Petrogenesis of lavas from Detroit Seamount: Geochemical differences between Emperor Chain and Hawaiian volcanoes



[1] The Hawaiian Ridge and Emperor Seamount Chain define a hot spot track that provides an 80 Myr record of Hawaiian magmatism. Detroit Seamount (∼76 to 81 Ma) is one of the oldest Emperor Seamounts. Volcanic rocks forming this seamount have been cored by the Ocean Drilling Program at six locations. Only tholeiitic basalt occurs at Site 884 on the eastern flank and only alkalic basalt, probably postshield lavas, occurs at Sites 883 and 1204 on the summit plateau. However, at Site 1203 the basement core (453 m penetration) includes four thick flows of pahoehoe alkalic basalt underlying ∼300 m of volcaniclastic rocks interbedded with submarine erupted tholeiitic basalt. The geochemical characteristics of these alkalic lavas indicate that phlogopite was important in their petrogenesis; they may represent preshield stage volcanism. The surprising upward transition from subaerial to submarine eruptives implies rapid subsidence of the volcano, which can be explained by the inferred near-ridge axis setting of the seamount at ∼80 Ma. A near-ridge axis setting with thin lithosphere is also consistent with a shallow depth of melt segregation for Detroit Seamount magmas relative to Hawaiian magmas, and the significant role for plagioclase fractionation as the Detroit Seamount magmas evolved in the crust. An important long-term trend along the hot spot track is that 87Sr/86Sr decreases in lavas erupted from ∼40 to 80 Ma. Tholeiitic basalt at Site 884 on Detroit Seamount is the extreme and overlaps with the 87Sr/86Sr-143Nd/144Nd field of Pacific mid-ocean ridge basalts (MORB). Complementary evidence for a depleted component in Detroit Seamount lavas is that relative to Hawaiian basalt, Detroit Seamount lavas have lower abundances of incompatible elements at a given MgO content. These lavas, especially from Sites 883 and 884, trend to extremely unradiogenic Pb isotopic ratios which are unlike MORB erupted at the East Pacific Rise. A component with relatively low 87Sr/86Sr and 206Pb/204Pb is required. Lavas erupted from a spreading center in the Garrett transform fault, 13°28′S on the East Pacific Rise, have this characteristic. A plausible hypothesis is mixing of a plume-related component with a component similar to that expressed in lavas from the Garrett transform fault. However, basaltic glasses from Detroit Seamount also have relatively high Ba/Th, a distinctive characteristic of Hawaiian lavas. We argue that all Detroit Seamount lavas, including those from Site 884, are related to the Hawaiian hot spot. Rejuvenated stage Hawaiian lavas also have high Ba/Th and define a trend to low 87Sr/86Sr and 206Pb/204Pb. We speculate that rejuvenated stage lavas and Detroit Seamount lavas sample a depleted mantle component, intrinsic to the plume, over the past 80 Myr.

1. Introduction

[2] The linear intraplate Hawaiian Ridge and Emperor Seamount Chain define an approximately 6000 km long, age-progressive hot spot track [Wilson, 1963], ranging from active volcanoes at the southeastern end of the Hawaiian Ridge to >80 Myr old extinct volcanoes at the northern end of the Emperor Seamount Chain (Figure 1a). Hawaiian volcanoes are predominantly constructed by eruptions of basaltic magma with major and trace element compositions and radiogenic isotopic ratios that distinguish them from mid-ocean ridge basalt (MORB) [e.g., Basaltic Volcanism Study Project, 1981]. An understanding of the origin and evolution of Hawaiian magmatism requires documentation and understanding of geochemical variations in hot spot-derived lavas on both short-term (<1 Myr) and long-term (>10 Myr) time scales. On a time scale of ∼1 Myr, volcanoes forming the Hawaiian islands evolve through a sequence of growth stages, characterized by the transition from alkalic to tholeiitic and back to alkalic magmatism [e.g., Clague and Dalrymple, 1987]. These transitions are consistent with passage of the Pacific plate over the hot spot [e.g., Chen and Frey, 1985]. Although there have been numerous studies of young volcanoes exposed on the Hawaiian Islands [e.g., Rhodes and Lockwood, 1995] (see Hawaii Scientific Drilling Project theme at, much less is known about the geochemistry of lavas forming the older volcanoes of the submarine Emperor Seamount Chain. Are lavas forming the Emperor Seamounts geochemically similar to lavas forming Hawaiian volcanoes, and did these old volcanoes evolve through a similar sequence of growth stages?

Figure 1a.

The Hawaii-Emperor Seamount Chain and the location of Detroit Seamount. Modified from Tarduno et al. [2002] by adding new 40Ar-39Ar ages for Detroit Seamount, Nintoku Seamount, and Koko Seamount from Duncan and Keller [2004] and for Suiko Seamount from Sharp and Clague [2002]. Magnetic anomaly identifications are following Mueller et al. [1997].

[3] Detailed study of lavas forming the Emperor Seamounts requires drilling to penetrate the thick sedimentary cover and to obtain samples whose relative ages are known. The first detailed sampling of the Emperor Seamounts was during Deep Sea Drilling Project Leg 55, which penetrated 387 m of basalt at Suiko Seamount (Figure 1a). Dalrymple et al. [1980] and Sharp and Clague [2002] reported ages of 64.7 ± 1.1 Ma and 61.3 ± 0.5 Ma, respectively, for Suiko Seamount lavas. Important results are that: (1) like Hawaiian volcanoes, Suiko Seamount has shield stage lavas of tholeiitic basalt overlain by alkalic basalt [Kirkpatrick et al., 1980]; (2) Initial 87Sr/86Sr ratios of tholeiitic basalt decrease from Daikakuji Seamount at the Hawaiian-Emperor bend to Suiko Seamount, which represent over 20 Myr of hot spot activity (∼42–61 Ma, Figure 1a) [Lanphere et al., 1980].

[4] More recently, the Ocean Drilling Program (ODP) Leg 145 drilled and cored the older Detroit Seamount (∼76–81 Ma, Table 1, Figures 1a and 1b) [Keller et al., 1995; Duncan and Keller, 2004]. The basaltic core from Site 883 and especially Site 884 of ODP Leg 145 yielded surprising results. Relative to primitive mantle and tholeiitic basalt from younger volcanoes on the Hawaiian Ridge-Emperor Seamount Chain, the ∼81 Ma basalt from Site 884 has lower ratios of highly to moderately incompatible elements, lower initial 87Sr/86Sr and higher initial 143Nd/144Nd ratios; in fact, Site 884 lavas are similar to MORB [Keller et al., 2000; Regelous et al., 2003]. These results have led to two contrasting interpretations for the origin and evolution of magma forming Detroit Seamount. Keller et al. [2000] noted that at ∼80 Ma, the Hawaiian hot spot was located near an active mid-ocean ridge [see Cottrell and Tarduno, 2003, Figure 5c]. They suggest that the MORB-like isotopic signature of Detroit Seamount lavas from Site 884 resulted from entrainment of MORB-related upper mantle (i.e., hot low viscosity asthenosphere and hot thin lithosphere) by the rising mantle plume. In contrast, Regelous et al. [2003] propose that the depleted characteristics of Site 884 lavas reflect sampling of a depleted and refractory plume component by unusually high extents of mantle melting beneath young and thin oceanic lithosphere.

Figure 1b.

Bathymetry of Detroit Seamount and drill site locations. Detroit Seamount refers to the region north of the red dashed line. Also shown is the outline of the Big Island of Hawaii (thick green line).

Table 1. Sampling of Detroit Seamount
HolePenetration Into Basement, mBasement Recovery, %
  • a

    The penetration of basement was 60.8 m at Hole 1204A. However, there was no recovery in the lower 22.4 m due to a clogged bit. The basement recovery rate is based on the upper 38.4 m. See Shipboard Scientific Party [2002b] for details.


[5] During ODP Leg 197 three additional cores with basement penetrations of 61 to 453 m were recovered from Detroit Seamount (Table 1) [Tarduno et al., 2002]. Lavas from Site 1203 yield 40Ar-39Ar ages of 76 ± 1 Ma [Duncan and Keller, 2004]. In this paper, we report major and trace element abundances for basaltic whole rocks recovered during ODP Leg 197 and basaltic glasses recovered during ODP Legs 197 and 145. In addition, Sr, Nd and Pb isotopic ratios are reported for whole-rock samples from Leg 197. We use these geochemical data, along with the petrographic and volcanological information to constrain the magmatic evolution at Detroit Seamount and to evaluate the two alternative hypotheses for the origin of its magmas.

[6] Important results are that (1) alkalic basalt at Sites 883 and 1204 may reflect postshield stage lavas, but the upward transition from intercalated alkalic and tholeiitic basalt to solely tholeiitic basalt at Hole 1203A may reflect the preshield to shield stage transition; (2) compared with Hawaiian shield stage lavas, tholeiitic basalt magmas erupted at Detroit Seamount segregated at a lower mean pressure; (3) Detroit Seamount lavas sampled a depleted component (i.e., low 87Sr/86Sr, Pb isotopic ratios and high 143Nd/144Nd) that is not present in Hawaiian shield stage lavas or East Pacific Rise (EPR) MORB. It is also present in Hawaiian rejuvenated stage and North Arch lavas [Frey et al., 2005]. We speculate that this depleted component has been associated with the Hawaiian hot spot for ∼80 Myr.

2. Detroit Seamount

2.1. Bathymetry

[7] A region of shallow seafloor, depths of 2400 to 3000 m, centered at 51°N in the northern part of the Emperor Seamount Chain has been referred to as Detroit Tablemount or Detroit Plateau [Lonsdale et al., 1993]. We refer to this region as Detroit Seamount (Figures 1a and 1b). The area, 10,000 km2, of the Detroit Seamount is similar to the subaerial part of the Big Island of Hawaii (Figure 1b). Like the Big Island, Detroit Seamount probably represents several coalesced volcanic shields [Duncan and Keller, 2004]. Three peaks, Detroit, Windsor and Wayne, reach shallower depths, up to 1388m (Figure 1b). Dredging of the northerly peak recovered lapilli hyalotuffs that included nepheline and perovskite-bearing lava fragments that were interpreted as rejuvenated stage volcanism, analogous to the highly alkalic rejuvenated stage of some Hawaiian volcanoes [Lonsdale et al., 1993]. These unsedimented peaks were avoided during ODP drilling whose objective was to sample lavas erupted during the shield construction stage; such lavas are buried beneath 460 to 850 m of sediment.

2.2. Stratigraphy and Petrography

2.2.1. Hole 1203A

[8] Hole 1203A, located in a valley on the eastern flank of Detroit Seamount, penetrated 452.6 m into basement, and recovered parts of 18 lava flow units (Figures 1b and 2). Five samples from four tholeiitic units (Units 11, 14, 16 and 24) have been dated using the 40Ar-39Ar technique, and yield an age of 76 ± 1 Ma [Duncan and Keller, 2004]. From top to bottom, these lava flow units range from dominantly non-vesicular pillow lavas to dominantly vesicular compound pahoehoe lavas. Glassy lobe margins are present in all lava units except for Units 6 and 24 whose lobe margins were not recovered. Unaltered glass occurs in the lobe margins throughout the upper pillow lava units; but only altered glass occurs in the lower pahoehoe lava units. Vesicles are variably filled with secondary minerals, such as carbonate, zeolites and sulfides. The vesicular compound pahoehoe lavas are inferred to have been erupted subaerially, whereas the upper non-vesicular pillow lavas and some of the associated volcaniclastic rocks (Figure 2) are inferred to be submarine eruptives [Shipboard Scientific Party, 2002a] (Figure 3). Volatile analyses of glasses from lobe margins of these pillow lavas indicate that they were erupted at variable water depths. In detail, uppermost pillow lavas, Units 3 and 8, are undegassed; hence they were erupted in water depths greater than 500 m. In contrast, the deeper pillow lavas, Units 14 and 18, are partially degassed, and inferred eruption water depths are less than 500 m (T. Thordarson et al., manuscript in preparation, 2004).

Figure 2.

Stratigraphy of basement core at Holes 1203A, 1204A, and 1204B. Unit boundaries, rock types, presence of glass, and phenocryst abundance are indicated. Modified from Tarduno et al. [2002].

Figure 3.

Inferred volcanic environment of four drill sites at Detroit Seamount. Modified from T. Thordarson et al. (manuscript in preparation, 2004). See text for details.

[9] The subsidence indicated by the upward transition from subaerial to submarine eruptives differs from that of a typical Hawaiian volcano. For example, submarine lavas underlie subaerial lavas in the drill core for Mauna Kea volcano recovered by the Hawaii Scientific Drilling Project (HSDP) [DePaolo et al., 1999]. That is, the accumulation rate, 9 to 18 mm/yr (W. D. Sharp and P. Renne, 40Ar/39Ar dating of core recovered by the Hawaii Scientific Drilling Project (phase 2) Hilo, Hawaii, submitted to Geochemistry, Geophysics, Geosystems, 2004), for the HSDP drill core from Mauna Kea was greater than subsidence rate of the Big Island, ∼2.4 mm/yr [Moore, 1987]. The dominance of subsidence rate over accumulation rate at Detroit Seamount can be explained by a lower accumulation rate and the proximity of the seamount to an active spreading center at ∼80 Ma [Rea and Dixon, 1983; Mammerickx and Sharman, 1988; Cottrell and Tarduno, 2003]. The relatively thin and weak oceanic lithosphere close to a spreading center subsides rapidly [e.g., Parsons and Sclater, 1977; Stein and Stein, 1992], so it is possible that during the growth of Detroit Seamount the subsidence rate was greater than the volcano accumulation rate.

[10] Lava flows from Hole 1203A are aphyric to olivine-and/or-plagioclase-phyric basalt (Table 2a). Plagioclase and olivine phenocrysts commonly occur as glomerocrysts. Large plagioclase phenocrysts/glomerocrysts, up to 1.3 cm in length, occur in Units 14 and 31. Trace amounts (<5%) of clinopyroxene phenocrysts occur in several units (Table 2a). Some plagioclase phenocrysts have rounded or embayed margins, which suggests that they are not in equilibrium with surrounding groundmass. Olivine rich zones, containing >10% olivine by volume, occur in Units 11 and 16. Some olivine phenocrysts are unaltered, but many are altered to clay, iddingsite, Fe-oxyhydroxide and carbonate, and they are recognized by their equant form and characteristic fracture patterns. The groundmass of these lavas consists of plagioclase, clinopyroxene, opaque minerals (mostly titanomagnetite) and glass (usually devitrified).

Table 2a. Whole-Rock Sample Information
Sample NameUnitUnit NameaBasalt TypebLava Flow TypeaUnit Thickness,a mDepth,a mbsfPhenocryst,c %
  • a

    Taken from Tarduno et al. [2002] with modification.

  • b

    See section 5.1 for details.

  • c

    Phenocryst content is estimated in volume%.

  • d

    Phenocrysts are low first-order yellow to gray and biaxial. They may be orthopyroxene.

  • e

    This sample contains a single clinopyroxene phenocryst (2.5 mm in diameter), which is attached to a plagioclase phenocryst.

