Journal of Petrology

This Article Abstract FREE Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Similar articles in ISI Web of Science Alert me to new issues of the journal Add to My Personal Archive Download to citation manager Search for citing articles in: ISI Web of Science (38) Request Permissions Google Scholar Articles by LEAT, P. T. Articles by PEARCE, J. A. Search for Related Content GeoRef GeoRef Citation Social Bookmarking          What's this?

Journal of Petrology Volume 41 Number 6 Pages 845-866 2000
© Oxford University Press 2000

Magma Supply in Back-arc Spreading Centre Segment E2, East Scotia Ridge

P. T. LEAT^1,*, R. A. LIVERMORE¹, I. L. MILLAR^1,3 and J. A. PEARCE^2,

¹BRITISH ANTARCTIC SURVEY, HIGH CROSS, MADINGLEY ROAD, CAMBRIDGE CB3 0ET, UK
²DEPARTMENT OF GEOLOGICAL SCIENCES, UNIVERSITY OF DURHAM, DURHAM DH1 3LE, UK
³c/o NERC ISOTOPE GEOSCIENCES LABORATORY, KINGSLEY DUNHAM CENTRE, KEYWORTH, NOTTINGHAM NG12 5GG, UK

Received March 12, 1999; Revised typescript accepted November 23, 1999

	ABSTRACT

TOP ABSTRACT INTRODUCTION REGIONAL SETTING THE EAST SCOTIA RIDGE SEGMENT E2 RESULTS DISCUSSION REFERENCES

 
Segment E2 is situated in the back-arc East Scotia Ridge. The segment is unusual in that it has an axial topographic high underlain by a seismically imaged melt lens. The axis of the segment, which is 70 km long, was sampled at 2 km spacing. There is strong correlation between compositions and bathymetry, and there is no evidence that lateral flow of magmas along dykes within the segment was more than 25 km. Magmas are more evolved towards the summit, indicating that magma fractionation took place within the imaged melt lens. Na_8·0 is roughly constant at 2·6, implying uniform degree of partial melting, except for some samples at the summit that have Na_8·0 = 2·2. Compositions closest to N-MORB occur at the segment tips, and LREE/HREE ratios increase toward the summit. None of the magmas were derived from depleted sub-arc mantle. Nevertheless, most compositions in the segment were modified by slab-derived components. The low-Na_8·0 samples have high Ba/Nb compared with the rest of the segment. Dredged lavas from the lateral flanks of the summit have the most extreme compositions, including ones derived from plume mantle, and are end-members for magma mixing processes that controlled compositions.

KEY WORDS: geochemistry; petrology; volcanism; back-arc; subduction

	INTRODUCTION

TOP ABSTRACT INTRODUCTION REGIONAL SETTING THE EAST SCOTIA RIDGE SEGMENT E2 RESULTS DISCUSSION REFERENCES

 
The back-arc East Scotia Ridge is spreading at an intermediate rate of 65–70 km/My (full rate), but has an axial morphology along most of its length typical of slow ocean ridge spreading rates. However, two segments, one toward the north end of the ridge, and one toward the south, have magmatically inflated morphologies with axial highs, typical of fast-spreading centres. The axial highs are a result of greater than normal magma production. As the spreading rate is essentially constant along the length of the ridge, the excess magma production may best be explained if mantle that is more easily melted than normal underlies the inflated segments (Livermore et al., 1997). This paper describes the results of a detailed geochemical survey of inflated segment E2, which was undertaken to determine why the excess mantle melting has occurred. The northern inflated segment, of 70 km length, has a central volcanic axial high with over 600 m of relief, a summit graben of 2 km width, and seismically imaged axial melt lens 3 km below the sea floor, all features that indicate that the segment is magmatically very active (Livermore et al., 1997). Bathymetric data indicate that the excess magmatism is a long-lived feature of this segment (Livermore et al., 1997, fig. 2a). The geochemical data are used to comment on the extent of lateral migration of magma along dykes within the segment, and the nature and origin of the mantle sources. The work forms part of the multidisciplinary British Antarctic Survey Sandwich Lithospheric and Crustal Experiment (SLICE) in the South Sandwich Arc and its associated back-arc (Larter et al., 1998).

	REGIONAL SETTING

TOP ABSTRACT INTRODUCTION REGIONAL SETTING THE EAST SCOTIA RIDGE SEGMENT E2 RESULTS DISCUSSION REFERENCES

 
There are three major plates in the Scotia Sea—the Antarctic, South American and Scotia plates—all of which are moving slowly (<22 km/My) relative to the global hotspot reference frame (Barker, 1995). The north and south edges of the Scotia plate, the North Scotia Ridge and South Scotia Ridge, respectively, are essentially amagmatic strike slip boundaries (Fig. 1). Within this system of slow-moving plates, the small Sandwich plate is overriding the South American plate at a rate of 70–85 km/My (Pelayo & Wiens, 1989). Earthquake solutions indicate the subducting South American slab dips at 45–55% to the west beneath the Sandwich plate, and is being torn from the non-subducting South American plate north of the arc in a scissor-like motion (Forsyth, 1975; Brett, 1977). Situated on the Sandwich plate, the South Sandwich Islands form the active volcanic arc related to this subduction. Both the Sandwich plate and the easternmost part of the Scotia plate are oceanic in character, and the South Sandwich Islands are an example of a fully intra-oceanic arc. The East Scotia Ridge (Livermore et al., 1997) is the back-arc spreading centre to the South Sandwich arc (Fig. 1). The arc–back-arc system has occupied this isolated intra-oceanic position for 10 My (Barker, 1995). Alvarez (1982) suggested that mantle in the vicinity is moving eastward, pushing the subducting slab and hence the trench–arc system in its path. However, Livermore et al. (1997) suggested that mantle is flowing in a westerly direction around the northern end of the slab, raising the possibility that the eastward motion of the Sandwich plate is powered entirely by slab rollback.

View larger version (36K): [in this window] [in a new window]   Fig. 1. (a) Tectonic setting of the East Scotia Ridge within the South Atlantic region. The small, entirely oceanic Sandwich plate is situated between the South American, Antarctic and Scotia plates. The North Scotia Ridge and the South Scotia Ridge are largely amagmatic transform plate boundaries. The slow-spreading South American–Antarctic Ridge joins the southern boundary of the Sandwich plate to the Bouvet triple junction. (b) Sketch map of the Sandwich plate and the East Scotia Ridge. The East Scotia Ridge is being generated by east–west divergence of the Scotia and Sandwich plates. It comprises 10 segments, from north to south E1–E10. Dredge sites DR.20–DR.24 and DR.56, -57 and -60 described in the literature are indicated. The South American plate is subducting beneath the Sandwich plate at the South Sandwich Trench. The South Sandwich Islands, from Zavadovski in the north to Thule in the south, form the volcanic arc related to this subduction.

 

The 11 main South Sandwich Islands form a crescent-shaped volcanic arc of 350 km length. The islands are small, typically 3–8 km (up to 20 km) across, and entirely volcanic in origin (Holdgate & Baker, 1979; Baker, 1990). Most of the islands have abundant evidence for young volcanic activity: the oldest radiometric age for the group is 3·1 Ma (Baker et al., 1977). The volcanic rocks are dominantly basalt and basaltic andesite, together with andesite and dacite. Some belong to the tholeiitic and some to the calc-alkaline series (Baker, 1968, 1990; Pearce et al., 1995a). Magnetic data indicate that the islands are built on ocean crust that formed at the East Scotia Ridge up to 10 My ago (Barker & Hill, 1981; Barker, 1995), and the arc is regarded as a classic example of an early stage of island arc development (Baker, 1968).

The South American–Antarctic Ridge is a very slow-spreading (9 km/My, half rate) mid-ocean ridge along which long transform faults are interspaced with short spreading segments (le Roex et al., 1985; Barker & Lawver, 1988). At its eastern end, this ridge joins the Mid-Atlantic and Southwest Indian ridges at the Bouvet triple junction (Fig. 1). Ridge morphology, bathymetry and geochemistry indicate that there are at least two mantle plumes near the triple junction. The Bouvet plume is thought to underlie Bouvet Island, a small (9 km by 7 km) ocean island situated near the Southwest Indian Ridge (Sun, 1980; le Roex & Erlank, 1982; Verwoerd, 1990). The Shona mantle plume is thought to underlie the southern mid-Atlantic Ridge at about 52°S, where the ridge shallows to 1·5 km depth and the geochemistry of dredged basalts indicates plume-related mantle (Douglass et al., 1999). Dredge samples from the South American–Antarctic Ridge have been used to suggest that mantle from at least one of these mantle plumes is migrating westward toward the Sandwich plate (le Roex et al., 1985; Kurz et al., 1998).

