XRF analyses and ElizabethKristofetz for the ICP-MS analyses. Improvements to the manuscriptwere made on the basis of helpful comments by Andrew Saunders,John Sinton, and Sondra Stewart.
REFERENCES*
Baker, P.E., Buckley, R, and Holland, J.G., 1974. Petrology and geochemistryof Easter Island. Contrib. Mineral. Petrol, 44:85-100.
Cann, J.R., 1970. Rb, Sr, Y, Zr, and Nb in some ocean floor basaltic rocks.Earth Planet. Sci. Lett., 10:7-11.
Castillo, PR., Floyd, RA., and France-Lanord, C, 1992. Isotope geochemistryof Leg 129 basalts: implications for the origin of the widespread Creta-ceous volcanic event in the Pacific. In Larson, R.L., Lancelot, Y., et al.,Proc. ODP, Sci. Results, 129: College Station, TX (Ocean Drilling Pro-gram), 405-413.
Davis, A.S., Pringle, M.S., Pickthorn, L.B.G., Clague, D.A., and Schwab,W C , 1989. Petrology and age of alkalic lava from the Ratak Chain of theMarshall Islands. J. Geophys. Res., 94:5757-5774.
Duncan, R.A., and Clague, D.A., 1985. Pacific plate motion recorded by linearvolcanic chains. In Nairn, A.E.M., Stehli, F.G., and Uyeda, S. (Eds.), TheOcean Basins and Margins (Vol. 7A): The Pacific Ocean: New York(Plenum), 89-121.
Fodor, R.V., Keil, K., and Bunch, T.E., 1975. Contributions to the mineral chem-istry of Hawaiian rocks. IV. Pyroxenes in rocks from Haleakala and WestMaui volcanoes, Maui, Hawaii. Contrib. Mineral. Petrol, 50:173-195.
Abbreviations for names of organizations and publications in ODP reference lists followthe style given in Chemical Abstracts Service Source Index (published by AmericanChemical Society).
Hart, S.R., 1984. A large-scale isotope anomaly in the Southern Hemispheremantle. Nature, 309:753-757.
Hart, S.R., Erlank, A.J., and Kable, E.J.D., 1974. Sea floor basalt alteration:some chemical and Sr isotopic effects. Contrib. Mineral. Petrol, 44:219-230.
Johnson, K.T.M., Sinton, J.M., and Price, R.C., 1986. Petrology of seamountsnorthwest of Samoa and their relation to Samoan volcanism. Bull. Vol-canol, 48:225-235.
Le Bas, MJ., 1962. The role of aluminum in igneous clinopyroxenes withrelation to their parentage. Am. J. Sci., 260:267-288.
LeMaitre, R.W., 1980. A generalized petrological mixing model program.Comput. Geosci., 7:229-247.
McNutt, M.K., and Fischer, K.M., 1987. The South Pacific superswell. InKeating, B.H., Fryer, P., Batiza, R., and Boehlert, G.W. (Eds.), Seamounts,Islands, and Atolls. Geophys. Monogr., Am. Geophys. Union, 43:25-34.
Minster, J.F., and Allegré, C.J., 1978. Systematic use of trace elements inigneous processes. Part III: inverse problem of batch partial melting involcanic suites. Contrib. Mineral Petrol, 69:37-52.
Palacz, Z.A., and Saunders, A.D., 1986. Coupled trace element and isotopeenrichment in the Cook-Austral-Samoa islands, southwest Pacific. EarthPlanet. Sci. Lett., 79:270-280.
Sager, W.W., Winterer, E.L., Firth, J.V., et al., 1993. Proc. ODP, Init. Repts.,143: College Station, TX (Ocean Drilling Program).
Schweitzer, E.L., Papike, J.J., and Bence, A.E., 1979. Statistical analysis ofclinopyroxenes from deep-sea basalts. Am. Mineral, 64:501-513.
Shipboard Scientific Party, 1981. Site 463: western Mid-Pacific Mountains. InThiede, J., Valuer, T.L., et al., Init. Repts. DSDP, 62: Washington (U.S.Govt. Printing Office), 33-156.
Smith, W.H.F., Staudigel, H., Watts, A.B., and Pringle, M.S., 1989. TheMagellan Seamounts: Early Cretaceous record of the South Pacific iso-topic and thermal anomaly. J. Geophys. Res., 94:10501-10523.
Staudigel, H., Park, K.-H., Pringle, M., Rubenstone, J.L., Smith, W.H.F., andZindler, A., 1991. The longevity of the South Pacific isotopic and thermalanomaly. Earth Planet. Sci. Lett., 102:24^4.