1203A 17R4W 43-471highly plagioclase-olivine-phyric basalttholeiiticpillow lava7.95457.991040
1203A 18R3W 116-1193highly plagioclase-olivine-phyric basalttholeiiticpillow lava24.60467.961031d
1203A 20R3W 10-143  pillow lava 486.1949
1203A 25R1W 37-415moderately plagioclase-olivine-phyric basalttholeiiticsheet lobe1.94531.47000
1203A 26R2W 5-96sparsely plagioclase-phyric basalttholeiiticsheet lobe4.78535.84000
1203A 26R3W 97-1016  sheet lobe 537.4000
1203A 30R1W 108-1128highly plagioclase-olivine-phyric basalttholeiiticpillow lava18.70570.8820.50
1203A 31R1W 46-508  pillow lava 579.86510
1203A 32R2W 32-3611plagioclase-olivine-phyric basalttholeiiticsheet lobe4.58590.6220
1203A 32R3W 58-6211  sheet lobe 592.3625355
1203A 32R4W 76-8011  sheet lobe 593.8220150
1203A 32R5W 60-6411  sheet lobe 594.7730.50
1203A 36R3W 25-2914moderately plagioclase-phyric basalttholeiiticvesicular pillow lava9.49620.55700
1203A 36R6W 18-2214  vesicular pillow lava 624.18800
1203A 37R2W 87-9116aphyric to highly-olivine-phyric basalttholeiiticsheet lobe10.37629.870.5160
1203A 37R3W 103-10716  sheet lobe 631.530.5450
1203A 38R1W 123-12616  sheet lobe 638.33140
1203A 39R5W 98-10218moderately plagioclase-olivine-phyric basalttholeiiticvesicular pillow lava14.57653.14530
1203A 42R1W 88-9219vesicular moderately olivine-phyric basalttholeiiticcompound pahoehoe19.63676.381071
1203A 42R5W 40-4419  compound pahoehoe 681.5917112
1203A 44R1W 75-7920moderately olivine-phyric basalttholeiiticvesicular pillow lava36.73695.45150
1203A 45R1W 35-3920  vesicular pillow lava 703.45150
1203A 46R4W 50-5420  vesicular pillow lava 717.48120
1203A 47R3W 50-5421moderately olivine-phyric basalttholeiiticvesicular pillow lava19.05725.48220
1203A 48R2W 96-10021  vesicular pillow lava 729.4200
1203A 49R3W 50-5421  vesicular pillow lava 735.05000
1203A 52R 6W 23-2723vesicular sparsely olivine-plagioclase-phyric to aphyric basaltalkaliccompound pahoehoe63.39768.09410
1203A 53R6W 123-12723  compound pahoehoe 779.13210
1203A 54R4W 74-7823  compound pahoehoe 785.09210
1203A 55R1W 109-11323  compound pahoehoe 790.792trace0e
1203A 58R2W 94-9823  compound pahoehoe 816.95311
1203A 58R4W 37-4124vesicular aphyric basalttholeiiticpahoehoe sheet lobe3.38819.2410.50
1203A 59R2W 69-7324  pahoehoe sheet lobe 820.550.500
1203A 62R1W 101-10526vesicular aphyric basaltalkaliccompound pahoehoe40.55848.41000
1203A 62R2W 88-9226  compound pahoehoe 849.78trace00
1203A 63R4W 19-2226  compound pahoehoe 861.65000
1203A 65R4W 9-1329vesicular aphyric basaltalkaliccompound pahoehoe8.38880.03210
1203A 66R2W 8-1030vesicular, sparsely plagioclase-phyric basaltalkaliccompound pahoehoe23.76887.38140
1203A 67R4W 10-1430  compound pahoehoe 899.90.50.50
1203A 68R4W 40-4331bhighly plagioclase-phyric basalttholeiiticpillow lava≫0.2909.86150.50
1204A 7R3W 12-162aphyric basaltalkalicvesicular pillow lava35.49822.440.500
1204A 8R1W 73-772  vesicular pillow lava 829.730.500
1204A 9R2W 55-592  vesicular pillow lava 840.65000
1204A 9R2W 76-802  vesicular pillow lava 840.860.500
1204A 9R3W 25-272  vesicular pillow lava 841.85000
1204A 10R2W 89-932  vesicular pillow lava 850.660trace0
1204A 10R2W 108-1122  vesicular pillow lava 850.850trace0
1204A 10R3W 77-812  vesicular pillow lava 851.68000
1204A 10R5W 44-482  vesicular pillow lava 854.260.50.50
1204B 1R4W 51-531aphyric basaltalkalicvesicular pillow lava27.07815.26<100
1204B 2R2W 76-801  vesicular pillow lava 822.36<100
1204B 2R4W 102-1051  vesicular pillow lava 824.82<100
1204B 2R5W 11-151  vesicular pillow lava 825.26<100
1204B 3R2W 41-441  vesicular pillow lava 831.76<100
1204B 4R2W 49-532Aaphyric basaltalkalicvesicular pillow lava30.53841.59000
1204B 6R1W 44-472A  vesicular pillow lava 859.04000
1204B 7R2W 39-432A  vesicular pillow lava 870.06000
1204B 7R3W 68-722A  vesicular pillow lava 871.85000
1204B 8R2W 85-892Bdiabasealkalicsheet lobe/internal pathway46.22880.15000
1204B 9R3W 28-322B  sheet lobe/internal pathway 888.97000
1204B 10R1W 11-142B  sheet lobe/internal pathway 888.91000
1204B 10R2W 50-542B  sheet lobe/internal pathway 890.8000
1204B 10R4W 43-472B  sheet lobe/internal pathway 892.79000
1204B 12R1W 65-692B  sheet lobe/internal pathway 906.95000
1204B 13R1W 118-1222B  sheet lobe/internal pathway 909.48000
1204B 14R1W 18-222B  sheet lobe/internal pathway 916.08000
1204B 14R1W 87-912B  sheet lobe/internal pathway 916.77000
1204B 17R1W 107-1103aphyric basaltalkalicvesicular pillow lava2.33945.97000
1204B 17R2W 11-153  vesicular pillow lava 946.19000

2.2.2. Holes 883E, 883F, 1204A, and 1204B

[11] Site 1204 is ∼27 km north of Site 1203, and within ∼0.5 km of Site 883. Both sites are located on the edge of a summit plateau (Figure 1b). Holes 1204A and 1204B were drilled within ∼100 m of each other. Five samples from Site 1204 and two from Site 883 yield no reliable 40Ar-39Ar ages [Keller et al., 1995; Duncan and Keller, 2004]. However, nannofossils in the sediment overlying and intercalated within the basalt from Site 1204 imply a minimal age of 71–76 Ma (cc22-23 [Shipboard Scientific Party, 2002b]). At Hole 1204A the basement penetration was 60.8 m, but a clogged bit limited recovery to 38.4 m of one vesicular pillow lava flow unit (Table 1; Figure 2). Hole 1204B penetrated 140.5 m into basement, and recovered 3 lava flow units; Unit 2 is ∼77 m thick (Figure 2). These lava flow units, except for subunit 2B, are multilobed vesicular pillow lavas, and unaltered glass was recovered from several glassy margins. Subunit 2B is composed of coarse grained (up to 6 mm) diabase: From the upper and lower margins toward the center of this subunit, grain size ranges from highly vesicular aphanitic to fine-grained to sparsely vesicular medium-grained with ophitic to subophitic texture. This subunit may be a submarine sheet lobe or it may represent an internal transport system that fed an active flow front (Figure 3; T. Thordarson et al., manuscript in preparation, 2004). On the basis of their high vesicularity, Site 1204 lavas, and nearby Site 883 lavas, were classified as subaerially erupted pahoehoe flows by Shipboard Scientific Party [2002b]. However, volatile analyses of glasses associated with these lavas indicate that these lavas were only partially degassed upon eruption. They probably erupted at relatively shallow water depths (<500–1000 m) (Figure 3; T. Thordarson et al., manuscript in preparation, 2004).

[12] The four lava flow units from Site 1204 are nearly aphyric, with some units containing trace amounts (<5%) of plagioclase phenocrysts and olivine and plagioclase microphenocrysts. These lava flows range from highly vesicular (up to 30% close to lobe margins) to non-vesicular (in diabase). Vesicles are variably filled with secondary minerals, such as carbonate, zeolites, sulfides and green clay. Olivine, in groundmass or as microphenocrysts, is altered to Fe-oxyhydroxide, iddingsite and carbonate. Glass in groundmass is altered to green and brown clays and Fe-oxyhydroxide.

2.2.3. Hole 884E

[13] In contrast to the relatively shallow water depth of Sites 1203, 1204 and 883 (2370–2593 m), Site 884 is located in deeper water (3284 m) on the northeastern flank of Detroit Seamount (Figure 1b). Only one of three samples from this drill site yields a meaningful 40Ar-39Ar plateau age: 81 ± 1 Ma [Keller et al., 1995]. Hole 884E was drilled 87 m into basement and recovered 13 cooling units [Shipboard Scientific Party, 1993]. From top to bottom, these lava flow units range from variably vesicular (0–30%) pahoehoe lavas to non-vesicular pillow lavas. Unaltered glass only occurs in the lobe margins of lower pillow lavas, and their volatile contents require eruption at deep water depth (>500 m) (Figure 3; T. Thordarson et al., manuscript in preparation, 2004). No volcaniclastic rock was recovered from Hole 884E.

[14] At Site 884 the upper pahoehoe lavas range from aphyric to highly plagioclase-phyric to megaphyric basalt. Lower pillow lavas range from highly plagioclase-phyric to megaphyric to highly plagioclase-olivine basalt [Shipboard Scientific Party, 1993]. The compositional zonation of cm-size plagioclase reflects magma mixing [Kinman and Neal, 2002].

3. Samples and Sample Preparation

[15] All lava flow units recovered from Detroit Seamount during ODP Leg 197 were sampled and multiple samples were collected from thick units. Table 2a shows the unit number, name and thickness, basalt type, lava flow type and phenocryst proportions for each sample. Unaltered glass was obtained from 12 units at Sites 883, 884, 1203 and 1204 (Table 2b).

Table 2b. Glass Sample Information
Sample NameUnitaDepth, mbsf
  • a

    Most units are multilobed, and are divided into subunits. Letters after numbers refer to subunits. See Tarduno et al. [2002] for a detailed documentation of subunits.

1203A 18R-2 17-211f: base of Unit 1 overlying sediment465.47
1203A 18R-2 21-271f: glass at the top of Unit 2 sediment465.51
1203A 19R-2 38-423k: glass in middle of lobe 3k475.28
1203A 20R-1 106-1073x: base of lobe, overlying 3y484.06
1203A 20R-2 58-613ab top of lobe485.08
1203A 31R-1 105-1088k: glass in middle of lobe580.45
1203A 31R-2 1-68m top of lobe580.91
1203A 39R-6 57-6318f top of lobe653.97
1203A 40R-1 0-318h top of lobe656.2
1203A 40R-1 65-7018h bottom of lobe656.85
1203A 40R-2 102-10518i base of lobe657.92
1203A 40R-2 142-14618j base of lobe658.32
1203A 40R-4 57-6518n top of lobe660.29
1203A 40R-4 80-9018o top/side of lobe660.52
1203A 45R-2 32-3520s top of lobe704.92
1203A 46R-4 0-420aj top of lobe716.98
1204B 1R-3 69-81Unit 1 glass fragments814.39
1204B 3R-2 97-100Unit 1 lobe boundary832.2
1204B 3R-2 100-106Unit 1 lobe boundary832.35
1204B 3R-2 106-110Unit 1 lobe boundary832.41
1204B 15R-1 112-116Unit 2c lobe margin926.72
883E 20R-5 46-50Unit 8 top 
883E 22R-2 105-110Unit 17 lobe margin 
883E 22R-4 112-120Unit 19-20 boundary 
884E 10R-5 59-62Unit 11 top 
884E 10R-6 38-40Unit 11 base 
884E 10R-6 118-121Unit 13 top 

[16] Highly altered regions, such as amygdules and veins, were removed from all samples selected for whole-rock analysis. The trimmed whole-rock samples were wrapped in plastic and crushed with a hammer into small pieces (<0.5 cm). Again, pieces containing filled vesicles and/or veins were removed. The remaining chips were rinsed with deionized water several times before pulverizing in an agate shatter box. Fragments of unaltered glass were hand-picked using a binocular microscope. To remove alteration phases, these fragments were leached in a 1:1 mixture of 2N HCl and H2O2 for 10 min.

4. Analytical Procedures

4.1. Major and Trace Element Analyses

4.1.1. Whole-Rock Samples

[17] Abundance of major and some trace elements (Table 3a) were analyzed by X-ray fluorescence at the University of Massachusetts at Amherst following the procedure described by Rhodes [1996]. Other trace elements (Table 3a) were analyzed by inductively coupled plasma mass spectrometry (ICP-MS) at Massachusetts Institute of Technology using the procedure described by Huang and Frey [2003]. Whole rocks from Sites 883 and 884 have been studied by Regelous et al. [2003]. For comparison of their data with our data on Sites 1203 and 1204 lavas, their data have been normalized using BHVO standard values reported by Regelous et al. [2003] and Huang and Frey [2003].

Table 3a. Whole-Rock Analysesa
Sample NameUnitUnit NameDepth, mbsfLOIbSiO2TiO2Al2O3Fe2O3*cMnOMgOCaONa2OK2OP2O5Totald        
  • a

    Major oxide contents are in %, and trace element abundances are in ppm.

  • b

    L.O.I., loss on ignition. ∼0.2 g samples were ignited at 1000°C for over 8 hours. L.O.I. was calculated using (weight loss during ignition/weight before ignition)*100%.

  • c

    Samples were oxidized prior to analyses, so all iron is reported as Fe3+, i.e., Fe2O3.

  • d

    Total is sum of oxides measured after ignition and oxidation.