	THE EAST SCOTIA RIDGE

TOP ABSTRACT INTRODUCTION REGIONAL SETTING THE EAST SCOTIA RIDGE SEGMENT E2 RESULTS DISCUSSION REFERENCES

 
The East Scotia Ridge back-arc oceanic spreading centre (55°15'S, 29°30'W to 60°30'S, 29°30'W) consists of nine main segments, called E1 (north) to E9 (south) (Fig. 1) (Livermore et al., 1997; Larter et al., 1998). A southernmost segment, E10, is anomalous, being situated on the south margin of the Sandwich plate. The spreading history and form of the ridge is well documented from magnetic, satellite-derived gravity, and HAWAII Mapping and Researcher 1 (MR1)-derived bathymetric and side-scan data (Barker & Hill, 1981; Livermore et al., 1994, 1995). Segments E3–E8 have axial rifts with their rift floors at 4 km water depth and are typical of slow-spreading ocean ridges. Segments E2 and E9 have axial highs rising to 2·5 km depth, typical of inflated segments of fast-spreading ocean ridges. In view of the essentially constant spreading rate along the East Scotia Ridge, the excess magma production in segments E2 and E9 implies that these two segments are underlain by different, more readily melted mantle than the rest of the ridge (Livermore et al., 1997). Segments E2 and E1 have been described in detail by Livermore et al. (1997), using bathymetric, side-scan sonar and seismic data. Segment E1 is a wide, trough-like feature that increases in depth from 4·2 km in the south to 5·5 km at its northern end, where it joins the trench. It is separated from segment E2 by a non-transform offset at 55°45'S.

	SEGMENT E2

TOP ABSTRACT INTRODUCTION REGIONAL SETTING THE EAST SCOTIA RIDGE SEGMENT E2 RESULTS DISCUSSION REFERENCES

 
This segment is 70 km long, is robust, and is extending southward at the expense of segment E3 (Livermore et al., 1997). Segment E2 decreases in depth from 3·5 km at the segment tips to 2·6 km at the axial high summit. The north–central part of the segment is occupied by an area of positive topography at <3·1 km depth, which is formed by a constructive volcanic edifice (Fig. 2). This feature (Mermaid’s Purse: Livermore et al., 1997) stands some 0·6 km high, and has a north–south trending summit graben of 2 km width. Side-scan sonar images and Towed Ocean Bottom Instrument (TOBI) data show that the summit area is the site of the most recent lava extrusion (Livermore et al., 1997, 1999). Seismic reflection data indicate that the Mermaid’s Purse is underlain by a melt lens (Livermore et al., 1997). The melt lens is narrow (<1 km) in an east–west direction, some 20 km long and about 3 km below the sea floor, comparable with a seismically imaged melt lens beneath the back-arc Valu Fa Ridge, Lau Basin (Collier & Sinha, 1990). The absence of a rift valley, the presence of the Mermaid’s Purse constructional edifice and the presence of the melt lens are all features of mid-ocean spreading centres that are fast spreading [Batiza (1996) and references therein].

View larger version (108K): [in this window] [in a new window]   Fig. 2. MR1-derived bathymetry of segment E2 and the northern part of segment E3, showing wax core sample stations and dredge sites. Contours represent water depth in km. Warm colours represent shallower areas, blue colours deeper areas.

 

Previous geochemical work
No previous geochemical work has been carried out on segment E2. Most previous geochemical discussion on the East Scotia Ridge has been based on four dredge sites recovered by R.R.S. Shackleton in 1974, all south of segment E2 (Fig. 1b): dredge DR.20 (northern part of segment E3); dredge DR.22 (segment E5); dredges DR.23 and DR.24 (segment E9). Petrographic details, major and trace element abundances including rare earth elements (REE), Sr, Nd, Pb, O and C isotope ratios, and volatile abundances have been reported for these samples (Hawkesworth et al., 1977; Tarney et al., 1977; Saunders & Tarney, 1979; Muenow et al., 1980; Cohen & O’Nions, 1982; Saunders et al., 1982; Mattey et al., 1984; Newman & Stolper, 1996; Eiler et al., 1997). Pearce et al. (1995b) reanalysed a suite of samples from dredges DR.20, DR.22, DR.23 and DR.24 for major elements, trace elements by inductively coupled plasma mass spectrometry (ICP-MS), and Sr, Nd and Pb isotopes, and analysed previously unstudied samples from dredge sites DR.56, DR.57 and DR.60 collected by R.R.S. Shackleton in 1981 from east of the axis of segment E1. This dataset is used in this paper where comparisons are made between segment E2 and other parts of the East Scotia Ridge.

Most of the earlier studies concluded that the basalts of the East Scotia Ridge represent weak contamination of mid-ocean ridge basalt (MORB) source mantle by fluids and/or sediments derived from the subducting slab, and represent, in generalized composition, a position intermediate between MORB and island arc tholeiite. Nevertheless, it has also been recognized that some of the East Scotia Ridge basalts are enriched in ratios such as light REE/heavy REE (LREE/HREE) relative to both MORB and the South Sandwich Islands (Hawkesworth et al., 1977; Saunders & Tarney, 1979). This kind of relationship, which holds true for ratios such as Nb/Yb, which are thought not to be affected by fluxes from the slab, led Pearce et al. (1995b) to suggest that mantle plume material, possibly derived from the Bouvet hot spot, is present in the back-arc.

Sampling procedure
The sampling was carried out in 1996 as part of cruise JR12 of the British Antarctic Survey vessel R.R.S. James Clark Ross. Navigation data were recorded continuously by differential global positioning system (GPS), which allowed determination of the ship’s position at all times within 1 m. Most sites (WX.1–38) were sampled using the ‘wax coring’ or ‘rock coring’ technique in which wax-filled metal cups attached to a heavy metal frame are dropped at speed (110 m/min) onto the rock surface, and detached chips are retrieved in the wax. The retrieved material consists in most cases of glass chips derived from fractured lava flow surfaces. Samples collected by this method are referred to as ‘wax cores’ in this paper. The method allows an order of magnitude greater precision in sampling than standard dredging techniques (e.g. Reynolds et al., 1992), and hence sampling in much greater detail than most investigations of spreading centres hitherto (Batiza, 1996). The greatest error in determining the position of sample sites was caused by ocean currents, which induced deviations from the vertical in the cable from which the corer was suspended. Compound horizontal errors are therefore between 10 and 100 m.

Thirty wax cores yielding fresh lava glass were obtained at 2 km spacing from the axis of the median rift-like structure on segment E2 (Fig. 2), two wax cores were obtained from the northern part of segment E3, and one was obtained from segment E1. Two sites, DR.157 and DR.158, on the steep (>20%) lateral flanks of segment E2 axial high (Fig. 2) were dredged using standard methods. These dredge hauls consisted dominantly of vesicular, basaltic material, with abundant pillowed and ropy surfaces, typical of submarine lava flows. Several separate samples were analysed from each dredge haul.

Most samples are vesicular, but a group of samples from the southern end of the segment are non-vesicular. Vesicularity correlates inversely with water depth, a depth of between 3340 and 3450 m separating vesicular and non-vesicular types (see Fig. 6d, below). Nevertheless, the non-vesicular samples also tend to be more mafic and less enriched in incompatible elements, as we shall detail below.