Sun, S.-S., and McDonough, W.F., 1989. Chemical and isotopic systematicsof oceanic basalts: implications for mantle composition and processes. In
PETROLOGY AND GEOCHEMISTRY OF IGNEOUS ROCKS
Saunders, A.D., and Norry, MJ. (Eds.), Magmatism in the Ocean Basins. Zindler, A., and Hart, S., 1986. Chemical geodynamics. Annu. Rev. EarthGeol. Soc. Spec. Publ. London, 42:313-345. Planet. Sci., 14:493-571.
Woodhead, J.D., 1992. Temporal geochemical evolution in oceanic intra-platevolcanics: a case study from the Marquesas (French Polynesia) and com-parison with other hot spots. Contrib. Mineral. Petrol, 111:458-467.
Woodhead, J.D., and Devey, C.W., 1993. Geochemistry of the Pitcairn sea- D a t e o f i n i t i a l receipt: 1 December 1993mounts, I: source character and temporal trends. Earth and Planet. Sci. Date of acceptance: 6 July 1994Lett., 116:81-99. Ms 143SR-216
257
RE. BAKER, P.R. CASTILLO, E. CONDLIFFE
Table 6. Whole-rock XRF and ICP-MS analyses of basaltic lavas from Hole 866A.
Core, section:
Interval (cm):
SiO2TiO^A12O3Fe2O3MnO'MgOCaON a , 0K 20P2O5LOITotal
171R-3
70-73
49.632.91
17.2911.500.052.411.220.858.880.285.10
100.12
171R-3
143-146
50.252.77
16.319.540.081.682.590.68
10.070.225.97
100.16
Rare earth elements (in ppm) by ICP-MS:LaCePrNdSmEuTbHoErYbLu
1 1.326.1
3.5217.13.811.250.620.701.791.260.1 6
12.225.804.06
19.24.641.430.700.711.731.360.19
Other trace elements (in ppm) by XRF:BaCoCrCuNbNiRbSeSrThVYZnZr
20347
5943722
386592963
5151
19168147
25440
6026922
297512271
6210
2076
143
174R-1
3-5
47.052.65
15.2510.800.156.41
11.073.000.550.303.06
100.29
14.737.2
5.1025.0
5.931.900.911.062.922.450.40
11253
4168124
2468
24580
7251
27100168
177R-1
10-13
46.842.76
15.4213.110.176.989.662.980.470.291.79
100.472
14.435.04.65
22.34.971.830.851.102.972.340.35
9254
1767621
1355
24396
6306
2799
151
179R-1
108-112
45.513.42
15.1212.220.257.457.013.250.830.264.56
99.88
13.634.64.81
24.05.841.940.900.962.452.090.28
10961
2925326
1995
24552
6274
26109160
I79R-2
13-18
46.873.45
15.2812.490.237.986.023.260.750.253.60
100.16
13.433.5
4.7222.7
5.301.950.850.912.461.910.24
11661
3025527
2007
22538
6289
25111164
179R-2
137-140
47.232.52
16.6111.290.389.422.482.322.4550.215.68
100.72
9.724.7
3.4316.63.951.540.640.762.201.680.25
14553
207120
17137
1629
2836
3001982
128
179R-5
32-35
46.442.75
16.8411.520.247.058.203.010.710.253.16
100.17
11.429.8
4.1520.34.941.730.800.952.421.940.31
9950
1828020
1367
26468
6
2791
145
179R-5
51-59
46.622.61
16.8011.390.246.739.943.030.410.241.79
99.80
11.329.04.04
19.94.551.690.720.892.521.970.26
9646
1756718
1366
19480
8259
2484
134
179R-5
131-135
46.212.51
16.6911.160.166.24
10.712.800.400.242.33
99.45
10.827.83.81
18.54.451.600.670.862.401.880.26
9748
1627717
1315
14495
7241
2479
130
Notes: Major oxides in weight percent (%), and trace elements in parts per million (ppm). LOI = loss on ignition.
Reasonable fitReasonable fitReasonable fitReasonable fit
N o t g o o d b a n d c
Not goodc
Reasonable fitReasonable fit
Not goodc
Not goodc
N o t g o o d b a n d c
Not goodc
Not goodc
Not goodc
N o t g o o d b a n d c
N o t g o o d b a n d c
N o t g o o d b a n d c
Reasonable fit
Note: F (%) = amount of liquid remaining in the system after the removal of crystals. Oliv = olivine, Cpx = clinopyroxene, Plag = Plagioclase, Tmag = titanomagnetite, Ilm = ilmenite,and Orth = orthoclase.
a Orthoclase is generally a secondary mineral and is usually not present in fresh mafic lavas.b Large residual errors.c Unlikely mineral assemblage.