1203A 17R4W 43-471highly plagioclase-olivine-phyric basalt457.992.1349.41.9916.        
1203A 18R3W 116-1193highly plagioclase-olivine-phyric basalt467.962.1848.61.9515.312.70.175.9011.72.390.640.2099.5        
1203A 20R3W 10-143highly plagioclase-olivine-phyric basalt486.11.3949.31.9515.611.30.167.0011.        
1203A 25R1W 37-415moderately plagioclase-olivine-phyric basalt531.473.1848.12.0416.        
1203A 26R2W 5-96sparsely plagioclase-phyric basalt535.844.4947.72.0016.112.40.1810.08.262.320.270.2199.4        
1203A 26R3W 97-1016sparsely plagioclase-phyric basalt537.44.1347.91.9115.712.50.219.998.452.380.360.2399.7        
1203A 30R1W 108-1128highly plagioclase-olivine-phyric basalt570.880.37049.01.9515.612.20.215.6312.        
1203A 31R1W 46-508highly plagioclase-olivine-phyric basalt579.861.5549.51.9415.910.30.246.4612.32.510.250.1999.6        
1203A 32R2W 32-3611plagioclase-olivine-phyric basalt590.62.3250.12.0115.711.30.128.598.223.050.370.2099.7        
1203A 32R3W 58-6211plagioclase-olivine-phyric basalt592.361.5644.70.949.9413.30.2021.        
1203A 32R4W 76-8011plagioclase-olivine-phyric basalt593.821.9948.41.5416.311.70.157.8911.        
1203A 32R5W 60-6411plagioclase-olivine-phyric basalt594.772.2749.31.8915.711.        
1203A 36R3W 25-2914moderately plagioclase-phyric basalt620.551.5248.91.6817.99.800.175.7412.92.370.100.1699.7        
1203A 36R6W 18-2214moderately plagioclase-phyric basalt624.181.3749.81.7718.        
1203A 37R2W 87-9116aphyric to highly-olivine-phyric basalt629.873.2347.31.2314.411.80.1613.39.681.340.100.1099.5        
1203A 37R3W 103-10716aphyric to highly-olivine-phyric basalt631.531.5143.80.727.7812.40.1828.45.410.860.080.0699.7        
1203A 38R1W 123-12616aphyric to highly-olivine-phyric basalt638.331.7848.71.4415.412.10.178.3111.31.740.160.1299.4        
1203A 39R5W 98-10218moderately plagioclase-olivine-phyric basalt653.143.0848.41.7415.911.00.246.5813.        
1203A 42R1W 88-9219vesicular moderately olivine-phyric basalt676.385.3547.91.5615.511.00.178.4911.        
1203A 42R5W 40-4419vesicular moderately olivine-phyric basalt681.593.5947.91.5415.311.00.199.8911.42.500.080.15100.0        
1203A 44R1W 75-7920moderately olivine-phyric basalt695.455.2347.32.1715.        
1203A 45R1W 35-3920moderately olivine-phyric basalt703.458.0145.62.1316.        
1203A 46R4W 50-5420moderately olivine-phyric basalt717.4810.944.32.2115.212.50.303.6516.23.261.670.2799.5        
1203A 47R3W 50-5421moderately olivine-phyric basalt725.483.2849.12.2115.911.90.209.438.512.250.310.20100.0        
1203A 48R2W 96-10021moderately olivine-phyric basalt729.42.9348.12.1715.413.30.198.339.502.470.250.2099.8        
1203A 49R3W 50-5421moderately olivine-phyric basalt735.051.8347.82.1914.913.50.208.3410.        
1203A 52R 6W 23-2723vesicular sparsely olivine-plagioclase-phyric to aphyric basalt768.093.0248.22.5916.        
1203A 53R6W 123-12723vesicular sparsely olivine-plagioclase-phyric to aphyric basalt779.132.5247.62.5815.312.70.196.9010.82.900.180.4899.6        
1203A 54R4W 74-7823vesicular sparsely olivine-plagioclase-phyric to aphyric basalt785.091.4848.12.8815.712.40.195.9811.12.940.130.5499.9        
1203A 55R1W 109-11323vesicular sparsely olivine-plagioclase-phyric to aphyric basalt790.791.9847.52.6915.512.60.206.4210.82.760.400.5199.4        
1203A 58R2W 94-9823vesicular sparsely olivine-plagioclase-phyric to aphyric basalt816.951.9647.82.6615.        
1203A 58R4W 37-4124vesicular aphyric basalt819.241.8349.01.7215.511.80.187.7511.22.630.240.17100.1        
1203A 59R2W 69-7324vesicular aphyric basalt820.551.9848.51.7615.511.90.157.5611.52.410.090.1799.5        
1203A 62R1W 101-10526vesicular aphyric basalt848.414.8545.23.3013.713.10.305.4514.42.531.000.6899.7        
1203A 62R2W 88-9226vesicular aphyric basalt849.783.9747.42.9315.        
1203A 63R4W 19-2226vesicular aphyric basalt861.653.2349.03.0515.611.        
1203A 65R4W 9-1329vesicular aphyric basalt880.033.7947.63.6015.        
1203A 66R2W 8-1030vesicular, sparsely plagioclase-phyric basalt887.384.8346.53.4414.815.40.157.566.762.741.790.6399.7        
1203A 67R4W 10-1430vesicular, sparsely plagioclase-phyric basalt899.93.6247.53.1914.613.        
1203A 68R4W 40-4331bhighly plagioclase-phyric basalt909.863.2248.81.4019.310.40.105.7210.        
1203A 17R4W 43-471 8.212427.32499.010880271265 40.69.2124430.91248.6941.07.4119.43.0814.3
1203A 18R3W 116-1193 8.112326.223811.610971247297 38.811.422929.21208.6935.27.0218.42.9514.0
1203A 20R3W 10-143 8.012226.52460.411398236305 39.80.73123229.01188.6030.16.8018.42.9013.8
1203A 25R1W 37-415 8.212924.92313.888130232257 35.13.8122728.21289.0254.97.4019.43.0714.6
1203A 26R2W 5-96 7.712324.22271.385133259244 33.81.4621426.91228.5251.96.9718.32.9513.8
1203A 26R3W 97-1016 8.112825.62152.884150238215 30.92.9820128.11278.9855.07.4720.03.0514.6
1203A 30R1W 108-1128 7.511726.22433.211189260298 40.13.5323028.91128.0325.06.6517.52.8413.5
1203A 31R1W 46-508 7.411726.52531.011097250303 43.41.2225229.51148.1537.96.6717.72.8113.5
1203A 32R2W 32-3611 7.411725.22383.410976290304 43.23.4423228.41158.3548.46.9518.02.9013.7
1203A 32R3W 58-6211 3.45512.81641.786812651140 21.11.8715013.7543.9123.
1203A 32R4W 76-8011 5.99220.72351.281120264200 33.11.3122322.5896.2629.95.2514.02.2410.6
1203A 32R5W 60-6411 6.810722.02291.610676262298 40.71.6221724.01047.4740.65.9515.72.5412.1
1203A 36R3W 25-2914 6.49722.32660.38777272263 36.90.39025724.3946.7627.15.6914.72.4311.7
1203A 36R6W 18-2214 6.710322.12710.47981306284 38.10.68925724.31007.2436.56.1015.92.5312.1
1203A 37R2W 87-9116 3.06517.51551.184376895187 31.11.2114418.9653.4317.33.118.551.477.39
1203A 37R3W 103-10716 1.94110.7901.28513221442124 20.00.9908311.7412.0911.51.955.290.9144.67
1203A 38R1W 123-12616 4.08121.11951.389164353232 34.51.5118623.5804.3425.44.0410.91.879.33
1203A 39R5W 98-10218 6.810323.22605.389112268249 34.75.1524024.81017.3326.56.1516.12.5612.2
1203A 42R1W 88-9219 5.38520.020116.685174372203 31.916.819722.5845.7327.54.8412.82.1310.1
1203A 42R5W 40-4419 5.38520.32420.684212487231 32.70.79423322.4855.6716.14.8512.52.0310.0
1203A 44R1W 75-7920 7.211928.220815.9105119247297 44.817.322233.71268.2437.36.5217.62.9114.1
1203A 45R1W 35-3920 6.711327.724427.49799245339 48.230.927235.31258.1551.96.7917.62.8613.8
1203A 46R4W 50-5420 6.711327.332420.610593277336 47.023.235634.31227.8948.46.4516.92.8013.7
1203A 47R3W 50-5421 7.312227.61911.2102114242306 46.81.2320634.11328.6733.06.6018.32.9514.5
1203A 48R2W 96-10021 7.312129.21892.3102102238296 47.42.8019835.51298.3943.56.7417.82.9014.1
1203A 49R3W 50-5421 7.112129.41970.097115273299 49.70.30220635.81288.3327.76.6417.92.9314.3
1203A 52R 6W 23-2723 13.621435.42350.8112104183299 44.00.84425043.422915.446.412.332.94.9823.6
1203A 53R6W 123-12723 13.921639.72320.012692183295 42.80.48124448.323215.647.012.734.55.2925.2
1203A 54R4W 74-7823 15.524445.32540.013785200328 49.20.34326054.825617.352.214.438.35.8828.0
1203A 55R1W 109-11323 14.622841.82373.713194187305 44.04.3624550.623916.146.413.636.15.5226.6
1203A 58R2W 94-9823 14.122240.12280.512182180298 38.11.2822946.822515.550.212.834.75.3425.4
1203A 58R4W 37-4124 5.810324.22212.884107267238 38.53.0221127.01006.3844.05.6714.92.4611.9
1203A 59R2W 69-7324 5.910724.82330.086106249239 36.90.38621927.11036.5332.65.6315.32.5112.0
1203A 62R1W 101-10526 18.528454.12198.013766129335 38.18.0121062.328319.462.317.646.16.9433.1
1203A 62R2W 88-9226 15.924641.62250.512086136310 33.70.91321646.724516.855.214.438.15.6826.8
1203A 63R4W 19-2226 16.725538.12390.810895145331 34.00.89322542.925717.861.314.939.25.8727.3
1203A 65R4W 9-1329 15.821141.223617.815075179363 39.017.823046.120316.262.714.037.25.6126.8
1203A 66R2W 8-1030 14.919938.322524.114263167384 35.123.021142.119215.057.613.434.85.1924.6
1203A 67R4W 10-1430 14.018536.92349.414092160337 36.59.3722641.818515.058.412.332.75.0924.4
1203A 68R4W 40-4331b 4.67919.921827.49372277217 31.327.721221.8764.8328.95.1913.32.1810.6
Sample NameUnitICP-MSSmEuGdTbDyHoErTmYbLuHfTaPbThU      
1203A 17R4W 43-471 4.251.555.050.8815.301.092.940.4282.650.3863.130.6090.8340.5600.769      
1203A 18R3W 116-1193 4.131.514.890.8455.051.052.850.4252.540.3613.120.6130.7080.5520.286      
1203A 20R3W 10-143 4.001.464.840.8265.001.022.780.4092.510.3743.010.5890.8140.5420.149      
1203A 25R1W 37-415 4.151.524.770.8234.901.002.760.4122.510.3593.140.5900.6740.5550.182      
1203A 26R2W 5-96 3.981.474.600.7874.710.9742.670.3902.420.3523.080.5530.6330.5270.172      
1203A 26R3W 97-1016 4.131.494.800.8194.881.002.730.4192.450.3633.150.5750.9370.5610.182      
1203A 30R1W 108-1128 4.071.504.870.8455.101.052.910.4232.580.3713.000.5510.7260.5120.183      
1203A 31R1W 46-508 3.981.474.690.8124.901.012.800.3992.520.3572.900.5210.6900.4960.266      
1203A 32R2W 32-3611 4.031.504.880.8435.081.022.860.4302.570.3743.010.5990.9910.5210.179      
1203A 32R3W 58-6211 1.880.7102.290.3992.410.5001.340.2021.250.1851.410.2730.3360.2410.0809      
1203A 32R4W 76-8011      
1203A 32R5W 60-6411 3.571.364.250.7264.410.8892.410.3672.220.3232.770.5260.6020.4670.548      
1203A 36R3W 25-2914 3.451.324.060.7094.220.8762.410.3462.120.3032.490.4950.5760.4490.164      
1203A 36R6W 18-2214 3.511.354.200.7144.260.8882.400.3662.150.3122.590.5260.6460.4690.236      
1203A 37R2W 87-9116 2.390.9373.040.5333.320.6911.880.2811.720.2501.730.2440.7020.2050.0726      
1203A 37R3W 103-10716 1.450.5801.850.3312.060.4281.200.1751.080.1601.070.2620.2950.1240.0460      
1203A 38R1W 123-12616 2.981.163.760.6604.020.8352.300.3492.060.3032.160.3350.4820.2570.0902      
1203A 39R5W 98-10218 3.571.344.190.7254.360.9082.460.3702.200.3252.560.5050.5970.4690.241      
1203A 42R1W 88-9219      
1203A 42R5W 40-4419      
1203A 44R1W 75-7920 4.221.565.270.9125.541.173.200.4742.930.4323.200.5800.6770.5250.266      
1203A 45R1W 35-3920 4.271.555.200.9025.521.173.240.4822.970.4393.120.6400.6260.5050.394      
1203A 46R4W 50-5420 4.101.535.070.8925.391.133.080.4662.880.4313.020.5660.5450.4850.233      
1203A 47R3W 50-5421 4.401.615.290.9255.601.183.260.4812.950.4373.360.5840.8250.5470.185      
1203A 48R2W 96-10021 4.341.625.320.9375.641.183.300.5032.970.4483.230.5421.210.5210.172      
1203A 49R3W 50-5421 4.381.605.380.9425.761.193.290.4932.990.4463.210.5660.5910.5180.174      
1203A 52R 6W 23-2723 6.532.237.321.227.221.484.040.6023.640.5315.120.9610.8950.8910.451      
1203A 53R6W 123-12723 6.952.368.031.347.801.604.390.6463.860.5795.200.9880.9470.9230.318      
1203A 54R4W 74-7823 7.772.638.971.498.821.834.930.7194.400.6365.701.081.020.9900.380      
1203A 55R1W 109-11323 7.282.468.371.408.321.724.680.6754.030.6035.421.030.9770.9480.409      
1203A 58R2W 94-9823 6.982.378.101.357.931.644.440.6483.890.5725.180.9880.8770.8930.353      
1203A 58R4W 37-4124 3.551.344.360.7494.560.9322.570.3872.370.3532.640.4271.230.3930.131      
1203A 59R2W 69-7324 3.641.364.450.7654.570.9602.660.3942.400.3452.690.4400.5730.3960.134      
1203A 62R1W 101-10526 8.992.9610.51.7310.32.125.760.8384.960.7396.491.271.591.151.13      
1203A 62R2W 88-9226 7.282.448.231.378.071.654.460.6584.010.5975.571.100.9581.000.467      
1203A 63R4W 19-2226 7.282.468.031.367.841.614.420.6594.040.5865.911.      
1203A 65R4W 9-1329 7.422.588.501.418.251.664.510.6373.860.5685.      
1203A 66R2W 8-1030 6.752.387.651.287.501.534.110.6093.630.5324.691.030.9400.8930.453      
1203A 67R4W 10-1430 6.752.377.651.297.441.534.170.5823.580.5044.601.010.9530.8800.628      
1203A 68R4W 40-4331b      
Sample NameUnitUnit NameDepth, mbsfLOISiO2TiO2Al2O3Fe2O3*MnOMgOCaONa2OK2OP2O5Total        
1204A 7R3W 12-162aphyric basalt822.448.1744.12.2915.        
1204A 8R1W 73-772aphyric basalt829.738.9942.92.3214.910.20.153.4222.12.810.840.3099.9        
1204A 9R2W 55-592aphyric basalt840.658.3942.02.1914.613.40.203.6420.12.640.750.2899.8        
1204A 9R2W 76-802aphyric basalt840.863.4346.82.1915.        
1204A 9R3W 25-272aphyric basalt841.856.0245.12.5314.        
1204A 10R2W 89-932aphyric basalt850.663.8547.52.3316.        
1204A 10R2W 108-1122aphyric basalt850.853.4448.32.1916.711.        
1204A 10R3W 77-812aphyric basalt851.683.1147.62.3415.813.70.168.687.712.810.450.3099.5        
1204A 10R5W 44-482aphyric basalt854.263.9147.52.4115.413.        
1204B 1R4W 51-531aphyric basalt815.264.0646.32.3317.412.90.152.7614.42.770.720.28100.0        
1204B 2R2W 76-801aphyric basalt822.366.8044.32.3315.811.60.173.4418.62.540.560.2799.7        
1204B 2R4W 102-1051aphyric basalt824.823.6245.92.5617.913.80.202.1412.        
1204B 2R5W 11-151aphyric basalt825.263.1645.72.4016.514.80.233.0112.52.931.160.3499.6        
1204B 3R2W 41-441aphyric basalt831.762.6346.02.5017.714.50.192.3011.63.541.170.71100.2        
1204B 4R2W 49-532Aaphyric basalt841.595.1846.62.2715.712.70.163.4414.12.911.220.3399.4        
1204B 6R1W 44-472Aaphyric basalt859.044.6746.62.3615.913.00.123.8013.        
1204B 7R2W 39-432Aaphyric basalt870.064.4546.52.5814.413.00.185.5713.42.870.990.3199.8        
1204B 7R3W 68-722Aaphyric basalt871.852.6747.62.2616.        
1204B 8R2W 85-892Bdiabase880.154.0547.02.3915.412.90.176.1711.        
1204B 9R3W 28-322Bdiabase888.972.8747.42.2015.513.30.178.748.982.630.390.2599.6        
1204B 10R1W 11-142Bdiabase888.912.5947.82.2915.913.00.158.359.082.740.380.2599.9        
1204B 10R2W 50-542Bdiabase890.83.2047.32.2715.313.40.167.419.773.011.090.2799.9        
1204B 10R4W 43-472Bdiabase892.792.1747.72.2415.512.70.187.739.882.800.430.3199.4        
1204B 12R1W 65-692Bdiabase906.953.1047.62.4615.        
1204B 13R1W 118-1222Bdiabase909.483.1547.52.4415.        
1204B 14R1W 18-222Bdiabase916.083.3347.52.2516.        
1204B 14R1W 87-912Bdiabase916.773.7645.62.1614.516.        
1204B 17R1W 107-1103aphyric basalt945.974.5647.42.2015.312.70.195.5612.32.821.250.2699.9        
1204B 17R2W 11-153aphyric basalt946.196.2746.12.3615.412.20.204.5915.02.791.070.31100.0        
1204A 7R3W 12-162 8.213430.724311.09540144283 35.910.522632.81298.5869.97.7820.23.2315.8
1204A 8R1W 73-772 8.513229.92308.69540153295 37.68.3121632.01238.7166.07.6520.63.2615.7
1204A 9R2W 55-592 8.013529.324817.28851138300 35.316.423632.51328.6374.57.5719.63.1415.2
1204A 9R2W 76-802 8.113529.318921.89286185264 39.420.918132.91338.6965.87.5220.03.1715.4
1204A 9R3W 25-272 8.414333.022462.09940174352 44.161.521136.81419.2788.78.1521.13.4216.6
1204A 10R2W 89-932 8.514231.02069.38775414278 36.98.9919233.91379.1966.07.8621.03.3516.1
1204A 10R2W 108-1122 8.113628.32053.68995154299 35.63.5119431.21318.5559.37.5520.03.1114.9
1204A 10R3W 77-812 10.216435.21933.510079127283 34.63.5418239.016110.870.39.4925.23.8918.7
1204A 10R5W 44-482 8.914732.519621.99475284296 41.021.518436.91489.6989.08.4322.33.5617.1
1204B 1R4W 51-531 9.114732.22849.310662166325 37.59.1726635.71419.4667.38.2621.93.4816.9
1204B 2R2W 76-801 8.614532.32877.89974160288 38.97.4427735.31389.1657.97.9221.53.3516.1
1204B 2R4W 102-1051 9.815935.429114.113970148373 40.513.727739.215510.31159.2424.43.8818.6
1204B 2R5W 11-151 9.215431.825319.012679164380 41.118.424135.71529.7992.68.5322.83.5617.3
1204B 3R2W 41-441 9.715736.828613.614272155400 42.213.126740.71519.9697.99.3623.53.7118.1
1204B 4R2W 49-532A 8.914531.023930.217264149272 37.030.223334.91429.5180.38.2021.63.4216.4
1204B 6R1W 44-472A 9.114932.323922.716755154305 38.422.422935.61439.6375.68.2921.93.4616.8
1204B 7R2W 39-432A 10.016536.320936.812563148331 42.636.619740.616110.681.19.2224.43.8318.6
1204B 7R3W 68-722A 8.314232.622256.0177111182334 40.054.220435.91379.1278.28.2220.83.3916.4
1204B 8R2W 85-892B 8.714532.520934.2166145128304 35.432.919535.91429.5276.68.4422.63.5617.3
1204B 9R3W 28-322B 8.514130.21943.58885149285 36.73.4717232.81368.9558.57.7620.73.2715.9
1204B 10R1W 11-142B 8.514330.32033.59094149292 38.53.5118633.91389.1260.37.9921.13.3516.2
1204B 10R2W 50-542B 8.514330.019125.2210140162299 37.224.317732.81409.3266.67.7921.13.3416.2
1204B 10R4W 43-472B 9.916033.62023.899105146285 36.43.6218436.515310.165.38.7523.63.7518.2
1204B 12R1W 65-692B 10.116935.819820.010781176334 38.419.318639.816010.897.89.4125.33.9318.9
1204B 13R1W 118-1222B 10.116837.021130.913479148293 36.929.119240.216311.084.89.8626.14.1119.8
1204B 14R1W 18-222B 8.614231.41973.69596151285 37.53.4618434.01389.0862.17.9321.03.3416.0
1204B 14R1W 87-912B 8.013529.818123.48077150267 37.522.517032.81338.691517.4619.83.1215.1
1204B 17R1W 107-1103 8.413929.421572.5148179139283 36.873.620933.51388.9185.27.8420.73.3215.7
1204B 17R2W 11-153 8.914731.423940.814067146274 38.242.524235.51429.5780.48.1722.13.4516.3
Sample NameUnitICP-MSSmEuGdTbDyHoErTmYbLuHfTaPbThU      
1204A 7R3W 12-162 4.651.695.440.9425.731.213.320.4972.970.4543.360.6060.9030.5810.394      
1204A 8R1W 73-772 4.531.635.350.9175.591.163.150.4752.930.4293.340.6181.240.6010.262      
1204A 9R2W 55-592 4.441.615.290.9135.551.173.220.4923.000.4393.240.5830.7470.5770.351      
1204A 9R2W 76-802 4.541.625.440.9255.561.183.190.4822.930.4333.380.5930.7870.5880.176      
1204A 9R3W 25-272 4.881.775.871.036.171.313.620.5353.240.4843.620.6240.8690.6131.06      
1204A 10R2W 89-932 4.691.715.540.9525.791.223.330.5063.100.4623.530.6370.6340.6040.218      
1204A 10R2W 108-1122 4.321.595.040.8735.321.133.100.4722.890.4183.320.5870.4800.5870.233      
1204A 10R3W 77-812 5.391.916.361.086.621.393.760.5663.520.5193.920.7320.5220.7430.222      
1204A 10R5W 44-482 5.001.815.971.046.241.323.650.5403.280.4853.740.6280.7260.6320.241      
1204B 1R4W 51-531 4.941.795.891.      
1204B 2R2W 76-801 4.701.725.570.9805.771.223.400.5063.060.4583.510.6170.7810.6201.79      
1204B 2R4W 102-1051 5.431.986.471.136.651.403.890.5853.510.5284.020.6960.9340.7010.583      
1204B 2R5W 11-151 5.071.846.      
1204B 3R2W 41-441 5.231.906.301.096.481.393.830.5843.520.5273.900.6900.9780.6810.608      
1204B 4R2W 49-532A 4.801.725.700.9775.901.233.390.5123.070.4583.540.6360.8520.6220.321      
1204B 6R1W 44-472A 4.901.785.791.      
1204B 7R2W 39-432A 5.501.946.471.126.711.413.900.5833.540.5354.030.7011.070.6890.705      
1204B 7R3W 68-722A 4.731.735.781.005.971.283.510.5373.220.4813.540.6140.4230.6120.671      
1204B 8R2W 85-892B 5.081.856.      
1204B 9R3W 28-322B 4.621.695.440.9565.741.193.260.4943.000.4523.470.5970.8340.6050.184      
1204B 10R1W 11-142B 4.681.725.560.9625.811.223.390.5003.110.4593.540.6100.6650.5990.194      
1204B 10R2W 50-542B 4.681.705.630.9615.831.223.290.5002.990.4423.600.6401.250.6240.357      
1204B 10R4W 43-472B 5.261.876.261.076.341.333.600.5453.320.4953.810.6760.9570.6990.178      
1204B 12R1W 65-692B 5.471.916.511.126.781.413.960.6033.640.5314.070.7251.300.7420.372      
1204B 13R1W 118-1222B 5.662.006.711.176.931.464.010.6033.670.5414.100.7401.330.7470.487      
1204B 14R1W 18-222B 4.721.735.570.9805.841.223.320.5033.020.4483.460.6021.140.6030.199      
1204B 14R1W 87-912B 4.451.615.290.9195.471.163.170.4812.890.4313.330.5890.6240.5790.160      
1204B 17R1W 107-1103 4.561.675.490.9625.651.193.260.4922.980.4393.410.5910.8120.6050.263      
1204B 17R2W 11-153 4.771.745.660.9795.911.243.420.5133.150.4593.590.6480.8620.6420.633      