Analytical techniques
Major elements on the wax core samples were determined on glass chips using a CAMECA SX100 wavelength-dispersive electron microprobe at the Department of Earth Sciences, Open University, UK. Operating conditions were as follows: accelerating voltage, 20 kV; beam current, 20 nA; beam diameter, 20 µm. Reported analyses are, in most cases, averages of three determinations. Major elements on glassy rinds of three dredge samples were analysed by standard X-ray fluorescence (XRF) procedures at the Department of Earth Sciences, University of Keele, UK (Floyd, 1985). Major elements on other dredge samples were measured by inductively coupled plasma optical emission spectrometry (ICP-OES) at Durham University. Trace elements on all samples were determined by ICP-MS plasmaquad at Durham University, using the methods of Pearce et al. (1995a). Sr, Nd and Pb isotopes were analysed on a Finnigan MAT 262 mass spectrometer at the NERC Isotope Geoscience Laboratories, Keyworth. Sr isotope analysis followed the methods of Pankhurst & Rapela (1995). ⁸⁷Sr/⁸⁶Sr ratios are normalized to ⁸⁶Sr/⁸⁸Sr = 0·1194. Twenty analyses of the NBS987 standard across the time of sample analysis gave 0·710212 ± 0·000007 (1). Results presented in Table 3 (below) are normalized to a value of 0·710240 for NBS987, to allow comparison with earlier datasets. Chemical methods for Nd analysis followed Pankhurst & Rapela (1995). ¹⁴³Nd/¹⁴⁴Nd ratios are normalized to ¹⁴⁶Nd/¹⁴⁴Nd = 0·7219. Twenty analyses of the La Jolla standard across the time of sample analysis gave 0·511869 ± 0·000008 (1). Long-term reproducibility of Sr and Nd isotope ratios in standard solutions is 15–20 ppm (1), but on rock standards this rises to 30–40 ppm. Pb analyses followed procedures summarized by Kempton et al. (1997). All whole-rock powders were leached in 6 M HCl before dissolution. Pb isotope ratios are corrected for fractionation using the standard values of Todt et al. (1984). Long-term reproducibility of Pb isotope ratios in the NBS981 solution standard is better than ±0·1% (1). Long-term reproducibility of the BHVO-1 rock standard is better than ±0·2% (1). Laboratory blanks for Sr, Nd and Pb at the time of analysis were better than 500, 250 and 150 pg, respectively.

View this table: [in this window] [in a new window]   Table 3: Sr, Nd and Pb isotopic data for samples from segments E2 and E3, East Scotia Ridge

 

	RESULTS

TOP ABSTRACT INTRODUCTION REGIONAL SETTING THE EAST SCOTIA RIDGE SEGMENT E2 RESULTS DISCUSSION REFERENCES

 
Major and trace element analyses on wax core samples are reported in Table 1, and dredge sites DR.157 and DR.158 are reported in Table 2. Sr, Nd and Pb isotope analyses are presented in Table 3.

View this table: [in this window] [in a new window]   Table 1: Major and trace element analyses of wax core samples from segments E1, E2 and E3 of the East Scotia Ridge

 

View this table: [in this window] [in a new window]   Table 2: Major and trace element analyses of dredge samples from segment E2 of the East Scotia Ridge

 

Major elements
The samples range from basalt to andesite, but nearly all samples are basaltic andesite (Fig. 3). The compositions are significantly higher in Si and lower in Mg than ‘normal’ for mid-ocean lavas (e.g. Sigurdsson, 1981). Although the wax core samples define a trend of decreasing Si with increasing Mg (Fig. 3), they do not constitute a suite related directly by fractional crystallization, as we shall detail below. Two of the samples are andesites, having 59·0 and 60·5 wt % SiO₂, similar to evolved lavas on the Valu Fa Ridge in the back-arc Lau Basin (Vallier et al., 1991). In the K₂O vs SiO₂ plot (Fig. 3b), the wax core samples form a scatter of low-K tholeiite and medium-K calc-alkaline compositions. The dredge samples are significantly enriched in K relative to the wax core samples at similar SiO₂ abundances.

View larger version (25K): [in this window] [in a new window]   Fig. 3. Major element variation diagrams for samples from segment E2. In (b), the line dividing medium-K (MK) and low-K (LK) rock series is shown. Generalized trends for low-K tholeiite, tholeiite and calc-alkaline series from the South Sandwich Islands (Pearce et al., 1995a) are indicated by dashed lines (LTS, TS and CS, respectively). All values in (a) and (b) are recalculated to volatile-free totals of 100 wt %.

 

In the wax core samples, major element compositions vary with bathymetry. Mg and Ca are positively correlated and Si, Fe, Mn, K, Na, Ti and P negatively correlated with water depth (Fig. 4). In terms of degree of fractional crystallization, the more evolved Si- and Fe-rich magmas erupted near the segment summit, and the more primitive, Mg-rich magmas erupted at the segment tips. This is the reverse relationship to that noted in some mid-ocean ridge segments, where the highest MgO lavas have been recovered from the centres of segments (Batiza, 1996; Lawson et al., 1996; Reynolds & Langmuir, 1997). Na_8·0 values {Na₂O abundances recalculated, using a fractional crystallization model, to an MgO content of 8 wt %: Na_8·0 = [Na₂O + 0·115(8 - MgO)]/[1 + 0·133(8 - MgO)]; Plank & Langmuir, 1992} are inversely related to degree of mantle melting. In segment E2, Na_8·0 values show little variation with bathymetry (Fig. 4f), indicating approximately constant degree of partial melting beneath different parts of the segment. Na_8·0 values for the segment lie within the ranges for the Lau and Marianas back-arc spreading centres (Pearce et al., 1994; Gribble et al., 1998) and within global arrays for mid-ocean ridges on Na_8·0–depth plots (Klein & Langmuir, 1987) (Fig. 5). Most of the samples have Na_8·0 values in the relatively narrow range 2·35–2·79. A subset of five samples, taken mainly near the segment summit, has lower Na_8·0 (2·13–2·25) and plots as a distinct group (hereinafter the ‘low-Na_8·0 group’) in Fig. 4f. In the main group of samples, Na_8·0 exhibits a slight negative correlation with water depth, the reverse of the global relationship for averaged compositions from mid-ocean ridges (Klein & Langmuir, 1987; Brodholt & Batiza, 1989) (Fig. 5).

View larger version (28K): [in this window] [in a new window]   Fig. 4. Plots of major elements and Na8·0 against water depth for wax core samples from segment E2. Na8·0 is calculated using the method of Plank & Langmuir (1992). Abundances are in wt %.

 

View larger version (20K): [in this window] [in a new window]   Fig. 5. Plot of Na8·0 vs water depth for wax cores from segment E2 (symbols as in Fig. 4), relative to MORB data from the oceans (after Klein & Langmuir, 1987).

 

Selected major element parameters and bathymetry are plotted against latitude in Fig. 6. This plot clearly shows that SiO₂ is negatively correlated and MgO positively correlated with water depth: compositions closest to basaltic MORB occupy the southern part of the segment, whereas samples from the summit region are more fractionated. Low-Na_8·0 samples are restricted to the summit and the axis south of the summit.

View larger version (22K): [in this window] [in a new window]   Fig. 6. Plots of selected major elements and water depth against latitude, segment E2. Only wax core samples from the segment axis are plotted. Na8·0 is calculated by the method of Plank & Langmuir (1992). In (a) and (b), dashed lines show generalized overall geochemical trends. The lateral extent of the seismically imaged melt lens in (d) is after Livermore et al. (1997). Depth of the melt lens is not to scale. Symbols: •, , vesicular samples; , non-vesicular samples.

 

There is a clear relationship between the geochemistry of the lavas and the position of the seismically imaged melt lens. High Si and low Mg abundances correlate even more closely with the melt lens position than with the bathymetry (Fig. 6). The two andesite samples come from lavas that directly overlie the northern tip of the melt lens. A group of less fractionated, but still low Mg lavas overlie its southern tip. The result is a W-shaped geochemical trend for MgO in Fig. 6b, with the troughs overlying the tips of the melt lens. We suggest that the excellent correlation between composition and position of the melt lens is evidence that the fractional crystallization of magma required to generate the evolved compositions near the segment summit took place within the imaged magma chamber.

Trace elements
All samples from segment E2 are LREE enriched, with a total range in La_N/Lu_N of 1·00–5·11. Variations in chondrite-normalized REE patterns are shown in Fig. 7. The mafic wax core samples define a band of REE patterns in the figure. The andesite wax core samples have patterns that are parallel to this band, but at higher total REE abundances, consistent with having been generated by fractional crystallization of mafic magmas (Fig. 7a). Most of the dredge samples have REE patterns similar to those of the mafic wax core samples (Fig. 7a), although two (DR.157.1 and DR.157.2) are significantly more enriched in LREE. Wax core samples from the tips of segment E2 generally have flatter REE patterns and lower La/Lu than those from its centre. Sample WX.17 from the south of E2 is similar to sample WX.15 from segment E3, and an adjacent sample WX.14 is even more LREE depleted, having an N-MORB-like REE pattern. Sample WX.35 from the north of segment E2 has a similar, flat pattern.

View larger version (50K): [in this window] [in a new window]   Fig. 7. Chondrite-normalized REE plots for selected samples from segment E2 and the northern part of segment E3, with locations marked on sketch bathymetric map (b). In (a), selected dredge samples are compared with andesites and the range for mafic wax core samples from segment E2 (shaded field). In (c)–(h), clusters of samples from various parts of the segments are plotted.

 

Figure 8 illustrates MORB-normalized multi-element patterns for segment E2 and the northern tip of segment E3 compared with E-MORB, an arc basalt from the South Sandwich Islands and a basalt from Bouvet Island. Sample WX.14 (Fig. 8f) from segment E3 has trace element abundances closest to N-MORB, from which it is significantly different only by minor enrichment in Cs, Rb, Ba and Pb. The other samples from the northern and southern tips of segment E2 have MORB-normalized patterns similar to that of E-MORB, with an increase in normalized values toward the left of the plots, peaking at 10 times MORB. Samples from the central part of the segment are more enriched in the elements Cs, Rb and Ba, and are more variable in composition. The dredge samples show a particularly large scatter, some having patterns similar to that of the South Sandwich Islands basalt, and others having similar patterns to the Bouvet Island basalt. The large range (from positive to negative) in Nb–Ta anomalies in the dredge samples is well illustrated by the figure. The wax core samples may have no Nb–Ta anomalies, small negative ones (e.g. in the low-Na_8·0 group), or small positive ones (sample WX.37) (Fig. 8c).

View larger version (44K): [in this window] [in a new window]   Fig. 8. N-MORB normalized plots for segment E2 and the northern tip of segment E3, and comparative samples. Normalizing values are after Sun & McDonough (1989). The element order is essentially one of increasing incompatibility in spinel lherzolite from right to left. (a) E-MORB from Sun & McDonough (1989), Bouvet Island basalt sample 1972.0-148 (1) from Weaver et al. (1987), and South Sandwich Islands tholeiitic volcanic arc basalt sample SSS.5.4 from Pearce et al. (1995a).

 

Selected incompatible trace element ratios for wax core samples from segment E2 are plotted by latitude in Fig. 9. There is a correlation between Nb/Zr, Nb/Yb and La/Yb with bathymetry. Compositions closest to N-MORB are from samples at the segment tips, and ratios of the more to less incompatible elements increase toward the topographic summit.

View larger version (28K): [in this window] [in a new window]   Fig. 9. Plots of selected trace element ratios and water depth against latitude, segment E2. Only wax core samples from the segment axis are plotted. The lateral extent of the seismically imaged melt lens is after Livermore et al. (1997). Depth of the melt lens is not to scale. Symbols as in Fig. 4.

 

Radiogenic isotopes
The variations in Nd, Sr and Pb isotopes from segment E2 are much greater than analytical error (Fig. 10). All segment E2 samples have lower ¹⁴³Nd/¹⁴⁴Nd and higher ⁸⁷Sr/⁸⁶Sr, ²⁰⁶Pb/²⁰⁴Pb, ²⁰⁷Pb/²⁰⁴Pb and ²⁰⁸Pb/²⁰⁴Pb ratios than samples from segment E3 (Fig. 10; Table 3). Three dredge samples from segment E2 have even lower ¹⁴³Nd/¹⁴⁴Nd and higher ⁸⁷Sr/⁸⁶Sr, ²⁰⁶Pb/²⁰⁴Pb, ²⁰⁷Pb/²⁰⁴Pb and ²⁰⁸Pb/²⁰⁴Pb ratios than the E2 wax core samples (Table 3). There are strong correlations between radiogenic isotope ratios and trace element compositions in these segments. Figure 10 shows isotopic and trace element variations in wax cores from segment E2 and dredge and wax core samples from the northern part of segment E3. ¹⁴³Nd/¹⁴⁴Nd and ⁸⁷Sr/⁸⁶Sr show a strong negative correlation. Although it might be argued that the high ⁸⁷Sr/⁸⁶Sr ratios from segment E2 are a result of assimilation of hydrothermally altered crust by magmas, this is unlikely to be the case for Nd. We propose that the range in isotope ratios in Fig. 10a is the result of variations in the mantle sources underlying the segments.

View larger version (20K): [in this window] [in a new window]   Fig. 10. Plots of co-variations in Nd, Sr and Pb isotope ratios and trace element ratios for wax core samples from segments E2 and E3, and dredge samples from dredge DR.20, segment E3. In (a), error bars are shown for 143Nd/144Nd ratios; errors in 87Sr/86Sr are smaller than the width of the points in the diagram, as are errors in 206Pb/204Pb in (d). In all plots, the three segment E3 samples are encircled. Wax core data from both segments are from this paper; data for the two dredge samples, DR.20.36 and DR.20.42 (segment E3) are from J. A. Pearce et al. (unpublished data, 1998).

 

	DISCUSSION

TOP ABSTRACT INTRODUCTION REGIONAL SETTING THE EAST SCOTIA RIDGE SEGMENT E2 RESULTS DISCUSSION REFERENCES

 
Lateral migration of magma
Several workers have proposed on volcanological, seismic and petrological grounds that magma of mid-ocean ridges is commonly transported laterally in dykes (e.g. Michael et al., 1989; Embley & Chadwick, 1994; Fox et al., 1995; Batiza, 1996). Such proposals were encouraged by descriptions of lateral dyke injection events in rift zones in Hawaii and Iceland (Fiske & Jackson, 1972; Sigurdsson & Sparks, 1978). In the case of E2, it is evident from Figs 6, 7 and 9 that lateral transport of magma cannot explain the overall chemical variations in the segment. The summit lavas are distinct from those from the axial flanks and segment tips in terms of major elements, and trace element ratios such as Nb/Yb and La/Yb, which measure mantle melting processes. Moreover, the samples from the segment tips are distinct from axial flank samples, implying that the segment tips are tapping different magmas from the rest of the segment. There are several pieces of geochemical evidence that nevertheless permit a relatively minor role for lateral magma transport:

1. although compositional changes with latitude are generally progressive, there is some scatter. This could be explained by flow of lavas away from eruptive vents or lateral flow in dykes, among other processes.
   
2. The part of the segment between 56°13'S and 56°26'S, a distance of some 25 km, has lavas with similar major and trace element compositions. This produces the ‘flat’ segment on the plots of La_N/Yb_N and Nb/Yb vs latitude (Fig. 9). This part of the southern axial flank could have been fed by lateral flow along dykes.
   
3. The low-Na_8·0 samples are distributed across a north–south distance of 16 km. If they were derived from a single source beneath the segment summit, it is possible that magmas flowed in dykes for lateral distances of at least 16 km.
   There is no evidence that lateral migration of magma along dykes exceeded 25 km anywhere in the segment. These features do not negate the conclusion that the overall geochemical features of the segment are not a result of lateral magma migration.
   

Magma sources and mantle melting processes
The roughly constant Na_8·0 values in most samples along the length of the segment (increasing slightly toward the summit; Fig. 4) indicate an approximately constant degree of partial melting with latitude. The presence of the low-Na_8·0 samples recovered from near the summit implies that a relatively high degree of mantle melting took place locally. The low-Na_8·0 samples also have lower total REE than adjacent samples (Fig. 7): low-Na_8·0 sample WX.7 is the only sample having HREE <10x chondrite. The relative depletion of REE in these samples is consistent with generation by a relatively high degree of mantle melting. Conversely, total REE abundances are high in many other samples near the centre of the segment, notably in the andesites, a result of the large degree of fractional crystallization of magmas associated with the axial high and melt lens. Inasmuch as fractional crystallization in the segment clearly did not significantly alter ratios among incompatible elements, and the degree of partial melting as measured by Na_8·0 was roughly constant (excluding low-Na_8·0 samples), the increases in Nb/Yb and La/Yb toward the centre of the segment (Fig. 9) imply a systematic change in the composition of the magma supply between the segment tips and the summit of E2. Although the increase in La/Yb may be modelled as a result of addition of a subduction component to the mantle source, Nb/Yb cannot be so modelled (e.g. Pearce et al., 1995a; Elliott et al., 1997; Woodhead et al., 1998).