4.1.2. Glass Samples

[18] Major element compositions of glasses were obtained at the University of Bristol (UB) using a JEOL 8600 electron microprobe with 15keV accelerating voltage, a 10nA beam current and 5–10 μm spot size. The results reported in Table 3b are averages of 4 to 10 analyses for each glass.

Table 3b. Glass Analysesa
SampleIntervalDepth, mbsfUnitbSiO2TiO2Al2O3FeOMnOMgOCr2O3CaONa2OK2OP2O5Total              
  • a

    Major oxides were analyzed by electron microprobe, and are reported in %. Trace elements were analyzed by LA-ICP-MS, and are reported in ppm.

  • b

    See footnote of Table 2b.

1203A-11203A 18R-2 17-21465.471f: base of Unit 1 overlying sediment49.52.2113.613.              
1203A-21203A 18R-2 21-27465.511f: glass at top of Unit 2 sediment49.02.1913.613.              
1203A-31203A 19R-2 38-42475.283k: glass in middle of lobe 3k49.02.2013.613.              
1203A-41203A 20R-1 106-107484.063x: base of lobe, overlying 3y48.62.1913.313.              
1203A-4r   48.52.2013.313.              
1203A-51203A 20R-2 58-61485.083ab top of lobe48.11.9014.811.              
1203A-61203A 31R-1 105-108580.458k: glass in middle of lobe48.01.9014.811.              
1203A-71203A 31R-2 1-6580.918m top of lobe49.21.9714.              
1203A-81203A 39R-6 57-63653.9718f top of lobe49.01.9314.              
1203A-91203A 40R-1 0-3656.218h top of lobe48.92.2113.513.              
1203A-101203A 40R-1 65-70656.8518h bottom of lobe48.61.9114.911.              
1203A-111203A 40R-2 102-105657.9218i base of lobe48.61.8814.911.              
1203A-121203A 40R-2 142-146658.3218j base of lobe48.71.9215.              
1203A-131203A 40R-4 57-65660.2918n top of lobe48.71.8715.              
1203A-141203A 40R-4 80-90660.5218o top/side of lobe48.71.8915.              
1203A-151203A 45R-2 32-35704.9220s top of lobe47.92.3914.612.60.296.580.0310.              
1203A-161203A 46R-4 0-4716.9820aj top of lobe47.62.3814.412.90.216.520.0610.92.990.280.2498.5              
1204B-11204B 1R-3 69-81814.39Unit 1 glass fragments47.02.4715.              
1204B-21204B 3R-2 97-100832.2Unit 1 lobe boundary47.22.6214.413.              
1204B-31204B 3R-2 100-106832.35Unit 1 lobe boundary46.82.4015.              
1204B-41204B 3R-2 106-110832.41Unit 1 lobe boundary46.22.4314.713.              
1204B-4r   46.12.4614.613.              
1204B-51204B 15R-1 112-116926.72Unit 2c lobe margin46.62.4214.913.20.186.360.0211.03.360.380.3198.7              
883-1883E 20R-5 46-50 Unit 8 top46.62.7414.              
883-2883E 22R-2 105-110 Unit 17 lobe margin46.72.7214.              
883-3883E 22R-4 112-120 Unit 19-20 boundary46.92.7414.              
884-1884E 10R-5 59-62 Unit 11 top48.21.3314.710.80.197.770.0512.              
884-2884E 10R-6 38-40 Unit 11 base48.11.3814.810.70.167.800.0712.72.930.120.1398.9              
884-3884E 10R-6 118-121 Unit 13 top48.31.3714.910.70.217.780.0412.72.950.120.1299.2              

[19] In situ laser ablation ICP-MS trace element analyses were obtained at UB using a VG Elemental PlasmaQuad 3+ S-option instrument equipped with a 266 nm Nd-YAG laser (VG Microprobe II). The laser beam diameter at the sample surface was ∼50 μm, and the laser repetition rate was set to 10Hz. This choice of laser beam conditions reflected the need to obtain good sensitivity in order to achieve low detection limits. Helium gas and then an argon/helium mixture carried the ablated material from the sample cell to the plasma torch. All measurements were made using Thermo Elemental PlasmaLab “time-resolved analysis” data acquisition software with a total acquisition time of 100 s per analysis (50 s for a gas blank and 50 s for laser ablation). The data reduction algorithm is that of Longerich et al. [1996]. NIST 610 glass was used for calibration and NIST 612 glass was used to check on instrument linearity. Analyses of the NIST glass standards bracketed the analysis of unknowns in order to correct for instrument drift. The glass standards BCR-2g, BHVO-2g, BIR-1g were analyzed as unknowns as a check of accuracy. Ca was used as internal standard to correct for variations in ablation yield between and during individual analyses of both standards and samples. Data in Table 3b represent averages of a minimum of 6 analyses of each glass fragment, and for comparison with whole-rock data, the glass data have been normalized to the values of BHVO-2 obtained from solution ICP-MS by Huang and Frey [2003].

4.2. Isotopic Analyses

[20] Sr-Nd-Pb isotopic ratios were determined in seventeen whole-rock samples (Table 4a). Because isotopic ratios and parent/daughter ratios were likely affected by postmagmatic alteration processes [Regelous et al., 2003], Sr, Nd and Pb isotopic ratios and their parent/daughter ratios were determined on acid leached whole-rock powders.

Table 4a. Sr-Nd-Pb Isotopic Ratios and Parent/Daughter Ratios
SampleUnitWeight Loss,a %equation image±2 sigmaequation image±2 sigmaequation image±2 sigmaequation image±2 sigmaequation image±2 sigmaIn Leached ResidueIn Unleached Whole Rock
equation imageequation imageequation imageequation imageequation imageequation imageequation imageequation image
  • a

    Weight loss during acid leaching.

  • b

    Pb isotopic ratios in this sample are average of two independent analyses, and U/Pb and Th/Pb ratios are average of three independent analyses: Measured 206Pb/204Pb, 207Pb/204Pb, and 208Pb/204Pb are 19.072, 15.530, 38.415, and 19.074, 15.500, 38.345, respectively. Measured U/Pb ratios in residue are 1.04, 1.30, and 1.23, and Th/Pb ratios are 0.24, 0.25, and 0.29.

  • c

    Values are reported for standard samples: La Jolla Nd standard (for Nd), NBS-987 (for Sr) and NBS-981 (for Pb).

1203A 17 R4W 43-471480.5130800.0000060.7032180.00002318.7210.00315.4730.00237.9470.0050.410.00460.330.140.300.0380.920.67
1203A 20 R3W 10-143370.5130790.0000060.7031980.00001318.3620.00415.4870.00437.9860.0100.410.0010.060.0750.290.0030.180.67
1203A 31 R1W 46-508280.5130850.0000060.7031050.00001018.3120.00215.4760.00237.9270.0040.390.00250.130.130.300.0050.390.72
1203A 32 R4W 76-8011360.5130770.0000070.7031020.00000818.2660.00415.4620.00337.8380.0080.400.00160.
1203A 38 R1W 123-12616360.5131120.0000060.7030350.00001118.1770.00215.4540.00237.7330.0040.420.00340.140.150.320.0080.190.53
1203A 49 R3W 50-5421380.5130780.0000100.7030460.00000818.2990.00215.4690.00237.9030.0050.400.00050.0880.100.310.00100.290.88
1203A 54 R4W 74-7823400.5130470.0000060.7031620.00000818.5900.00215.4840.00238.1300.0050.360.00070.
1203A 59 R2W 69-7324330.5130930.0000070.7031130.00000818.2060.00215.4670.00237.7840.0040.430.00050.080.110.300.00200.230.69
1203A 63 R4W 19-2226540.5130650.0000090.7030850.00000818.6290.00815.4680.00638.0230.0140.330.0040.460.100.270.00400.611.02
1203A 65 R4W 9-13b29600.5130450.0000060.7031880.00001019.0730.00315.5150.00238.3800.0050.320.0311.
1203A 66 R2W 8-1030800.5130470.0000080.7032020.00001518.7750.00215.4900.00238.1400.0050.330.0320.720.080.270.1090.480.95
1203A 68 R4W 40-4331500.5130790.0000050.7031460.00000819.1110.00315.5340.00238.1960.0060.332E-050.910.790.300.1310.460.48
1204A 10 R2W 108-1122440.5130960.0000060.7029150.00000718.2610.00415.4570.00437.8230.0080.350.0130.280.560.290.0180.491.22
1204B 3 R2W 41-441400.5131010.0000050.7028530.00000818.2310.00315.4480.00337.8470.0070.370.0160.
1204B 7 R3W 68-722A400.5130900.0000070.7029990.00001018.3280.00315.4760.00237.8710.0050.380.0540.340.690.290.271.591.45
1204B 10 R4W 43-472B390.5130780.0000060.7029100.00000818.1620.00115.4570.00137.7930.0030.340.0090.
1204B 17 R1W 107-1103520.5130930.0000070.7029400.00001318.3180.00315.4470.00237.9280.0060.350.0250.
LE unleached  0.5130360.0000060.7032050.00000818.3080.00215.4690.00237.9050.0050.290.00590.220.65    
LE A    0.7031140.00000718.2850.00815.4720.00637.8810.0170.400.00270.150.30    
LE B  0.5130770.0000070.7031170.000018      0.400.00220.160.29    
LE C    0.7031240.00001018.2710.00315.4540.00337.8190.0070.390.00180.150.26    
LE D    0.7031180.00001818.2710.00315.4570.00237.8270.006        
LE E    0.7031200.000007              
LE F    0.7031240.000007      0.400.00160.160.27    
LE G    0.7031090.000007              
LE H    0.7031150.000007              
LE I  0.5130770.0000070.7031130.00000818.2660.00415.4620.00337.8380.0080.400.00160.160.27    
Standard sample valuesc                    
MIT  0.511856 0.710265 16.937 15.493 36.713         
From Keller et al.  0.511876 0.710248 16.937 15.493 36.705         
From Regelous et al.  0.511872 0.710223 16.9403 15.4974 36.7246         

[21] Whole-rock powders were strongly acid-leached before being digested in HF-HNO3 for Sr-Nd-Pb isotopic analyses. Specifically, ∼0.3 to 0.4 g whole-rock powder was weighed into a precleaned Teflon® beaker, and ∼10 mL 6N HCl was added. This beaker was placed in an ultra-sonic bath and the HCl was changed every 30 min until the acid was colorless or pale yellow. For most samples, 6 to 9 leaching steps were required. The residue was soaked in deionized H2O for 30 min, rinsed twice, and dried. The weight loss after acid leaching was between 28 to 80% (Table 4a). Electron microprobe analyses of several residual assemblages showed that only plagioclase, pyroxene and olivine were present.

[22] The leached sample powders were digested in HF-HNO3. Three separate dissolutions were prepared for: 1) Pb isotope analyses by Thermal Ionization Mass Spectrometry (TIMS); 2) Sr-Nd isotope analyses by TIMS; and 3) parent/daughter abundance ratio determination by ICP-MS. Pb was separated from the sample matrix by anion exchange technique in HBr acid. Sr and Nd were separated by cation exchange technique in HNO3 and HCl, respectively. Detailed TIMS analytical procedures are described by Schmitz and Bowring [2003]. Parent/daughter abundance ratios in the leached powder (Table 4a) were analyzed by ICP-MS using the procedure described in Huang and Frey [2003]. Initial Sr-Nd-Pb isotopic ratios in Detroit Seamount lavas are reported in Table 4b.

Table 4b. Age-Corrected (to 76 Ma) Sr-Nd-Pb Isotopic Ratiosa
SampleUnit143Nd/144Nd±2 sigma87Sr/86Sr±2 sigma206Pb/204Pb±2 sigma207Pb/204Pb±2 sigma208Pb/204Pb±2 sigma
  • a

    The 2 sigma uncertainty in parent/daughter ratios is taken to be 5%. The 2 sigma uncertainty in age-corrected isotopic ratios is estimated considering both uncertainty in TIMS analyses and uncertainty in parent/daughter ratio analyses.

1203A 17 R4W 43-4710.5129530.0000090.7032040.00002318.4770.01315.4610.00237.9130.005
1203A 20 R3W 10-1430.5129520.0000090.7031950.00001318.3200.00515.4850.00437.9680.010
1203A 31 R1W 46-5080.5129640.0000090.7030970.0000118.2160.00515.4710.00237.8960.004
1203A 32 R4W 76-80110.5129500.0000150.7030970.00000818.1520.00715.4570.00337.7740.009
1203A 38 R1W 123-126160.5129820.0000090.7030250.00001118.0750.00515.4490.00237.6970.004
1203A 49 R3W 50-54210.5129550.0000120.7030440.00000818.2330.00415.4660.00237.8790.005
1203A 54 R4W 74-78230.5129360.0000080.703160.00000818.4120.00915.4760.00238.0960.005
1203A 59 R2W 69-73240.5129600.0000100.7031110.00000818.1470.00415.4640.00237.7580.004
1203A 63 R4W 19-22260.5129630.0000110.7030730.00000818.2890.01915.4520.00637.9990.014
1203A 65 R4W 9-13290.5129460.0000080.7030950.00001118.1830.04515.4730.00338.3160.006
1203A 66 R2W 8-10300.5129450.0000100.7031060.00001618.2410.02715.4650.00238.1210.005
1203A 68 R4W 40-43310.5129760.0000070.7031460.00000818.4320.03415.5020.00338.0020.011
1204A 10 R2W 108-11220.5129880.0000080.7028760.00000718.0560.01115.4470.00437.6880.010
1204B 3 R2W 41-4410.5129860.0000080.7028050.00000818.1430.00515.4440.00337.7840.008
1204B 7 R3W 68-722A0.5129720.0000090.7028370.00001318.0780.01315.4640.00237.7050.010
1204B 10 R4W 43-472B0.5129720.0000080.7028830.00000818.0670.00515.4520.00137.7590.003
1204B 17 R1W 107-11030.5129840.0000090.7028650.00001418.2440.00515.4440.00237.8680.007
LE Unleached 0.5129450.0000110.7031870.00000918.1450.00815.4610.00237.7470.009
LE I 0.5129500.0000150.7030970.00000818.1520.00715.4570.00337.7740.009

4.3. Effects of Acid Leaching on Isotopic Ratios

[23] In order to assess the effects of alteration on isotopic ratios, we carried out a nine step-leaching experiment on one sample, 1203A 32R4W 76-80 (Table 4a). We prepared a series of precleaned centrifuge tubes labeled as LE-A through LE-I, and ∼0.4 g sample 1203A 32R4W 76-80 was weighed into each tube; 10 mL 6N HCl was added to each tube, and the tubes were shaken. Then the tubes were placed in an ultra-sonic bath. After 30 min, the acid was decanted from each tube. For Tube LE-A, the leaching step was followed by two steps of rinsing with 10 mL deionized H2O. Tube LE-B was subjected to two leaching steps, Tube LE-C to three leaching steps, etc. for Tubes LE-D to LE-I with Tube LE-I experiencing nine leaching steps. The color of HCl was colorless after six leaching steps. Sr-Nd-Pb isotopic compositions and Rb/Sr, Sm/Nd, Th/U and U/Pb ratios in this series of unleached and leached powders are reported in Table 4a. As indicated in Figure 4, these isotopic ratios and parent/daughter ratios are relatively constant after two (or three for Rb/Sr and Th/Pb) acid leaching steps. The higher Sm/Nd but lower Rb/Sr, Th/Pb and U/Pb ratios in the leached powders reflect removal of fine-grained groundmass and alteration phases during acid leaching (Figure 4). We assume that the parent/daughter ratios of the leached powders are those of the unaltered mineral assemblage and use them for age correction. After age-correction to 76 Ma, the unleached sample and sample LE I have similar (within 2 σ uncertainty) initial 143Nd/144Nd, 206Pb/204Pb and 207Pb/204Pb ratios, but significantly different 87Sr/86Sr and 208Pb/204Pb ratios (Table 4b; Figure 4).