It could be argued that the progressive elemental changes in the segment are the result of different partial melting behaviour in the mantle beneath the segment from that beneath the segment tips, for example by dynamic melting processes, explained inter alia by Elliott et al. (1991), McKenzie & O’Nions (1991) and Devey et al. (1994), and invoked to explain trace element variations in the South Sandwich Islands by Pearce et al. (1995a). Nevertheless, dynamic melting of a single source cannot explain the radiogenic isotope variations, which require a different, more enriched source beneath the summit from that beneath the tips. La_N/Yb_N and Nb/Yb are both strongly negatively correlated with ¹⁴³Nd/¹⁴⁴Nd (Fig. 10b and c), which demonstrates that the increase in these ratios toward the summit of segment E2 is related to differences in the underlying mantle. The negative correlation of ¹⁴³Nd/¹⁴⁴Nd with Nb/Yb demonstrates that this is not a result of addition of a subducted component that might have been responsible for the increase in La/Yb. The basalts from the segment summit have relatively high Nb/Yb, La/Yb and ⁸⁷Sr/⁸⁶Sr, and low ¹⁴³Nd/¹⁴⁴Nd, which are characteristics of melts of mantle plume-derived melts. The high ²⁰⁶Pb/²⁰⁴Pb ratio of one dredge sample from segment E2 (Table 3; Fig. 12, below), discussed below, supports a mantle plume-derived origin for some magmas erupted near the summit. We suggest that plume-related mantle underlies the central part of segment E2, but that ambient MORB-source mantle underlies the segment tips and segment E3.

View larger version (32K): [in this window] [in a new window]   Fig. 12. Plot of 207Pb/204Pb vs 206Pb/204Pb. Data sources as for Fig. 11 and: Bouvet Island, Sun (1980); Atlantic MORB, Ito et al. (1987); Southwest Indian Ridge, Mahoney et al. (1989); South American–Antarctic Ridge, Kurz et al. (1998); South Atlantic sediments subducting at the Sandwich Trench, Barreiro (1983); East Scotia Ridge (E3–E9), Cohen & O’Nions (1982). S, W, A and P are the approximate positions of putative end-members mantle wedge, subducted sediment, altered basaltic subducting slab and mantle plume, respectively. South Sandwich Islands are covered by data sources as for Fig. 11.

 

View larger version (41K): [in this window] [in a new window]   Fig. 11. Plots of Ba/Yb and Th/Yb vs Nb/Yb comparing segment E2 samples with those of the wider region. The shaded field defines the regional upper mantle where it is unaffected by the Sandwich subduction system, and includes average N-MORB (Sun & McDonough, 1989), dredge samples from the South American–Antarctic Ridge (le Roex et al., 1985), and hawaiites from Bouvet Island (le Roex & Erlank, 1982). The upper mantle beneath the South American–Antarctic Ridge is thought to be variably contaminated by mantle from the Bouvet plume, which has flowed westward relative to the plume and the South American and Antarctic plates. South Sandwich Islands data are from Pearce et al. (1995a). Dredge and wax core data from the East Scotia Ridge, segments E1, E3, E5, E7 and E9 are from J. A. Pearce et al. (unpublished data, 1998), this paper and P. T. Leat (unpublished data, 1998). The wax core and dredge data from segment E2 are from this paper. Vector 1 represents addition of plume-derived mantle to the source. Vector 2 represents addition of fluid-dominated subduction component represented by sample DR.157.3.

 
The ratios of Ba/Nb, La/Nb and Th/Nb examine the relative behaviour of pairs of elements that are little fractionated by partial melting but may be significantly fractionated by subduction processes. All these ratios for most samples remain approximately constant along the segment as would be expected were partial melting the sole cause of the variation (Fig. 9). However, a small subset of samples have significantly higher than predicted Ba/Nb and, to a lesser extent, La/Nb and Th/Nb. This subset includes most of the low-Na_8·0 group. Thus it is likely that the low-Na_8·0 samples had a subduction component in their source. Moreover, if their low Na_8·0 can be explained by a relatively high degree of partial melting, it is likely that the fluids that stimulated the partial melting were also transporting the Ba. Experimental results indicate that Ba is significantly more likely to be transported by aqueous fluids than are La and Th, which are more likely to be added from sediment either by bulk addition or by melts (Brenan et al., 1995; Keppler, 1996). The low-Na_8·0 group has ¹⁴³Nd/¹⁴⁴Nd and ⁸⁷Sr/⁸⁶Sr ratios that overlap with the rest of the wax core samples, but extend to lower ¹⁴³Nd/¹⁴⁴Nd and higher ⁸⁷Sr/⁸⁶Sr (Fig. 10). The low-Na_8·0 samples have higher ²⁰⁶Pb/²⁰⁴Pb, ²⁰⁷Pb/²⁰⁴Pb and ²⁰⁸Pb/²⁰⁴Pb than the other wax core samples (Table 3, Fig. 10d). There is a weak positive correlation of ⁸⁷Sr/⁸⁶Sr with Nb/Yb, but a strong positive correlation of ⁸⁷Sr/⁸⁶Sr with Ba/Nb (Fig. 10e). These relationships indicate that the fluid that carried Ba also carried Pb and Sr, which is consistent with experimental results (Keppler, 1996).

Comparison with regional chemical reservoirs
Trace element compositions from the Sandwich subduction system, the South American–Antarctica Ridge and Bouvet Island are compared in Fig. 11. In Fig. 11a, N-MORB, the South American–Antarctic Ridge and Bouvet Island define a narrow, ‘intraplate’ zone of weakly increasing Ba/Nb as Nb/Yb and Ba/Yb strongly increase. The spread of data from the South American–Antarctic Ridge along this trend was attributed by le Roex et al. (1985) to lateral flow from the Bouvet mantle plume beneath the ridge. The South Sandwich Islands have much higher Ba/Nb than this MORB–Bouvet trend, because of transport of Ba into the sub-arc mantle by the subduction process (Pearce et al., 1995a). Excluding the segment E2 data, the rest of the East Scotia Ridge forms a triangular array between compositions close to N-MORB, the South Sandwich Islands and Bouvet Island, and is interpreted broadly to represent variable subduction modification of mantle with a similar range of compositions to that beneath the South American–Antarctic Ridge (Pearce et al., 1995b). A similar diagram may be constructed substituting Th for Ba, although Th data are not available for the South American–Antarctic Ridge (Fig. 11b). A significant difference between the two plots is that whereas Ba/Nb decreases with increasing Ba/Yb in the data from the South Sandwich Islands, Th/Nb increases with increasing Th/Yb. The different behaviour is interpreted to be a result of enrichment of the mantle source of the arc magmas by two distinct components (see next section). The E2 wax core samples form arrays approximately parallel to the ‘intraplate’ trends in both plots (vector 1, Fig. 11). However, most of the wax core samples have similar Th/Nb, but their Ba/Nb increases with increasing Ba/Yb. These trends are different from those in the South Sandwich arc, indicating that different processes controlled the trends in the arc and back-arc environments. The wax core samples extend further toward the Bouvet component in both plots than any other segment of the East Scotia Ridge. Most of the dredge samples from segment E2 have higher Ba/Yb and Th/Yb than any of the wax core samples. Two samples (DR.157.1-2 and DR.157.2) have almost identical compositions (in these diagrams) to Bouvet Island hawaiites. Three samples (DR.158.4, DR.158.5 and DR.158.6) are similar to the South Sandwich Islands in having high Ba/Nb and Th/Nb ratios, but have much higher Ba/Yb than any known South Sandwich Islands sample. The remaining dredge samples and the low-Na_8·0 wax core samples are displaced from the main group of E2 wax core samples toward these three high Ba/Nb dredge samples, and dredge sample DR.157.3 (vector 2, Fig. 11), rather than toward the South Sandwich Islands samples. Whereas all South Sandwich Island samples are depleted in high field strength elements (HFSE) (e.g. have Nb/Yb < N-MORB, Fig. 11), none of the segment E2 samples are so depleted. Because such depletion is thought to be a characteristic of sub-arc mantle (Pearce et al., 1995a), the inference is that no sub-arc mantle is involved in magmagenesis in segment E2.

The new lead isotopic data for most segment E2 samples are similar to those previously obtained from the East Scotia Ridge (Fig. 12). Pearce et al. (1995a) suggested that basalts from the ridge and the South Sandwich Islands contain Pb from three end-member reservoirs: subducted sediment (S), subducted altered basaltic oceanic crust (A) and mantle wedge (W). Inspection of Fig. 12 indicates that end-member A can be typical Atlantic MORB, and is very similar to the composition of the crust now forming at the South American–Antarctic Ridge (Kurz et al., 1998) and presumably also being subducted beneath the South Sandwich arc. However, end-member W has lower ²⁰⁶Pb/²⁰⁴Pb than Atlantic MORB, and is close to ‘low 206’ MORB types from the Southwest Indian Ridge (Mahoney et al., 1989). The ‘low 206’ mantle reservoir is thought to have originated in the Indian Ocean during Gondwana break-up, arguably as a result of contamination of the upper mantle by material removed from continental lithospheric mantle underlying the supercontinent [Mahoney et al. (1998) and references therein]. The East Scotia Ridge samples have lower ²⁰⁶Pb/²⁰⁴Pb than the South Sandwich Islands at similar ²⁰⁷Pb/²⁰⁴Pb (Fig. 12), and may contain little lead from end-member A. Dredge sample DR.157.1 has much higher ²⁰⁶Pb/²⁰⁴Pb than other East Scotia Ridge samples, and indicates the presence of another mantle end-member within the back-arc, herein designated P (plume), and similar to Bouvet Island compositions, which are mantle plume-derived.

We propose that these arguments support the following interpretations:

1. the tips of segment E2 tapped mantle similar to that which underlies segment E3. The magmas produced were close to Indian Ocean N-MORB in composition.
   
2. E2 samples are not depleted in HFSE (e.g. have higher Nb/Yb than N-MORB) whereas all South Sandwich arc lavas are depleted. This implies that depleted, sub-arc mantle is not a source for segment E2 magmatism.
   
3. On the lateral flanks of the summit, lavas that are strongly enriched in the most incompatible elements were erupted (dredge samples). Two of these are similar to Bouvet Island mafic lavas, having received almost no chemical influence from the subduction zone. The rest of the dredge samples were clearly modified by components derived from the slab, having high Ba/Nb and Th/Nb ratios.
   
4. The wax core samples from the segment summit and flanks are intermediate in composition between the dredge samples and the MORB-like end-member (Fig. 11). The low-Na_8·0 samples are those that were most clearly modified by slab-derived components.
   

The dredge samples might have been derived by partial melting of sources situated either in lithospheric mantle (and therefore possibly formed in different tectonic environments from that of the present back-arc), or in the back-arc asthenosphere. In the case of the two Bouvet-like samples we suggest that it is unlikely that their source could have been situated in lithospheric mantle that remained chemically almost unaffected by the subduction system, and it is more likely that the mantle source was asthenosphere derived from Bouvet (or a chemically similar mantle plume). The other dredge samples, which were derived from strongly subduction-modified sources, could have been derived from either lithospheric or asthenospheric sources. Nevertheless, the similarity in radiogenic isotopes (e.g. Fig. 12) of these samples with other South Sandwich arc and back-arc samples indicates that any lithospheric components were recently generated, and not derived from significantly older lithosphere.

Composition of the slab-derived component
In Fig. 13, it is clear that there is no simple mixing trend in the wax core samples between N-MORB and the Bouvet-like plume component P on the one hand (both of which have low Ba/Th and Th/Nb ratios), and slab-modified compositions as defined by the South Sandwich Islands on the other. There is also a negative correlation between Ba/Th and Th/Nb in the South Sandwich Islands (Fig. 13). Both these ratios are higher in the South Sandwich Islands lavas than in MORB, and so both are recording the chemical modification of mantle by subducting slab. Their negative correlation indicates that high Ba/Th (trend A) is recording a different enrichment process from that recorded by high Th/Nb (trend B). We suggest that trend A is equivalent to addition of the high B, B/Be, Cs/Th, U/Th, Ba/Nb, Ba/Th, ²³⁸U/²³²Th component variously observed in the Kurile, Marianas, New Britain and Lesser Antilles arcs, whereas trend B is equivalent to the high La/Yb, Th/Nb, La/Sm, ¹⁰Be component observed in the same arcs (Ryan et al., 1995; Turner et al., 1996; Elliott et al., 1997; Woodhead et al., 1998). These two components have been interpreted to be an aqueous slab-derived fluid (trend A equivalents) and sediment (trend B equivalents; possibly a fused silicate liquid) by those workers in the Pacific arcs and Lesser Antilles arc (Ryan et al., 1995; Turner et al., 1996; Elliott et al., 1997; Woodhead et al., 1998). The steep trend formed by segment E2 wax cores in Fig. 13 lies along trend A, and indicates that slab-derived fluid was their dominant influence. Sample DR.157.3 (Ba/Th = 198) lies at the high Ba/Th end of this trend, and is a more plausible end-member for the dominant subduction component in the wax core samples than samples DR.158.4, -5 and -6, which lie close to trend B (dominated by sediment addition).

View larger version (27K): [in this window] [in a new window]   Fig. 13. Plot of Ba/Th vs Th/Nb. Data sources are as in Fig. 11, with additional data for the South Sandwich Islands from P. T. Leat (unpublished data, 1999). Trends A and B are putative enrichments interpreted to have been caused by fluid addition and sediment addition, respectively, from the subducting slab.

 

Magma mixing processes
End-member compositions were identified at the segment tips (where compositions close to MORB were erupted), and on the lateral flanks of the topographical summit of the segment (dredge samples). All other wax core samples appear to be intermediate in composition between these end-members. Our preferred model is that the majority of the compositions represented by the wax cores were generated by magma mixing processes. Wax core samples from the segment axial flanks contain high contributions from the MORB-source mantle, whereas those from the summit contain high contributions from the plume mantle. One explanation would be that plume mantle underlies the summit of the segment, and MORB-source mantle underlies the segment tips. Alternatively, the mantle below the summit might be heterogeneous, containing blebs of material enriched in the plume component within MORB-source mantle. Low-degree partial melting would generate trace-element-enriched magmas like the dredge samples from the blebs before the surrounding mantle generated MORB-like melts. Plumbing systems must have been sufficiently voluminous to allow mixing of the different magmas to generate the systematic trace element variations along the segment (Fig. 9). The magma flux can be assumed to be greatest along the axis of the topographic summit as the edifice itself demonstrates excess magmatism and there is a seismically imaged melt lens. Even here, however, magma homogenization was inefficient, allowing eruption of the distinctive low-Na_8·0 lavas. Magma flux is likely to be much less along the lateral flanks of the edifice where the dredge samples were erupted. We suggest that the absence of well-developed magmatic pathways here allowed small-volume magma batches to reach the surface without mixing with other magmas, to explain the distinctive, relatively extreme, end-member compositions of the dredge samples.

Tectonic models
There are two competing models for the interaction of the mantle with the subducting slab in the Sandwich arc. Alvarez (1982) thought that eastward flow of mantle is pushing the subducting slab towards the east, thus generating the relative eastward migration of the trench–arc system. Livermore et al. (1997) suggested that mantle is flowing from east to west around the lateral edge at the north end of the slab. This implies that the eastward migration of the trench is a consequence of slab rollback, which is pulling the Sandwich plate eastward. Our results for segment E2 support the interpretations of Livermore et al. (1997), indicating that rollback at the Sandwich Trench is taking place despite the slab having to migrate through a ‘head-wind’ of mantle flow, or mantle that is essentially static. The identification of the mantle plume end-member in the segment supports previous ideas that material from the Bouvet mantle plume is migrating westward beneath the South American–Antarctic Ridge (le Roex et al., 1985; Kurz et al., 1998). Similar flow of plume-derived mantle into a back-arc round a lateral slab edge is believed to be taking place at the north end of the Lau Basin (Turner & Hawkesworth, 1998). As mantle flows westward around the north edge of the South American slab (Fig. 14), selective contamination is likely to take place by fluids and sediments from the exposed edge of the slab. We believe that the heterogeneity of the dredge samples is a record of such contamination, in which some mantle patches are more strongly modified than others, and in which there are variable components derived from the slab.

View larger version (28K): [in this window] [in a new window]   Fig. 14. Sketch showing the relationship of the northern part of the Sandwich subduction system to mantle flow, as seen from the latitude of Zavodovski. The Scotia and Sandwich plates are shown by dashed lines. Mantle originating east of the subducting slab flows through a wedge-shaped gap opening as the South American plate is tearing in a scissor-like motion. This mantle enters the back-arc directly beneath segment E2, and is thought to be the main mantle source of that segment.