Figure 4.

Sr-Nd-Pb isotopic ratios and parent/daughter ratios in samples from different leaching steps. See section 4.3 text for details. The important result is that these isotopic and parent/daughter ratios are relatively constant after three acid leaching steps. The isotope data reported in Tables 4a and 4b were obtained on rock powders that were subjected to 6–9 leaching steps.

[24] The similar initial 143Nd/144Nd ratios in leached and unleached powders is consistent with the immobility of REE during postmagmatic alteration. The different initial 87Sr/86Sr and 208Pb/204Pb ratios in leached and unleached powders clearly reflect the disturbance of the Rb-Sr and Th-Pb systems during postmagmatic alteration (Table 4b). Given the evidence for Pb mobility [e.g., Bach et al., 2003; Regelous et al., 2003], the difference in initial 208Pb/204Pb probably reflects loss of Pb (and consequently high Th/Pb) during postmagmatic alteration. However, it is surprising that leached and unleached powders have similar initial 206Pb/204Pb and 207Pb/204Pb ratios (Table 4b; Figure 4c). Since the Th/Pb ratio in leached and unleached powders changed by more than a factor of 2, but U/Pb ratio changed by only a factor of ∼1.5 (Figures 4c and 4d), it is possible that both Pb and U were mobile during alteration, but that the U/Pb ratio was not changed dramatically.

5. Results

5.1. Classification of Basalt Type

[25] The whole-rock samples are altered to variable extent as indicated by their Loss on Ignition (L.O.I.) (Table 3a), and the presence of secondary minerals. Among the 69 analyzed whole-rock samples, 54 samples have L.O.I. > 2% (hereafter described as strongly altered samples), and 15 samples have 0.4% < L.O.I. < 2% (hereafter described as moderately altered samples). In contrast, the sums of major elements in the 36 glass samples are greater than 97.9%, and 24 samples have sums greater than 99% (Table 3b, T. Thordarson et al., manuscript in preparation, 2004). This confirms the visual inspections that these glasses are unaltered. Rock classification using the total alkalis–silica criteria can be unreliable for altered rocks, because of the mobility of Na and K during postemplacement alteration (Figure 5a). However, the composition of unaltered glasses enables unambiguous alkali-silica classification for 7 of the 22 flow units recovered at Detroit Seamount during ODP Leg 197 and 5 cooling units recovered during ODP Leg 145 (Figure 3b). Out of 29 whole-rock samples from Site 1204 and 6 from the nearby Site 883 (data from Regelous et al. [2003]), 30 samples are within the alkalic field (Figure 5a). The remaining 5 samples are in the tholeiitic field close to the Macdonald-Katsura line. They include 3 samples from Hole 1204B (2 from Unit 2 and 1 from Unit 1) and 2 from Unit 2 at Hole 1204A. However, glasses from Units 1 and 2 at Hole 1204B are within the alkalic field (Table 3b, Figure 5b). Therefore we infer that the three whole-rock samples from Hole 1204B in the tholeiitic field lost alkalis during post magmatic alteration. Although no glass is available from Unit 2 at Hole 1204A, 7 of the 9 Unit 2 whole-rock samples are within the alkalic field. Therefore we classify all lavas from Sites 1204 and 883 as alkalic basalt.

Figure 5.

Na2O + K2O versus SiO2 (all in wt%). The dashed line shows the alkalic-tholeiitic boundary of Macdonald and Katsura [1964], using the equation Total Alkalis = 0.37SiO2 − 14.43 [Carmichael et al., 1974]. The Macdonald-Katsura line is based on data for total iron measured as FeO and Fe2O3; consequently, for our analyses, total iron is assumed to be 10% Fe3+ and 90% Fe2+. (a) Whole-rock data. All the Detroit Seamount lavas with low L.O.I. (<2%) are within the tholeiitic field, and those with high L.O.I. (>2%) range from tholeiitic to alkalic basalt. See text for details. Shown for comparison is the field (shaded) for late shield to postshield stage lavas from Mauna Kea [Frey et al., 1990, 1991]. Whole-rock fields for Sites 883 and 884 are from Regelous et al. [2003]. (b) Glass data. Compositions of unaltered glasses (Table 3b) clearly show that all lavas from Sites 883 and 1204 are alkalic basalt, whereas lavas from Site 884 are tholeiitic. Glasses are available in several units (Units 1, 3, 8, 18, and 20) at Site 1203, and they are also tholeiitic. Detroit Seamount glass data are from this study and T. Thordarson et al. (manuscript in preparation, 2004). (c) Whole-rock–glass pairs from submarine lava flow units at Site 1203. Except for the three samples from Unit 20, which are highly altered (L.O.I. ranges from 5–11%), the other whole-rock samples (three highly altered and three moderately altered) are within the tholeiitic field. As discussed in the text, we classify Unit 20 as a tholeiitic unit (see Figure 5b) and conclude that the high total alkali contents in Unit 20 whole rocks are a result of alteration. (d) The subaerial lava flows Units 23, 26, 29, and 30 at Site 1203 are highly altered (L.O.I. > 2%) and are within the alkalic field or straddle the alkalic-tholeiitic boundary line. No unaltered glass is available from these units; however, on the basis of their high abundances of incompatible elements, they are classified as alkalic basalt (see text).

[26] In contrast, glasses from Sites 884 and 1203, whole rocks from Site 884 and moderately altered (L.O.I. < 2%) whole rocks from Site 1203 are tholeiitic basalt. Highly altered (L.O.I. > 2%) whole rocks from Site 1203 range from the tholeiitic to the alkalic field. Three highly altered whole-rock samples from Site 1203 Unit 20 (L.O.I. ranges from 5–11%) are within alkalic field (Figure 5c); however the two glass samples from this unit are within the tholeiitic field (Figure 5c). In addition, whole-rock and glass samples from this unit have incompatible element abundances similar to other tholeiitic basalts (Tables 3a and 3b). We classify Unit 20 as tholeiitic basalt.

[27] Whole-rock samples from the thick compound pahoehoe Units 23, 26, 29 and 30 at Site 1203 (Figure 2) lie within the alkalic field or are close to alkalic-tholeiitic boundary line (Figure 5d). No unaltered glass is available from these units, but whole-rock samples from these units have high abundances of incompatible elements, such as Ti and Zr (Table 3a); hence we conclude that they represent alkalic basalt underlying the main tholeiitic basalt succession in Hole 1203A (Figure 2).

5.2. Compositional Effects of Alteration

[28] Our objectives are to understand the magmatic evolution of lavas forming Detroit Seamount. Because these lavas are ∼76–81 Myr old, the igneous geochemical characteristics of the whole rocks have been affected by postmagmatic processes. For submarine erupted lavas, these are submarine alteration processes. For now submerged but subaerially erupted lavas, both subaerial and submarine alteration processes have occurred. In this section, we discuss the effect of alteration on whole-rock chemical compositions.

[29] A direct approach to evaluate compositional changes of whole rocks caused by postmagmatic alteration is comparison of unaltered glass and whole-rock compositions. Our discussion begins with elements that are known to be mobile, e.g., K and Rb, but also Ba and P, and concludes with assessment of changes in ratios of highly incompatible elements.

[30] K (as K2O) and Rb are incompatible elements, and their abundance in unaltered tholeiitic and alkalic basalt glasses are positively correlated with Th abundance (Figures 6a and 6b). However, whole-rock samples do not define linear trends in Th-K2O and Th-Rb diagrams, which is best explained as the result of K and Rb mobility during alteration. Most Site 1204 whole-rock samples are from oxidizing zones defined by Shipboard Scientific Party [2002b], and they have higher K2O and Rb content than the unaltered glass samples at a given Th abundance (Figures 6a and 6b). In contrast, tholeiitic whole-rock samples from Site 1203 range to higher and lower K2O and Rb contents than unaltered glass samples at a given Th abundance (Figures 6a and 6b). For example, three whole-rock samples from Unit 20 at Site 1203 deviate from the trend formed by unaltered glasses to higher K2O and Rb content. Their L.O.I. varies from 5% to 11%. Hence we conclude that K2O and Rb were added during alteration. Two samples from Unit 19 at Site 1203 exhibit contrasting alteration trends. Sample 1203A 42R1W 88-92 deviates from the glass trends to higher K2O and Rb content, and sample 1203A 42R5W 40-44 trends to lower K2O and Rb content (Figures 6a and 6b). Their L.O.I. are 5% and 4%, respectively (Table 3a). We infer that K2O and Rb were added to sample 1203A 42R1W 88-92, and lost from sample 1203A 42R5W 40-44 during alteration.

Figure 6.

Th (ppm) versus (a) K2O (wt.%), (b) Rb (ppm), (c) P2O5 (wt.%) and (d) Ba (ppm) in glasses and whole rocks.

[31] Among the four alkalic, incompatible element-enriched units at the bottom of Hole 1203A, Units 29 and 30 have higher contents of K2O and Rb than Units 23 and 26 (Figures 6a and 6b). K2O/P2O5 is sensitive to alteration, and unaltered Hawaiian lavas have K2O/P2O5 > 1 [e.g., Huang and Frey, 2003]. Therefore the low and variable K2O/P2O5 in Units 23 and 26 (8 of 9 samples range from 0.24 to 0.92) compared to that of Units 29 and 30 (2.3 to 2.9) indicate that K2O and Rb were removed from Units 23 and 26 during postmagmatic alteration. This inference is important in subsequent evaluation of the role of phlogopite during the petrogenesis of these alkalic lavas.

[32] P is an incompatible element, and its content (as P2O5) in unaltered glasses is positively correlated with Th abundance (Figure 6c). Most whole-rock samples form a positive trend in Th versus P2O5 diagram which overlaps with the trend formed by the unaltered glasses. Six whole-rock samples (labeled in Figure 6c) deviate from the trend to higher P2O5 content. Their P2O5 contents, 0.5 to 0.7%, could be explained by secondary apatite, up to 1 wt%.

[33] Mobility of Ba has been observed in altered oceanic crust [e.g., Staudigel et al., 1995]. Most whole rocks from Detroit Seamount form a positive trend in a Th-Ba plot, overlapping with the trend formed by unaltered glasses (Figure 6d). However, several whole-rock samples deviate from the trend to high Ba abundance (Figure 6d). The high Ba abundance in Sample 1204B 14R1W 87-91 has been confirmed by both ICP-MS and XRF analyses (Table 3a). This sample is highly altered and contains large irregular domains of clay after glass.

[34] Ce/Pb and Ba/Th ratios are not readily changed by most magmatic processes; consequently, Ce/Pb and Ba/Th ratios are relatively uniform in unaltered oceanic basalt [e.g., Hofmann and White, 1983; Hofmann, 1986]. However, due to the mobility of Pb and Ba during alteration processes, these ratios may vary significantly in altered whole rocks. Comparison of unaltered glasses and whole rocks shows that each of these ratios are quite variable in the whole-rock samples, even among samples from the same unit, but relatively uniform in the unaltered glasses (Figure 7). For example, the average value of Ce/Pb in unaltered glasses from Detroit Seamount is 31 ± 6, which is close to the average value of oceanic basalt (25 ± 5) [Hofmann, 1986] and that of Mauna Kea lavas (29 ± 8) [Huang and Frey, 2003]. The highly variable Ce/Pb ratio (6–66) in the whole-rock samples (Figure 7a) most likely results from the mobility of Pb during alteration. The Ba/Th ratio in unaltered glasses from Detroit Seamount is 98 ± 17, but the Ba/Th ratio ranges widely (50–257) in the whole-rock samples, reflecting Ba mobility during alteration (Figure 7b). Primitive mantle values of 74 and 82 are reported by Hofmann [1988] and Sun and McDonough [1989], respectively.

Figure 7.

Highly incompatible element ratios: Ce/Pb and Ba/Th. Blue diamond: whole-rock data. Pink square: glass data. Site 883 and 884 data are normalized using the BHVO standard values reported by Regelous et al. [2003] and Huang and Frey [2003]. The pink fields show the glass range. Data for primitive mantle [Hofmann, 1988; Sun and McDonough, 1989], average oceanic basalt [Hofmann, 1986], Garrett transform fault lavas [Wendt et al., 1999], and Mauna Kea shield lavas [Huang and Frey, 2003] are shown for comparison.

5.3. Major Elements: Comparison With Hawaiian Lavas and MORB

[35] We use MgO variation plots to compare and contrast the composition of the basalt from the four ODP drill sites at Detroit Seamount as well as to compare with mid-ocean basalt (MORB) from the East Pacific Rise (EPR) and basalt from Mauna Kea Volcano and Loihi Seamount, which have well studied shield to postshield and preshield to shield transitions, respectively. Our principal objective is to use the well known differences in major element composition between MORB and Hawaiian magmas [e.g., Albarede, 1992] to determine the affinities of the magmas erupted at Detroit Seamount.

5.3.1. SiO2

[36] For whole rocks and glasses with 5–10% MgO, SiO2 contents increase from Mauna Kea and Loihi alkalic basalt (postshield and preshield stages, respectively) to Mauna Kea (low SiO2 group) and Loihi tholeiitic basalt (shield stages). The highest SiO2 contents (>50%) are in EPR MORB and the high SiO2 group lavas of Mauna Kea Shield lavas (Figure 8a). In contrast to the negative MgO-SiO2 trend defined by Hawaiian tholeiitic shield lavas, a trend reflecting olivine addition, subtraction and fractionation [e.g., Yang et al., 1996; Rhodes and Vollinger, 2004], most Detroit Seamount lavas define a broad positive trend (Figure 8a). Only the olivine-rich samples from Units 11 and 16 at Site 1203 fall on an olivine-control trend. Alkalic basalt, whole rocks and glasses, from Sites 883 and 1204 has relatively low SiO2 content that overlaps the range of alkalic postshield lavas from Mauna Kea and preshield stage lavas from Loihi. Tholeiitic glasses from Sites 884 and 1203 overlap with the low-SiO2 tholeiitic shield lavas from Mauna Kea volcano (Figure 8a).

Figure 8.

MgO content versus other oxide contents (all in wt.%). For each element, there are two panels: whole-rock data and glass data. Only samples with L.O.I. less than 7% are plotted. Whole-rock data are plotted after converting Fe2O3 to FeO. Data sources are as follows: Detroit Seamount whole rocks (Sites 1203 and 1203: this study; Sites 883 and 884: Regelous et al. [2003]); Detroit Seamount glasses (this study and Thordarson et al., manuscript in preparation, 2004); EPR N-MORB whole rock (downloaded data from PET DB); EPR N-MORB glass [Niu and Batiza, 1997; Niu et al., 1999; Regelous et al., 1999]; Loihi glass [Garcia et al., 1993, 1995, 1998]; Mauna Kea whole rock [Rhodes, 1996; Rhodes and Vollinger, 2004], and Mauna Kea glass [Stolper et al., 2004]. In the SiO2 panels the Mauna Kea fields are divided into Low and High SiO2 groups; these groups are combined in other panels.

Figure 8.


Figure 8.


Figure 8.