 

The progressively opening wedge-shaped gap is likely to direct mantle flow into the back-arc. Flow from beneath the relatively thick South American plate to beneath the thin lithosphere of the back-arc spreading centre may trigger decompression melting in excess of that generated by extension alone. This may explain the excess magmatism in segment E2 compared with the rest of the East Scotia Ridge. Nevertheless, the low-Na_8·0 samples, which have high Ba contents suggestive of slab-derived fluid in their sources, provide good evidence that hydration of mantle as it passed over the lateral slab edge triggered excess melting, as postulated by Livermore et al. (1997). The flow of mantle from the slab gap to segment E2 furthermore explains the absence, noted above, of depleted sub-arc mantle beneath the segment.

	ACKNOWLEDGEMENTS

 
We thank Richard Dingle, Anya Reading, Phil Jones, Peter Morris, Julie Ferris and Pat Cooper (British Antarctic Survey) for assistance during sampling operations. We are grateful to Chris Ottley (Durham) for assistance with the ICP-MS analyses, Kay Green (Open University) for her excellent work in making the probe mounts, and Andy Tindle and Phil Potts (Open University) for assistance with the microprobe analyses. R. Larter, J. Reynolds, A. Saunders and S. Turner provided constructive comments on the paper.

	FOOTNOTES

 
^*Corresponding author. Telephone: +44-1223 221400. Fax: +44-1223 362616. e-mail: p.leat{at}bas.ac.uk

Present address: Department of Earth Sciences, Cardiff University, PO Box 914, Cardiff CF10 3YE, UK

	REFERENCES

TOP ABSTRACT INTRODUCTION REGIONAL SETTING THE EAST SCOTIA RIDGE SEGMENT E2 RESULTS DISCUSSION REFERENCES

 
Alvarez, W. (1982). Geological evidence for the geographical pattern of mantle return flow and the driving mechanism of plate tectonics. Journal of Geophysical Research 87, 6697–6710.

Baker, P. E. (1968). Comparative volcanology and petrology of the Atlantic islands arcs. Bulletin Volcanologique 32, 189–206.

Baker, P. E. (1990). South Sandwich Islands. In: LeMasurier, W. E. & Thomson, J. W. (eds) Volcanoes of the Antarctic Plate and Southern Oceans. American Geophysical Union, Antarctic Research Series 48, 361–395.

Baker, P. E., Buckley, F. & Rex, D. C. (1977). Cenozoic volcanism in the Antarctic. Philosophical Transactions of the Royal Society of London, Series B 279, 131–142.

Barker, P. F. (1995). Tectonic framework of the East Scotia Sea. In: Taylor, B. (ed.) Backarc Basins: Tectonics and Magmatism. New York: Plenum, pp. 281–314.

Barker, P. F. & Hill, I. A. (1981). Back-arc extension in the Scotia Sea. Philosophical Transactions of the Royal Society of London, Series A 300, 249–262.

Barker, P. F. & Lawver, L. A. (1988). South American–Antarctic plate motion over the past 50 Myr, and the evolution of the South American–Antarctic ridge. Geophysical Journal 94, 377–386.[Web of Science]

Barreiro, B. (1983). Lead isotopic compositions of South Sandwich Island volcanic rocks and their bearing on magmagenesis in intra-oceanic island arcs. Geochimica et Cosmochimica Acta 47, 817–822.

Batiza, R. (1996). Magmatic segmentation of mid-ocean ridges: a review. In: MacLeod, C. J., Tyler, P. A. & Walker, C. L. (eds) Tectonic, Magmatic, Hydrothermal and Biological Segmentation of Mid-Ocean Ridges. Geological Society, London, Special Publication 118, 103–130.

Brenan, J. M., Shaw, H. F., Ryerson, F. J. & Phinney, D. L. (1995). Mineral–aqueous fluid partitioning of trace elements at 900°C and 2·0 GPa—constraints on the trace element chemistry of mantle and deep-crustal fluids. Geochimica et Cosmochimica Acta 59, 3331–3350.[Web of Science]

Brett, C. P. (1977). Seismicity of the South Sandwich Islands region. Geophysical Journal of the Royal Astronomical Society 51, 453–464.[Web of Science]

Brodholt, J. P. & Batiza, R. (1989). Global systematics of unaveraged mid-ocean ridge basalt compositions: comment on ‘Global correlations of ocean ridge basalt chemistry with axial depth and crustal thickness’ by E. M. Klein and C. H. Langmuir. Journal of Geophysical Research 94, 4231–4239.

Cohen, R. S. & O’Nions, R. K. (1982). Identification of recycled continental material in the mantle from Sr, Nd and Pb isotope investigations. Earth and Planetary Science Letters 61, 73–84.

Collier, J. & Sinha, M. (1990). Seismic images of a magma chamber beneath the Lau Basin back-arc spreading centre. Nature 346, 646–648.

Devey, C. W., Garbe-Schönberg, C.-D., Stoffers, P., Chuvel, C. & Mertz, D. F. (1994). Geochemical effects of dynamic melting beneath ridges: reconciling major and trace element variations in Kolbeinsey (and global) mid-ocean ridge basalt. Journal of Geophysical Research 99, 9077–9095.

Douglass, J., Schilling, J.-G. & Fontignie, D. (1999). Plume–ridge interactions of the Discovery and Shona mantle plumes with the southern Mid-Atlantic Ridge. Journal of Geophysical Research 104, 2941–2962.

Eiler, J. M., Kessel, R., Farley, K. A., Stolper, E. M. & Valley, J. W. (1997). Oxygen isotope constraints on the sources of oceanic arc volcanism: the Mariana and South Sandwich arcs. EOS Transactions, American Geophysical Union 78, F825.

Elliott, T. R., Hawkesworth, C. J. & Grönvold, K. (1991). Dynamic melting of the Iceland plume. Nature 351, 201–206.

Elliott, T., Plank, T., Zindler, A., White, W. & Bourdon, B. (1997). Element transport from slab to volcanic front at the Mariana arc. Journal of Geophysical Research 102, 14991–15019.

Embley, R. W. & Chadwick, W. W., Jr (1994). Volcanic and hydrothermal processes associated with a recent phase of seafloor spreading at the northern Cleft segment: Juan de Fuca ridge. Journal of Geophysical Research 99, 4741–4760.

Fiske, R. S. & Jackson, E. D. (1972). Orientation and growth of Hawaii volcanic rifts: the effect of regional structure and gravitational stresses. Proceedings of the Royal Society of London, Series A 329, 299–326.

Floyd, P. A. (1985). Petrology and geochemistry of intraplate sheet-flow basalts, Nauru Basin, Deep Sea Drilling Project leg 89. In: Moberley, R., Schlanger, S. O. et al. (eds) Initial Reports of the Deep Sea Drilling Project 89. Washington, DC: US Government Printing Office, pp. 471–497.

Forsyth, D. W. (1975). Fault plane solutions and tectonics of the South Atlantic and Scotia Sea. Journal of Geophysical Research 80, 1429–1443.

Fox, C. G., Radford, W. E., Dziak, R. P., Lau, T.-K., Matsumoto, H. & Schreiner, A. E. (1995). Acoustic detection of a seafloor spreading episode on the Juan de Fuca ridge using military hydrophone arrays. Geophysical Research Letters 22, 131–134.

Gribble, R. F., Stern, R. J., Newman, S., Bloomer, S. H. & O’Hearn, T. (1998). Chemical and isotopic composition of lavas from the northern Mariana Trough: implications for magmagenesis in back-arc basins. Journal of Petrology 39, 125–154.

Hawkesworth, C. J., O’Nions, R. K., Pankhurst, R. J., Hamilton, P. J. & Evensen, N. M. (1977). A geochemical study of island arc and back-arc tholeiites from the Scotia Sea. Earth and Planetary Science Letters 36, 253–262.

Holdgate, M. W. & Baker, P. E. (1979). The South Sandwich Islands: I. General description. British Antarctic Survey Scientific Reports 91, 1–76.

Ito, E., White, W. & Göpel, C. (1987). The O, Sr, Nd and Pb isotope geochemistry of MORB. Chemical Geology 62, 157–176.[Web of Science]

Klein, E. M. & Langmuir, C. H. (1987). Global correlations of ocean ridge basalt chemistry with axial depth and crustal thickness. Journal of Geophysical Research 92, 8089–8115.