5.3.2. Total Iron as FeO*

[37] For whole rocks with 7–10% MgO, the trend of increasing FeO* contents is opposite to that for SiO2; i.e., EPR N-MORB have the lowest FeO* abundances and Hawaiian tholeiitic basalt generally has lower FeO* abundance than Hawaiian alkalic basalt. For Detroit Seamount lavas, the whole-rock data for FeO* are scattered widely. The FeO* content of tholeiitic basalt from Sites 1203 and 884 varies from ∼7% to 13%, which is similar to EPR N-MORB. The olivine-rich units 11 and 16 at Site 1203 overlap with Mauna Kea shield stage tholeiitic basalt (Figure 8b). The alkalic lavas from Sites 883, 1203 and 1204 overlap with the Mauna Kea tholeiitic field but a few samples range to high FeO* (>13%) and are within the field of Mauna Kea alkalic basalt. A negative MgO-FeO* trend is well defined by Detroit Seamount glasses with the highest FeO* in alkalic basalt from Sites 883 and 1204. This trend parallels the trends for EPR N-MORB and Mauna Kea glasses (Figure 8b).

5.3.3. Al2O3

[38] In the context of the MORB-related versus Hawaiian plume-related alternatives proposed for Detroit Seamount lavas [Keller et al., 2000; Regelous et al., 2003], MgO versus Al2O3 is especially important because the parental magmas for N-MORB and Hawaiian shield lavas differ significantly in Al2O3. For MgO >8%, the Al2O3 content of MORB exceeds that for Hawaiian shield lavas (e.g., Figure 8c) [see also Albarede, 1992, Figure 2]. Unfortunately, this distinction diminishes as MgO content decreases, because at ∼7% MgO the positive Al2O3-MgO trend of MORB caused by plagioclase fractionation intersects the negative trend of Hawaiian shield lavas caused by olivine fractionation (Figure 8c). The ∼5.8 to 8.0% MgO glasses from Detroit Seamount are in the region of overlap between the Mauna Kea and EPR N-MORB fields. However, the positive MgO-Al2O3 trend for Detroit Seamount glasses is similar to the plagioclase fractionation trend of N-MORB. In contrast, the three high MgO (olivine-rich) samples from Units 11 and 16 at Site 1203 define a negative trend that is intermediate between the Mauna Kea shield and EPR MORB fields.

[39] The five lavas with atypically high Al2O3 at ∼5.7 to 7.2% MgO (Site 1203, Units 14 and 31 and samples 145-10 and 145-11 from Site 884, Figure 8c) are rich (>7 vol.%) in plagioclase phenocrysts (Table 2a) [Regelous et al., 2003; Shipboard Scientific Party, 2002a].

5.3.4. TiO2

[40] It is well known that the primary magmas for Hawaiian tholeiitic lavas contain higher abundances of TiO2 (and incompatible trace elements) than primary magmas for N-MORB [e.g., Frey and Roden, 1987; Albarede, 1992]. For Detroit Seamount lavas, there is a range in TiO2 at a given MgO content: TiO2 content increases in the order Site 884 < Site 1203 (tholeiitic) < Site 1204 ∼ Site 883 < Site 1203 (alkalic) (Figure 8d). Tholeiitic lavas from Sites 884 and 1203 overlap with the N-MORB field. Note that the alkalic lavas from Site 1203 have lower TiO2 content than alkalic lavas from Mauna Kea (Figure 8d).

5.3.5. Summary

[41] Alkalic basalt with relatively low SiO2 content and high contents of FeO* and TiO2 at Detroit Seamount is consistent with the tholeiitic to alkalic basalt transition that is characteristic of the late shield and postshield stages of Hawaiian volcanoes [e.g., Frey et al., 1990, 1991]. On the other hand, Site 884 lavas with low TiO2 and P2O5 contents are similar to MORB. Although MORB and Hawaiian shield stage lavas with >8% MgO content differ significantly in Al2O3 [Albarede, 1992], evolved MORB and Hawaiian lavas with ∼6–8% MgO have similar Al2O3 content. Therefore we cannot use Al2O3 content to establish a MORB or Hawaiian affinity for the glasses from Detroit Seamount which have ∼6–8% MgO. However, as in MORB, plagioclase fractionation was an important process controlling the compositions of Detroit Seamount lavas.

5.4. Incompatible Elements

[42] In both whole rocks and glasses, abundances of incompatible elements that are immobile during alteration, such as Nb, Zr and TiO2, are positively correlated with Th abundance (Figures 9a–9c). Generally, the lowest abundance of these incompatible elements are in Site 884 lavas and the three high MgO lavas from Site 1203 (from Units 11 and 16) (Tables 3a and 3b). Alkalic lavas from Sites 1204 and 883 have incompatible element abundance intermediate between tholeiitic lavas from Sites 1203 and 884 and alkalic lavas from Site 1203. The highest abundances are in the 4 thick compound pahoehoe units, Units 23, 26, 29 and 30, in the lower part of the Site 1203 core (Figure 2). The combination of high incompatible element abundances and their location within the alkalic field or straddling the alkalic-tholeiitic boundary line in Figure 5d shows that these lavas are alkalic basalt. These 4 alkalic basalt units define two separate trends in the Zr and TiO2 versus Th plots (e.g., Figures 9b and 9c). The Ti/Zr ratio in most lavas from Detroit Seamount ranges from 86–109, which is close to the primitive mantle value of 116 [Sun and McDonough, 1989]; however, alkalic Units 23 and 26 from Site 1203 have a lower Ti/Zr ratio (67–72). Also Zr/Hf in Units 23 and 26 (44–45) differ from those in Units 29 and 30 (40–41) (Table 5).

Figure 9.

Th abundance (in ppm) versus selected incompatible element abundances (in ppm or %). For each element, there are two panels: whole-rock data and glass data. Only samples with L.O.I. less than 7% are plotted. Whole-rock trace element abundances have been adjusted using Plotted Value = Measured Value/(1-L.O.I.). Whole-rock fields are shown in glass panels for comparison. Site 883 and 884 whole-rock data are from Regelous et al. [2003].

Table 5. Ti/Zr, Ba/Th, and Zr/Hf in Alkalic Lavas at Site 1203
231203A 52R 6W 23-2768.751.544.4
231203A 53R6W 123-12766.751.245.0
231203A 54R4W 74-7867.252.245.4
231203A 55R1W 109-11367.548.144.4
231203A 58R2W 94-9870.955.243.5
261203A 62R1W 101-10569.953.943.6
261203A 62R2W 88-9272.054.943.7
261203A 63R4W 19-2271.257.643.5
291203A 65R4W 9-13105.363.540.2
301203A 66R2W 8-10107.464.540.9
301203A 67R4W 10-14103.467.540.5

[43] Abundance of other moderately incompatible elements, such as Na (as Na2O) and Lu, versus Th abundance define more complex trends (Figures 9d and 9e). Except for the three high MgO lavas (from Site 1203 Units 11 and 16), which have the lowest Th, Lu and Na2O abundance, tholeiitic lavas (glasses and whole rocks) from Sites 1203 and 884 have similar Lu and Na2O abundance. This result is surprising because Site 884 lavas have the lowest abundance of highly incompatible elements. Also, whole-rock data for Site 884 tholeiitic lavas and Site 1203 alkalic lavas overlap in Na2O content. However, alkalic glasses from Sites 883 and 1204 have higher Na2O content than tholeiitic glasses from Sites 884 and 1203.

[44] Trends defined by Ba and Sr versus Th abundance are also complex (Figures 6d and 9f). Site 1203 alkalic lavas have variable Th abundances but relatively uniform Sr and Ba abundances. The Sr data for glasses show that Sr abundances are similar in alkalic glasses from Sites 883 and 1204 and tholeiitic glasses from Site 1203 (Figure 9f). This is a surprising result given the relative enrichment of other incompatible elements in these alkalic glasses.

[45] In a primitive mantle normalized trace element diagram, the relative enrichment in highly incompatible elements decreases in the order Mauna Kea alkalic lavas > Site 1203 alkalic lavas > Sites 1204 and 883 alkalic lavas (Figure 10a). Also relative to alkalic basalt from Mauna Kea, Detroit Seamount alkalic basalt has more pronounced depletion in Sr and flatter patterns. For tholeiitic lavas, incompatible element abundance decrease in the order Mauna Kea tholeiitic lavas > Site 1203 tholeiitic lavas > Site 884 tholeiitic lavas (Figure 10b). As pointed out by Regelous et al. [2003], the Site 884 tholeiitic lavas, both whole rocks and glasses, form convex upward trends which are similar to average MORB (Figure 10b). However, the relative depletion in Sr that is characteristic of average MORB does not occur in Mauna Kea or Detroit Seamount lavas.

Figure 10.

Primitive mantle normalized incompatible element abundances for lavas from Detroit Seamount: (a) alkalic lavas and (b) tholeiitic lavas. In order to minimize the effect of crystal fractionation, only lavas with 6% < MgO < 9% are plotted, except for Site 883 glass samples, whose MgO contents are ∼5.8%. Both glass and whole-rock data are plotted. In each group, glass data overlap with whole-rock data. Mauna Kea data are from Huang and Frey [2003]. Primitive mantle values are from Hofmann [1988]. Average N-MORB data are from Hofmann [1988] and Sun and McDonough [1989].

5.5. First Series Transition Metals: Ni and Sc

[46] Abundance of Ni is positively correlated with MgO content in the glasses; tholeiitic glasses from Sites 884 and 1203 typically have higher Ni abundance than alkalic glasses from Sites 883 and 1204 (Figure 11a). The highest Ni abundances are in the three olivine-rich whole-rock samples from Units 11 and 16 at Site 1203.

Figure 11.

MgO content (in wt%) versus Ni and Sc abundances (in ppm). For each element, there are two panels: whole-rock data and glass data. Only samples with L.O.I. less than 7% are plotted. Whole-rock fields are shown in glass panels for comparison. As shown, three samples contain large proportion of olivine phenocryst (>15 vol.%), and the positive trend in Figure 11a is a result of olivine addition. See the Figures 8a8d caption for other data sources. Koolau data are from Frey et al. [1994].

Figure 11.


[47] An inverse MgO-Sc trend reflecting olivine control is typical of Hawaiian shield lavas [e.g., Frey et al., 1994; Huang and Frey, 2003]. The three MgO-rich samples from Units 11 and 16 at Site 1203 show this trend (Figure 11b). Most importantly, at a given MgO content, the Sc abundance in EPR N-MORB exceeds that of Hawaiian lavas. The majority of Detroit Seamount samples (whole rocks and glasses) have >35 ppm Sc; hence they are within the MORB field (Figure 11b).

5.6. Summary: Incompatible Elements and First Series Transition Metals

[48] The abundance of highly incompatible, relatively immobile elements, such as Nb, La, and Th are highly correlated in Detroit Seamount lavas. There is, however, a large range from Site 884 tholeiitic basalt which are MORB-like to moderately incompatible element-rich alkalic basalt in the lower part of the Hole 1203A core. For a given lava type, tholeiitic or alkalic, lavas from Detroit Seamount are less enriched in highly incompatible elements than Mauna Kea lavas. In detail, the alkalic lavas from Site 1203 have several characteristics, relative depletion in Ba and Sr and two groups of Ti/Zr and Zr/Hf ratios, which reflect a distinctive petrogenesis. Finally, at a given MgO content, Sc abundances distinguish MORB and Hawaiian tholeiitic basalt. Lavas from Detroit Seamount have relatively high Sc abundance which overlaps with the MORB field.

5.7. Sr-Nd-Pb Isotopic Ratios

[49] Initial Sr-Nd-Pb isotopic ratios in Detroit Seamount lavas are reported in Table 4b. Because of low Rb/Sr ratios in the leached residues, which are plagioclase, pyroxene and olivine, the age correction on 87Sr/86Sr in most samples is very small (Tables 4a and 4b). Correction for 143Nd/144Nd are significant but the variability is similar for unleached and leached samples (Tables 4a and 4b). In contrast, the scatter of Pb isotopic ratios, especially 206Pb/204Pb, decreases after age-correction (Tables 4a and 4b).

[50] In a Sr-Nd diagram, Sites 883, 884 and 1204 lavas plot close to the age-corrected EPR MORB field (Figure 12a). Among Detroit Seamount lavas, Site 884 lavas have the highest 143Nd/144Nd and lowest 87Sr/86Sr ratios, and overlap with the 76 Ma MORB field. In contrast, Site 1203 tholeiitic and alkalic lavas are similar and overlap with the age-corrected Hawaiian rejuvenated stage lava and North Arch lava field which is intermediate between the EPR MORB and Mauna Kea shield fields (Figure 12a).

Figure 12.

(a) Sr-Nd isotopic ratios for Detroit Seamount lavas. (b) Pb-Pb isotopic ratios for Detroit Seamount lavas. (c) Values of 206Pb/204Pb versus 87Sr/86Sr for Detroit Seamount lavas. For comparison, fields are shown for EPR MORB, Garrett transform fault lavas, two extreme Hawaiian shields (Mauna Kea and Koolau), and Hawaiian rejuvenated stage lavas. All data points and fields have been age-corrected to 76 Ma, except for Site 884 lavas, which have been age-corrected to 81 Ma. A 2-sigma error bar is indicated unless the symbol is larger than the error bar. For the relatively young Hawaiian shield stage lavas and EPR MORB, parent/daughter ratios for age corrections should be those of the magma source. As a crude estimate of these ratios for tholeiitic basalt, we use average parent/daughter ratios in unaltered lavas (see Table 6). For the Sr and Pb isotopic systems this approach leads to overestimates for the age correction, and for the Nd isotopic system the age corrections are underestimates. Parent/daughter ratios in Koolau and Mauna Kea lavas are average values of relatively unaltered samples (K2O/P2O5 > 1.3) from Frey et al. [1994] and Huang and Frey [2003]. Parent/daughter ratios in Mauna Loa are average values of relatively unaltered samples (K2O/P2O5 > 1.3) from Hofmann and Jochum [1996]. Parent/daughter ratios in EPR-MORB are average N-MORB values from Sun and McDonough [1989]. Parent/daughter ratios in Garrett transform fault lavas are average values of lavas with 206Pb/204Pb < 18 [Wendt et al., 1999]. Because rejuvenated stage and North Arch alkalic basalt were formed by low extents of melting [e.g., Yang et al., 2003], which may lead to significant changes in parent/daughter ratios, we used two sets of parent/daughter ratios for Hawaiian Rejuvenated Stage lavas: N-MORB values for the orange solid line and Mauna Kea values for the black dashed line. Our selection of U/Pb and Th/Pb ratios in EPR MORB results in 238U/204Pb and 232Th/204Pb values of 10 and 26, respectively, which are about 2 times the values used by Regelous et al. [2003]. Data sources: Koolau: Roden et al. [1994], Lassiter and Hauri [1998]; Mauna Kea: J. G. Bryce et al. (Sr, Nd, and Os isotopes in a 2.84 km section of Mauna Kea Volcano: Implications for the geochemical structure of the Hawaiian plume, submitted to Geochemistry, Geophysics, Geosystems, 2004; hereinafter referred to as Bryce et al., submitted manuscript, 2004), Lassiter et al. [1996], Abouchami et al. [2000], Eisele et al. [2003]; Mauna Loa: Abouchami et al. [2000]; Loihi: Garcia et al. [1993, 1995, 1998], Norman and Garcia [1999]; EPR MORB: Niu et al. [1999], Regelous et al. [1999], Castillo et al. [2000]; Hawaiian rejuvenated stage and North Arch lavas: Stille et al. [1983], Roden et al. [1984], Tatsumoto et al. [1987], Chen and Frey [1985], Lassiter et al. [2000], Frey et al. [2000]; Garrett transform fault lavas: Wendt et al. [1999]; Site 883 and 884 lavas: Keller et al. [2000], Regelous et al. [2003]. EPR MORB fields in Pb-Pb plots are taken from Regelous et al. [2003].

Figure 12.


Table 6. Parent/Daughter Ratios Used to Calculate Initial Ratios for Fields in Figure 12
 KoolauMauna KeaMauna LoaLoihiEPR-MORBGTFa
  • a

    GTF, Garrett transform fault lavas.


[51] In Pb isotopic diagrams, Site 883, 1203 and 1204 lavas scatter around the Hawaiian shield field defined by lavas from Mauna Kea and Mauna Loa (Figure 12b). Compared with Site 1203 tholeiitic lavas, Site 1203 alkalic lavas are offset to higher 208Pb/204Pb ratio at a given 206Pb/204Pb ratio. In detail, Units 23 and 26 lie on an extension of the preshield stage Loihi field, which has higher 208Pb/204Pb ratio at a given 206Pb/204Pb ratio than the Mauna Kea field; Units 29 and 30 have even higher 208Pb/204Pb ratio at a given 206Pb/204Pb ratio (Figure 12b). It is important to point out that Sample 1203A 65R4W 9-13 from Unit 29 at Site 1203, which has the highest age-corrected 208Pb/204Pb ratio, has been carefully analyzed with two separate analyses for the isotopic ratios and three separate analyses for parent/daughter ratios. These multiple analyses yield similar Pb isotopic ratios and parent/daughter ratios (Table 4a). Compared with Site 1204 lavas, Site 883 lavas are offset to slightly lower 207Pb/204Pb ratio at a given 206Pb/204Pb ratio.

[52] Site 884 lavas are offset from the Hawaiian shield field to lower 207Pb/204Pb and 208Pb/204Pb ratios at a given 206Pb/204Pb ratio (Figure 12b). Relative to the EPR MORB field defined by data obtained by the triple spike technique, Site 884 lavas clearly have lower 207Pb/204Pb and trend to lower 206Pb/204Pb and 208Pb/204Pb. Consequently, Regelous et al. [2003] concluded that Site 884 lavas are not similar to EPR MORB or Hawaiian shield stage lavas. However, as indicated in Figure 12b, Detroit Seamount lavas form trends pointing to the field defined by Garrett transform fault lavas, which have low Pb isotopic ratios [Wendt et al., 1999]. The positive trend of Detroit Seamount in a 206Pb/204Pb versus 87Sr/86Sr diagram is unlike the negative trend defined by Hawaiian shield stage lavas, and it extrapolates toward the field defined by Garrett transform fault lavas (Figure 12c).

[53] In summary, lavas recovered from 4 drill sites are characterized by different age-corrected isotopic ratios. Even in the same drill hole, Hole 1203A, alkalic lavas and tholeiitic lavas have different age-corrected 208Pb/204Pb ratios. Compared with modern Hawaiian shields, Detroit Seamount lavas have lower 87Sr/86Sr and higher 143Nd/144Nd ratios, that range from MORB-like to similar to that in Hawaiian rejuvenated stage lavas and North Arch lavas. Trends of Detroit Seamount lavas in 206Pb/204Pb versus 208Pb/204Pb and, especially, 207Pb/204Pb extend to lower ratios than typical of EPR MORB, but they extend toward the field defined by Garrett transform fault lavas which have 206Pb/204Pb less than 18.

6. Discussion

6.1. Role of Crystal Fractionation

[54] Although the isotopic heterogeneity (Figure 12) and compositional changes caused by postmagmatic alteration preclude definition of liquid lines of descent, some geochemical trends provide compelling evidence that olivine and plagioclase fractionation and accumulation were significant processes during the evolution of lavas forming Detroit Seamount. For example, olivine fractionation is shown by the MgO-Ni correlation (Figure 11a) and plagioclase fractionation is indicated by the positive MgO-Al2O3 correlation for glasses (Figure 8c), the hyperbolic trend of Th-Sr for whole rocks and glasses (Figure 9f) and the positive Eu/Eu*-Sr/Nd correlation (Figure 13). This is a significant result because the compositions of Hawaiian shield lavas are dominantly controlled by olivine fractionation with a relatively minor role for plagioclase and clinopyroxene [e.g., Huang and Frey, 2003; Rhodes and Vollinger, 2004]. In contrast, like Detroit Seamount lavas, MORB compositions are typically controlled by plagioclase and olivine [e.g., Langmuir et al., 1992].

Figure 13.

(Sr/Nd)PM versus Eu/Eu* for Detroit Seamount lavas (whole rocks). Eu/Eu* = 2*[Eu]PM/([Gd]PM + [Sm]PM). Subscript PM indicates normalization to primitive mantle values from Sun and McDonough [1989]. These ratios are sensitive to plagioclase fractionation, and the range from >1 to <1 indicates accumulation and loss of plagioclase, respectively. Because Gd abundances for glasses are not reported, glasses are not included. The labeled units with high (Sr/Nd)PM also have anomalously high Al2O3 (Figure 8c).

6.2. Significance of Alkalic Basalt Deep in the Hole 1203A Core

6.2.1. Preshield or Postshield Stage Alkalic Lavas?

[55] In Hole 1203A, there is an upward change from intercalated alkalic and tholeiitic basalt to solely tholeiitic basalt (Figure 2). Specifically, four thick, ranging from 8 to 63 m, subaerially erupted compound pahoehoe alkalic flow units underlie a ∼300 m thick sequence of submarine erupted tholeiitic lava flows interbedded with volcaniclastic rocks. Although this alkalic to tholeiitic transition is expected in the preshield to shield transition at a single volcano, it could also reflect interfingering of lavas from two adjacent volcanoes at different growth stages. For example, shield stage tholeiitic lavas from Mauna Loa volcano overlie late-shield to postshield stage alkalic lavas from Mauna Kea volcano in the drill cores recovered by the Hawaii Scientific Drilling Project (HSDP) [Rhodes, 1996; Rhodes and Vollinger, 2004]. Did alkalic and tholeiitic lavas at Site 1203 erupt from the same volcano? Although there are erosional contacts between some units, such as Units 5 and 6, Units 23 and 24, Units 27 and 28, we cannot confidently use the contact relations in the non-continuous core containing numerous volcaniclastic units to determine if there is a regional uncomformity between the alkalic and overlying tholeiitic lavas. Therefore we evaluate isotopic differences between these units. The tholeiitic lavas from Mauna Loa in the HSDP core are clearly isotopically distinct from the underlying alkalic to transitional Mauna Kea lavas [e.g., Lassiter et al., 1996, Figure 6]. Therefore the similar age-corrected 87Sr/86Sr and 143Nd/144Nd ratios of the Site 1203 alkalic and tholeiitic lavas (Figure 12a) are evidence against the hypothesis of interfingering between two adjacent volcanoes at different growth stages. We cannot preclude the possibility that lavas from two adjacent volcanoes were similar in Sr and Nd isotopic ratios (e.g., Kilauea and Mauna Kea in Figure 6 of Lassiter et al. [1996]), but the subtle isotopic differences between alkalic and tholeiitic lavas at Site 1203 (Figures 12a and 12b) are typical of isotopic differences found in lavas from a single volcano (e.g., Lassiter et al. [1996] and Blichert-Toft et al. [2003] for Mauna Kea). Particularly interesting, these alkalic lavas have high 208Pb/204Pb at a given 206Pb/204Pb (Figure 12b), which is a distinct characteristic of preshield stage Loihi lavas [e.g., Garcia et al., 1993, 1995].

[56] Could these alkalic lava flows represent preshield stage lavas of Detroit Seamount? There are two examples of preshield stage lavas among Hawaiian volcanoes: Loihi Seamount and the submarine part of Kilauea volcano [Garcia et al., 1993, 1995; Lipman et al., 2002; Sisson et al., 2002]. In contrast to these submarine eruptives, the four alkalic flow units at Hole 1203A are subaerially erupted compound pahoehoe lavas. However, modern Hawaiian volcanoes are built on Cretaceous oceanic crust with a water depth of ∼4 km; in contrast, the near-ridge setting of Detroit Seamount at ∼80 Ma indicates that the seamount initiated in a shallow water depth environment. This inference is consistent with the inferred eruption environment of Site 1203 lavas (Figure 3). Since the maximum thickness of Loihi Seamount is ∼3.5 km, and the alkalic to tholeiitic transition occurs within the top several hundred meters [Garcia et al., 1995], it is possible that the preshield stage includes 2–3 km of alkalic basalt. The average water depth of zero age oceanic crust is less than 3 km in both North Pacific and North Atlantic oceans [e.g., Parsons and Sclater, 1977; Stein and Stein, 1992]. Therefore, if Detroit Seamount grew rapidly relative to subsidence of the young oceanic lithosphere, it is possible that its preshield stage included a subaerial growth stage. However, as discussed earlier, the upward transition from subaerial to submarine eruptives at Hole 1203A requires a subsequent decrease in accumulation rate relative to subsidence rate as the volcano aged.

6.2.2. Petrogenesis of Alkalic Lavas at Site 1203

[57] The four alkalic units form two compositional groups whose differences provide keys to understanding their petrogenesis. Specifically, relative to lavas from Units 29 and 30, lavas from Units 23 and 26 have lower K2O and Rb abundances (Figures 6a and 6b), lower Ba/Th, Ti/Zr and higher Zr/Hf ratio (Table 5). Common phenocryst phases, such as plagioclase and olivine, do not affect these ratios, but phlogopite, a mineral that fractionates Ba/Th, Ti/Zr and Zr/Hf (Table 7), occurs in preshield stage alkalic Kilauea lavas [Sisson et al., 2002]. For these Site 1203 alkalic lavas, a difficulty with inferring a petrogenetic role for phlogopite as a residual or fractionating mineral is that phlogopite is not a phenocryst phase, presumably because of the low K2O contents (<1.8%) (Table 3a). Melts saturated with phlogopite typically have more than 2% K2O (Table 7); however, Mengel and Green [1986] infer that only 1.6% K2O is required to saturate a SiO2-undersaturated nephelinite with phlogopite. The K2O content in Units 29 and 30 ranges from ∼1.5 to 1.8%, but is less than 1% in Units 23 and 26 (Figure 6a; Table 3a). As discussed in section 5.2, the low K2O content in Units 23 and 26 reflects K2O loss during alteration. If we assume that K2O contents of Units 29 and 30 reflect magmatic ratios (K2O/P2O5 = 2.3–2.9, with K2O = 1.5–1.8%), it is possible that phlogopite was a residual phase during partial melting. Although phlogopite may not be stable during the high temperature melting typically inferred for a plume environment [Class and Goldstein, 1997; Sisson et al., 2002], the presence of fluorine increases the thermal stability of phlogopite [e.g., Foley et al., 1986].

Table 7. Partition Coefficients of Selected Elements Between Phlogopite and Melt
ReferenceMelt TypeBaThZrHfTiSrLaCeYK2O, %
In MeltIn Phlogopite
Villemant et al. [1981]trachyte10.0 2.51.8 0.7     
Adam et al. [1993]basanite2.9 0.13  0.22  <12.747.28
LaTourrette et al. [1995]basanite3.680.00140.0170.191.7680.1590.028 0.0182.157.91
Foley et al. [1996]alkaline lamprophyre3.480.01450.023  0.183 0.00780.0072.58.87
Schmidt et al. [1999]lamproite1.030.000020.012 0.870.038 0.00002 8.79.9
 lamproite1.610.0000050.017 0.920.058 0.00002 99.8
 lamproite0.560.00020.009 0.8270.016 0.0002 8.69.7
Green et al. [2000]basanite3.34 0.0080. 0.0182.159.16

[58] A specific scenario was postulated by Sisson et al. [2002] to explain the generation of submarine nephelinite to basanite lavas at Kilauea Volcano. They suggested that these alkalic lavas formed by low degree melting of phlogopite-bearing garnet pyroxenite cumulates formed during high pressure solidification of plume-derived magmas. If these melts bearing a residual phlogopite signature, i.e., relatively low Ti/Zr and Ba/Th and high Zr/Hf, interact with peridotite or mix with a peridotite-derived melt, their high K2O content would be diminished. Hence they would no longer be phlogopite-saturated, but they could retain the trace element signature of residual phlogopite. This hypothesis is consistent with the geochemical data for the alkalic lavas from Site 1203.

6.3. Melt Segregation Pressure of Detroit Seamount Lavas

[59] On the basis of plate reconstruction, Mammerickx and Sharman [1988] and Lonsdale [1988] suggested that Detroit Seamount was built on a relatively young and thin lithosphere (Figure 1a). The paleolatitude [Tarduno et al., 2003] and the oceanic paleodepth of Detroit Seamount [Caplan-Auerbach et al., 2000] are consistent with this inference. Therefore during formation of Detroit Seamount decompression melting is likely to have continued to a shallower depth and the mean pressure of melt segregation was lower than that during formation of modern Hawaiian volcanoes which are constructed on thick (∼100–110 km [Li et al., 2004]) Cretaceous oceanic lithosphere (Figure 14). Are the compositional data for Detroit Seamount lavas consistent with this interpretation? If the mean pressure of melt segregation increases from MORB to Detroit Seamount lavas to Mauna Kea lavas (Figure 14), these lava groups should show systematic differences in compositional parameters sensitive to pressure.

Figure 14.

Schematic diagram showing melting column length for EPR MORB, Detroit Seamount, and Mauna Kea, i.e., the depth interval from onset of melting to the lithosphere. The mean pressure of melt segregation increases from MORB to Detroit Seamount lavas to Mauna Kea lavas.

6.3.1. SiO2 and Total Iron

[60] The SiO2 and FeO contents of partial melts of peridotite are a function of melt segregation pressure and extent of melting [e.g., Walter, 1998; Stolper et al., 2004, Figure 13]. Regelous et al. [2003] noted that compared to young Emperor Seamounts, whole-rock tholeiitic basalt compositions from Detroit Seamount are offset to lower total iron at a given MgO content; consequently, they inferred that melt segregation at Detroit Seamount occurred at relatively low pressure beneath thin lithosphere. Consistent with this inference, whole-rock FeO and SiO2 contents for tholeiitic basalt from Detroit Seamount overlap with MORB field, but inexplicably, the SiO2 content of the associated glasses do not (Figures 8a and 8b).

[61] There is also evidence that the depth of melt segregation changes during growth of a Hawaiian volcano. Compared with tholeiitic shield stage lavas from Mauna Kea, postshield stage alkalic lavas from Mauna Kea have higher total iron content but lower SiO2 content (Figures 8a and 8b). This result was used to infer that these alkalic lavas were generated by lower extent of partial melting at higher pressure [e.g., Yang et al., 1996]. Similarly, at Detroit Seamount, the alkalic lavas have higher FeO* and lower SiO2 contents than tholeiitic lavas (Figures 8a and 8b). Hence we infer that compared with tholeiitic lavas from Detroit Seamount, the alkalic lavas were generated by a lower degree of melting at higher pressure.

6.3.2. Sc

[62] Sc is compatible in garnet and clinopyroxene [e.g., Irving and Frey, 1978; Hart and Dunn, 1993; Blundy et al., 1995, 1998; van Westrenen et al., 1999]. Hawaiian lavas have lower Sc abundance than MORB at a given MgO content [Frey et al., 1994] (Figure 11b). If the source of Hawaiian lavas has a Sc abundance similar to the source of MORB, the lower Sc abundances in Hawaiian shield lavas can be explained as a result of more residual garnet [Frey et al., 1994]. At a certain MgO content, Detroit Seamount lavas have higher Sc abundance than Mauna Kea and Koolau lavas (Figure 11b), which implies less residual garnet for Detroit Seamount lavas. This result is consistent with the hypothesis that Detroit Seamount lavas segregated at lower mean pressure than Hawaiian shield stage lavas.

[63] The similar Sc abundance of alkalic and tholeiitic basalt (Figure 11b) from Detroit Seamount is inconsistent with the inferred higher pressure of melt segregation for the alkalic basalt based on SiO2 and total iron contents. However, Sc is an incompatible element during melting of peridotite; therefore the unexpectedly high Sc abundance of the alkalic basalt may reflect the combination of a lower extent of melting and higher pressure of melt segregation. In fact, alkalic and tholeiitic basalt from Mauna Kea Volcano also have similar Sc abundances [see Huang and Frey, 2003, Figure 5e].

6.3.3. Na2O/TiO2 and Tb/Yb

[64] During mantle melting, the partition coefficient of sodium (as Na2O) between clinopyroxene and melt increases with increasing pressure [e.g., Kinzler, 1997; Putirka, 1999; Longhi, 2002]; that is, the proportion of jadeite component in clinopyroxene increases with increasing pressure. In contrast, the partition coefficient of titanium (as TiO2) is not strongly dependent on pressure; it may decrease slightly with increasing pressure [e.g., Longhi, 2002]. Therefore the Na2O/TiO2 ratio yields information on the pressure of melt segregation with relatively lower Na2O/TiO2 ratio indicating higher mean pressure of melt segregation [e.g., Putirka, 1999, Figure 5]. The ratio of Tb/Yb in a melt can also be used as an indicator for pressure of melt segregation because (DTb/DYb)garnet/melt ≪ (DTb/DYb)clinopyroxene/melt [e.g., van Westrenen et al., 1999]. Hence a higher proportion of residual garnet results in higher Tb/Yb ratio due to the compatibility of Yb in garnet.

[65] Compared with EPR N-MORB, Mauna Kea lavas have lower Na2O/TiO2 but higher Tb/Yb ratios at a certain MgO content (Figure 15). These differences in Na2O/TiO2 and Tb/Yb ratios show that relative to EPR N-MORB, Mauna Kea lavas segregated at a higher pressure, although source heterogeneity in these ratios may also contribute to this difference. As indicated in Figure 15, Na2O/TiO2 decreases and Tb/Yb increases from Site 884 tholeiitic lavas to Site 1203 tholeiitic lavas and Site 883 and 1204 alkalic lavas to Site 1203 alkalic lavas. Site 884 lavas are distinct and overlap with the low pressure end of the EPR N-MORB field. All other Detroit Seamount lavas overlap with the low Na2O/TiO2 and high Tb/Yb (relatively high pressure) end of the MORB field. This observation is consistent with the hypothesis that the mean pressure of melt segregation decreases from Mauna Kea lavas to Detroit Seamount lavas to MORB. The lower Na2O/TiO2 and higher Tb/Yb ratios in Site 1203 alkalic lavas relative to Site 1203 tholeiitic lavas indicate that the former segregated at higher pressure.

Figure 15.

(a) MgO versus Na2O/TiO2 and (b) Tb/Yb. Both whole-rock and glass data from Detroit Seamount are plotted. See the Figures 8a8d and 10 captions for data sources.

6.3.4. Summary

[66] In summary, compositional indicators of mean pressure of melt segregation, such as Sc abundance and ratios of Na2O/TiO2 and Tb/Yb, show that tholeiitic lavas from Detroit Seamount are more similar to MORB than young Hawaiian shield lavas. This similarity to MORB is inferred to reflect segregation of Detroit Seamount lavas at low pressure beneath thin lithosphere. The extremes are expressed in Site 884 lavas (lowest pressure) and Site 1203 alkalic lavas (highest pressure).

6.4. Did Detroit Seamount Lavas Sample a Depleted MORB-Related Component or a Depleted Plume Component?

6.4.1. Controversy About the Origin of Detroit Seamount Lavas

[67] Studies of lavas recovered from Sites 883 and 884 during ODP Leg 145 have led to two different interpretations for the depleted component found in Detroit Seamount lavas. Keller et al. [2000] used age-corrected Sr-Nd-Pb isotopic ratios in Detroit Seamount lavas, especially Site 884 lavas, to conclude that lavas forming Detroit Seamount are “indistinguishable from mid-ocean ridge basalt”. Hence they proposed that the Hawaiian plume interacted with young and hot lithosphere and “entrained enough of the isotopically depleted upper mantle to overwhelm the chemical characteristics of the plume itself”. In contrast, using a high precision triple spike technique, Regelous et al. [2003] argued that the Pb isotopic compositions of Site 884 lavas are not similar to either EPR MORB or modern Hawaiian shield stage lavas [Regelous et al., 2003, Figure 6] (Figure 12b). They argued that the Site 884 lavas were derived from a depleted plume component that is not sampled by modern Hawaiian volcanoes. On the basis of Nd-Hf isotopic correlation, Kempton et al. [2002], Thompson et al. [2002], and Frey et al. [2005] also suggest that Detroit Seamount lavas sampled a depleted plume component. With additional data for lavas from Detroit Seamount, our objective is to evaluate these alternative hypotheses.

6.4.2. Constraints From Sr-Nd Isotopes

[68] Age-corrected Sr and Nd isotopic ratios of basalt from 3 of the 4 drill sites on Detroit Seamount (Sites 883, 884 and 1204) overlap with the field calculated for 76 Ma EPR MORB (Figure 12a). A new observation is that both tholeiitic (Site 884) and alkalic basalt (Sites 883 and 1204) overlap with the MORB field, but alkalic basalt has slightly higher 87Sr/86Sr and lower 143Nd/144Nd than tholeiitic basalt. The simplest interpretation is that a depleted MORB-like component, perhaps slightly heterogeneous in 87Sr/86Sr and 143Nd/144Nd, was melted to various extents to form Detroit Seamount lavas.

6.4.3. Constraints From Pb Isotopes

[69] In plots of Pb isotopic ratios, our new data for Sites 1203 and 1204 basalt confirm the observation by Regelous et al. [2003] that for a given 206Pb/204Pb ratio, most lavas from Detroit Seamount have lower 207Pb/204Pb ratios than EPR MORB data obtained using the triple-spike technique (Figure 12b). Regelous et al. [2003] noted that measured 207Pb/204Pb in Site 884 lavas, ranging from 15.43 to 15.46, are lower than data for most EPR MORB. Consequently, the difference in 207Pb/204Pb between Site 884 lavas and EPR MORB is not a result of an inaccurate age-correction. In addition, as emphasized by Regelous et al. [2003] the 206Pb/204Pb for lavas from Site 883 and 884 extend to considerably lower ratios (<18) than the EPR MORB field defined by data obtained using the triple-spike technique.

[70] Is the field defined by triple-spike data representative of young Pacific MORB? In order to answer this question, we used the Web database Pet DB ( Three important results arise from assessing Pb isotopic ratios for the 996 samples (Figure 16). (1) The EPR MORB field defined by data obtained by the triple-spike technique is representative of the Pacific MORB database, but it does not include the extremes. (2) Although the database for old Pacific MORB (distant from active spreading centers) is sparse, Pb isotopic ratios of such lavas are similar to young Pacific MORB (Figure 16). (3) Pacific MORB with 206Pb/204Pb < 18 are very unusual. The majority of Pacific MORB with low 206Pb/204Pb (<18) erupted at a spreading center in the Garrett transform fault (Figure 16). Wendt et al. [1999] proposed that Garrett transform fault lavas reflect two-stage melting of a MORB source; i.e., initial melting at the EPR axis followed by melting of the residue to generate the incompatible element-depleted Garrett transform fault lavas. Therefore, if the low 206Pb/204Pb (<18) component in Detroit Seamount lavas is a MORB-related component, as suggested by Keller et al. [2000], it is not typical Pacific MORB. Sampling of this atypical component may occur only at high extents of melting, such as achieved in the two-stage melting scenario proposed for lavas erupted in the Garrett transform fault [Wendt et al., 1999], and when the Hawaiian plume was near a ridge axis, such as Detroit Seamount at ∼80 Ma [Keller et al., 2000].

Figure 16.

Pb isotopic plots for Pacific MORB. Data are downloaded from PeT DB. The red field with dots is the EPR MORB field defined by data obtained using the triple-spike technique [Galer et al., 1999]. The low 206Pb/204Pb (<18) end of the Pacific MORB field is defined by Garrett transform fault lavas [Wendt et al., 1999]. Pacific MORB with 206Pb/204Pb< 18 are also reported by Barrett [1983], White et al. [1987], Fornari et al. [1988], Hanan and Schilling [1989], Castillo et al. [1998], Sturm et al. [1999], and Haase [2002]. Old Pacific MORB are shown for comparison, and they do not differ from young Pacific MORB. Data sources: Bass et al. [1973], Hickey-Vargas [1991], Castillo et al. [1991, 1992, 1994], and Janney and Castillo [1997].

6.4.4. Constraints From Incompatible Element Abundance Ratios

[71] In both MORB and Hawaiian tholeiitic basalt many incompatible trace element ratios are correlated with radiogenic isotopic ratios [e.g., Niu et al., 2002; Huang and Frey, 2003]. Among Hawaiian shield lavas, La/Nb is positively correlated with 87Sr/86Sr and negatively correlated with 143Nd/144Nd; Detroit Seamount tholeiites, like EPR MORB, define opposite correlations (Figures 17a and 17b). Obviously, Detroit Seamount lavas sampled a depleted, low 87Sr/86Sr, 206Pb/204Pb and high 143Nd/144Nd, La/Nb component that was not sampled by Hawaiian shield stage lavas. Is this depleted component similar to that sampled by Garrett transform fault lavas (Figure 17), as argued on the basis of isotope compositions?

Figure 17.

(a and b) La/Nb versus 87Sr/86Sr and 143Nd/144Nd for Detroit Seamount lavas (whole rocks). Unit 31, a plagioclase-rich unit (Table 2a), at Site 1203 has a high La/Nb ratio, which has been confirmed by three ICP-MS analyses (1.06, 1.07, and 1.10). Note that Hawaiian shield lavas and Detroit Seamount lavas define opposing slopes in the Sr and Nd panels. (c) La/Nb versus Ba/Th for Detroit Seamount glasses. Only Detroit Seamount glass data are plotted to avoid the alteration effect on whole-rock Ba data. The important point is that in this panel, Site 884 data do not overlap with the EPR MORB field. Data sources: Koolau: Frey et al. [1994], Roden et al. [1994], Lassiter and Hauri [1998], S. Huang and F. A. Frey (Temporal geochemical variation within the Koolau shield: A trace element perspective, submitted to Contributions to Mineralogy and Petrology, 2004); Mauna Kea: Bryce et al. (submitted manuscript, 2004), Eisele et al. [2003], Huang and Frey [2003]; Garrett transform fault: Wendt et al. [1999]; EPR N-MORB: Niu et al. [1999], Regelous et al. [1999]. The fields for Hawaiian shields are from Huang and Frey [2003].

[72] As discussed earlier, tholeiitic basalt from Site 884 has MORB-like incompatible element patterns (Figure 10). There is, however, an important difference between Site 884 lavas and MORB. Relative to MORB, Hawaiian lavas have high Ba/Th (>100) [Hofmann and Jochum, 1996; Huang and Frey, 2003; Yang et al., 2003]; such ratios are interpreted as characteristic of recycled oceanic lithosphere in the Hawaiian plume [Hofmann and Jochum, 1996; Huang et al., 2000; Sobolev et al., 2000]. High Ba/Th ratios are also characteristic of Detroit Seamount glasses, and Site 884 glasses are at the high end of the Ba/Th range in Detroit Seamount lavas (85-113) (Figure 17c). In contrast, Garrett transform fault lavas have low Ba/Th (35-57) which overlap with those of N-MORB [Wendt et al., 1999] (Figure 17c). We therefore speculate that the depleted component sampled by Site 884 lavas has high Ba/Th (>100), and is unrelated to the low 87Sr/86Sr and 206Pb/204Pb source of Garrett transform fault lavas. Rather, it is an intrinsic component of the plume.

[73] Young Hawaiian rejuvenated stage lavas also contain a component with low 87Sr/86Sr, 206Pb/204Pb and high 143Nd/144Nd and Ba/Th. Such lavas have relatively high incompatible trace element concentrations, and are thought to represent small degrees of melting of plume mantle mixed with melts of a depleted source [e.g., Yang et al., 2003]. Frey et al. [2005] argue that the low 87Sr/86Sr component in these lavas is the same as that present in Detroit lavas, i.e. it is intrinsic to the plume, and is sampled during a later stage of melting of the Hawaiian plume that previously melted to produce the tholeiitic shields. If this interpretation is correct, it implies that the depleted component has been an integral part of the Hawaiian plume for at least 80 Ma.

6.4.5. Summary

[74] 1. We agree with Regelous et al. [2003] that the Pb isotopic data, especially 206Pb/204Pb (<18), for Detroit Seamount lavas reflect a depleted component that is not commonly present in either ancient or recent Pacific MORB. Such a component is present in lavas erupted from a spreading center in the Garrett transform fault. Therefore if the depleted component arises from a MORB source, it is only sampled under unusual conditions such as the two-stage melting proposed for Garrett transform fault lavas [Wendt et al., 1999] or the proximity of the Hawaiian plume to a ridge axis (Detroit Seamount).

[75] 2. The occurrence at Sites 883 and 1204 of alkalic lavas that have depleted isotopic characteristics shows that the depleted component is not restricted to tholeiitic basalt. In the context of the Regelous et al. [2003] hypothesis the depleted plume component with a high solidus temperature was melted to varying extents.

[76] 3. High Ba/Th is a characteristic feature of Hawaiian lavas and the most depleted lavas from Detroit Seamount have Ba/Th ratios exceeding those of MORB (Figure 17c). Their high Ba/Th implies that their source is also characterized by high Ba/Th. Consequently, this source is unlike the source of Garrett transform fault lavas in terms of Ba/Th.

[77] 4. Support for the hypothesis of Regelous et al. [2003] that the depleted component is plume related arises from rejuvenated stage Hawaiian lavas. Their Pb isotopic ratios also extend to low 206Pb/204Pb ratios [Fekiacova and Abouchami, 2003; Frey et al., 2005], and they have high Ba/Th [Yang et al., 2003]. Therefore a depleted plume component may be sampled by rejuvenated stage lavas during a second stage of plume melting [Ribe and Christensen, 1999] and during ascent of the plume beneath young and thin oceanic lithosphere, i.e., at Detroit Seamount.

6.5. A Hypothesis for the Geochemical and Age Differences Between Lavas From Site 884 and Site 1203

[78] The age of Site 884 lavas is ∼81 ± 1 Ma on the basis of 40Ar-39Ar data for a reverse polarity, tholeiitic basalt [Keller et al., 1995]. At Site 1203, Duncan and Keller [2004] obtained an average age of ∼76 ± 1 Ma for five normal polarity, tholeiitic lavas. On the basis of its large area (Figure 1b), it is likely that Detroit Seamount consists of several coalesced shields, much like the Big Island of Hawaii. However, an age difference of 5 Myr between Site 1203 and Site 884 is much larger than that for the five shields on the Big Island of Hawaii which all formed over the past 1 Myr. What is the explanation for an age difference of 5 Myr for tholeiitic lavas erupted at sites separated by only 48 km?

[79] It is well-established that near-axis plumes interact with the spreading ridge axis. Using the Galapagos Platform as an analogy, plume-related volcanism in a near-ridge environment is more dispersed and long-lived [e.g., Sinton et al., 1996] than at an intraplate location, such as Hawaii. We propose a scenario that can explain both the old age of Site 884 lavas and their relatively shallow, MORB-like, melt segregation depths (Figure 15). Sleep [2002] noted that “off-axis hot spots appear to shut off at the time that an on-axis hot spot becomes active along an axis-approaching track”; i.e., there is a time when volcanism ceases above the hot spot and partial melting of plume material does not occur until ascent at the spreading center (Figure 18a). If the plate moved faster than the plume motion, at some later time the plume was overlain by plume-related lithosphere created at the spreading center (Figure 18b). Perhaps at ∼76 Ma, the Hawaiian plume was distant from the spreading center, and intraplate volcanism represented by lavas from Site 883, 1203 and 1204 erupted in close proximity to older Site 884 lavas which erupted at the spreading center (Figure 18c). Melting to a shallow depth at a spreading center for Site 884 lavas implies a large melting extent, and consequently Site 884 magmas sampled a depleted and refractory component with low 206Pb/204Pb.

Figure 18.

Plume-ridge interactions (adapted from Sleep [2002]). Figure 18a shows that when a plume is near a ridge axis, the plume material flows toward the ridge axis, where it melts upon ascent. This scenario may be appropriate for 81 Ma lavas at Site 884. Figure 18b shows the situation where the plume and the lithospheric plate are moving in the same direction, but the plate is migrating faster than the plume. In this case, Figure 18c shows that eventually plume-related magmas and lithosphere created at the ridge axis (Figure 18a) may override the plume and young (76 Ma lavas at Site 1203) lavas may erupt in proximity to old (81 Ma) lavas.

7. Conclusions

[80] On the basis of studies of drill cores from four drill sites at Detroit Seamount and the comparisons with EPR N-MORB and Mauna Kea lavas, our major conclusions are as follows:

[81] 1. The rapid subsidence rates inferred for Detroit Seamount support the view that this seamount formed on young and thin oceanic lithosphere close to a spreading ridge axis. Also consistent with this inference, major and trace element compositions of Detroit Seamount lavas show that their parental magmas were segregated at a lower pressure than Hawaiian lavas which formed under old and thick oceanic lithosphere.

[82] 2. The upward transition from intercalated alkalic and tholeiitic basalt to solely tholeiitic basalt at Site 1203 may reflect the preshield to shield stage transition, whereas alkalic basalts from Sites 883 and 1204 probably represent postshield stage lavas. Like preshield stage alkalic Kilauea lavas, phlogopite appears to have played an important role in generating the preshield stage alkalic lavas at Site 1203. Also, similar to preshield stage Loihi lavas, Site 1203 alkalic lavas are characterized by high 208Pb/204Pb at a given 206Pb/204Pb.

[83] 3. Both tholeiitic and alkalic Detroit Seamount lavas have lower concentrations of highly incompatible elements, lower 87Sr/86Sr and higher 143Nd/144Nd than the corresponding lava types from young Hawaiian volcanoes. Also Detroit Seamount lavas contain a component with unradiogenic Pb that is not present in Hawaiian shield stage lavas or EPR MORB. Such a component is present in lavas erupted from a spreading center within the Garrett transform fault. However, Site 884 lavas with the most depleted isotopic characteristics have high Ba/Th unlike those of MORB and Garrett transform fault lavas. If the depleted component has high Ba/Th, it is plume-related. Hawaiian rejuvenated stage lavas also have high Ba/Th and define a trend to low 87Sr/86Sr and 206Pb/204Pb, similar to the trend of Detroit lavas (Figure 12c). Rejuvenated stage melts have been proposed as second-stage melts of the Hawaiian plume [Ribe and Christensen, 1999]. Therefore this depleted component may be a refractory component of the plume that is only sampled at high degrees of melting, such as occurred at ∼81 Ma when the plume was near a spreading center.

[84] 4. The surprisingly large age difference (∼81 and 76 Ma) between lavas at two drill sites on Detroit Seamount may reflect the complexity of plume-related volcanism in a near ridge axis environment.


[85] We thank the crew and scientific staff of the Joides Resolution during ODP Leg 197 for their help in acquiring and describing lavas from Detroit Seamount. This research was supported by the U.S. Science Support Program. The Ocean Drilling Program is sponsored by the National Science Foundation and participating countries under the management of Joint Oceanographic Institutions, Inc. A grant from the Deutsche Forchungsgemeinschaft is gratefully acknowledged by M. Regelous. We thank S. Bowring for access to the MIT Mass Spectrometer facility and especially F. Dudas for his dedicated efforts obtaining isotopic data, M. Rhodes for access to the XRF facility at UMASS and M. Vollinger for his help in XRF analysis, and B. Grant and R. Kayser for their assistance in ICP-MS analysis. M. Regelous thanks B. Paterson and J. Wade for their help with the LA-ICP-MS and electron probe analyses. This paper benefited from reviews by P. Castillo and T. Sisson, the editorial comments of W. White and R. Duncan, and discussion with S. Parman, A. Hofmann, W. Abouchami, S. Galer, and C. Hawkesworth.