Kempton, P. D., Downes, H. & Embey-Isztin, A. (1997). Mafic granulite xenoliths in Neogene alkali basalts from the western Pannonian Basin: insights into the lower crust of a collapsed orogen. Journal of Petrology 38, 941–970.

Keppler, H. (1996) Constraints from partitioning experiments on the composition of subduction-zone fluids. Nature 380, 237–240.

Kurz, M. D., le Roex, A. P. & Dick, H. J. B. (1998). Isotope geochemistry of the enriched mantle near the Bouvet triple junction. Geochimica et Cosmochimica Acta 62, 841–852.

Larter, R. D., King, E. C., Leat, P. T., Reading, A. M., Smellie, J. L. & Smythe, D. K. (1998). South Sandwich slices reveal much about arc structure, geodynamics and composition. EOS Transactions, American Geophysical Union 79, 281–285.

Lawson, K., Searle, R. C., Pearce, J. A., Browning, P. & Kempton, P. (1996). Detailed volcanic geology of the MARNOK area, Mid-Atlantic Ridge north of Kane transform. In: MacLeod, C. J., Tyler, P. A. & Walker, C. L. (eds) Tectonic, Magmatic, Hydrothermal and Biological Segmentation of Mid-Ocean Ridges. Geological Society, London, Special Publication 118, 61–102.

le Roex, A. P. & Erlank, A. J. (1982). Quantitative evaluation of fractional crystallisation in Bouvet Island lavas. Journal of Volcanology and Geothermal Research 13, 309–338.

le Roex, A. P., Dick, H. J. B., Reid, A. M., Frey, F. A., Erlank, A. J. & Hart, S. R. (1985). Petrology and geochemistry of basalts from the American–Antarctic Ridge, Southern Ocean: implications for the westward influence of the Bouvet mantle plume. Contributions to Mineralogy and Petrology 90, 367–380.

Livermore, R., McAdoo, D. & Marks, K. (1994). Scotia Sea tectonics from high-resolution satellite gravity. Earth and Planetary Science Letters 123, 255–268.

Livermore, R. A., Larter, R. D., Cunningham, A. P., Vanneste, L., Hunter, R. J. and the JR09 team (1995). Hawaii-MR1 sonar survey of the East Scotia ridge. BRIDGE Newsletter 8, 51–53.

Livermore, R., Cunningham, A., Vanneste, L. & Larter, R. (1997). Subduction influence on magma supply at the East Scotia Ridge. Earth and Planetary Science Letters 150, 261–275.

Livermore, R., Bruguier, N., Cunningham, A., Domaschk, U., Eagles, G., Fretzdorff, S., German, C., Maldonado, A., Morris, P. and the JR39b team (1999). JR39b: deep towed sonar and seismic survey on the East Scotia Ridge. InterRidge News 8, 34–37.

Mahoney, J. J., Natland, J. H., White, W. M., Poreda, R., Bloomer, S. H., Fisher, R. L. & Baxter, A. N. (1989). Isotopic and geochemical provinces of the western Indian Ocean spreading centres. Journal of Geophysical Research 94, 4033–4052.

Mahoney, J. J., Frei, R., Tejada, M. L. G., Mo, X. X., Leat, P. T. & Nägler, T. F. (1998). Tracing the Indian Ocean mantle domain through time: isotopic results from old west Indian, east Tethyan, and South Pacific seafloor. Journal of Petrology 39, 1285–1306.

Mattey, D. P., Carr, R. H., Wright, I. P. & Pillinger, C. T. (1984). Carbon isotopes in submarine basalts. Earth and Planetary Science Letters 70, 196–206.

McKenzie, D. & O’Nions, R. K. (1991) Partial melt distributions from inversion of rare earth element concentrations. Journal of Petrology 32, 1021–1091.[Abstract/Free Full Text]

Michael, P. J., Chase, R. L. & Allan, J. F. (1989). Petrologic and geological variations along the southern Explora ridge, northeast Pacific Ocean. Journal of Geophysical Research 94, 13895–13918.

Muenow, D. W., Liu, N. W. K., Garcia, M. & Saunders, A. D. (1980). Volatiles in submarine volcanic rocks from the spreading centre axis of the East Scotia Sea back-arc basin. Earth and Planetary Science Letters 47, 272–278.

Newman, S. & Stolper, E. (1996). Volatiles in magmas from the East Scotia Sea, South Atlantic Ocean. JAG/RAS Meeting on Subduction Zone Structure, Dynamics and Magmatism, London, 1996. London: Joint Association for Geophysics, Geological Society.

Pankhurst, R. J. & Rapela, C. R. (1995). Production of Jurassic rhyolite by anatexis of the lower crust of Patagonia. Earth and Planetary Science Letters 134, 23–36.

Pearce, J. A., Ernewein, M., Bloomer, S. H., Parson, L. M., Murton, B. J. & Johnson, L. E. (1994). Geochemistry of Lau Basin volcanic rocks: influence of ridge segmentation and arc proximity. In: Smellie, J. L. (ed.) Volcanism Associated with Extension at Consuming Plate Margins. Geological Society, London, Special Publication 81, 53–75.

Pearce, J. A., Baker, P. E., Harvey, P. K. & Luff, I. W. (1995a). Geochemical evidence for subduction fluxes, mantle melting and fractional crystallization beneath the South Sandwich island arc. Journal of Petrology 36, 1073–1109.[Abstract/Free Full Text]

Pearce, J. A., Leat, P., Barker, P. F. & Millar, I. L. (1995b). Geochemical evidence for mantle dynamics and mantle melting beneath the Scotia Sea and surrounding region. EOS Transactions, American Geophysical Union 76, F542.

Pelayo, A. M. & Wiens, D. A. (1989). Seismotectonics and relative plate motions in the Scotia Sea region. Journal of Geophysical Research 94, 7293–7320.

Plank, T. & Langmuir, C. H. (1992). Effects of the melting regime on the composition of the oceanic crust. Journal of Geophysical Research 97, 19749–19770.

Reynolds, J. R. & Langmuir, C. H. (1997). Petrological systematics of the Mid-Atlantic Ridge south of Kane: implications for ocean crust formation. Journal of Geophysical Research 102, 14915–14946.

Reynolds, J. R., Langmuir, C. H., Bender, J. F., Kastens, K. A. & Ryan, B. F. (1992). Spatial and temporal variability in the geochemistry of basalts from the East Pacific Rise. Nature 359, 493–499.

Ryan, J. G., Morris, J., Tera, F., Leeman, W. P. & Tsvetkov, A. (1995). Cross-arc geochemical variations in the Kurile arc as a function of slab depth. Science 270, 625–627.[Abstract/Free Full Text]

Saunders, A. D. & Tarney, J. (1979). The geochemistry of basalts from a back-arc spreading centre in the east Scotia Sea. Geochimica et Cosmochimica Acta 43, 555–572.

Saunders, A. D., Tarney, J., Weaver, S. D. & Barker, P. F. (1982). Scotia Sea floor: geochemistry of basalts from the Drake Passage and South Sandwich spreading centres. In: Craddock, C. (ed.) Antarctic Geoscience. Madison, WI: University of Wisconsin Press, pp. 213–222.

Sigurdsson, H. (1981). First-order major-element variation in basalt glasses from the Mid-Atlantic Ridge: 29°N to 73°N. Journal of Geophysical Research 86, 9483–9502.

Sigurdsson, H. & Sparks, R. S. J. (1978). Rifting episode in North Iceland in 1874–1875 and the eruptions of Askja and Sveinagja. Bulletin Volcanologique 41, 149–167.

Sun, S.-S. (1980). Lead isotopic study of young volcanic rocks from mid-ocean ridges, ocean islands and island arcs. Philosophical Transactions of the Royal Society of London, Series A 297, 409–445.

Sun, S.-S. & McDonough, W. F. (1989). Chemical and isotopic systematics of oceanic basalts: implications for mantle compositions and processes. In: Saunders, A. D. & Norry, M. J. (eds) Magmatism in the Ocean Basins. Geological Society, London, Special Publication 42, 313–345.

Tarney, J., Saunders, A. D. & Weaver, S. D. (1977). Geochemistry of volcanic rocks from the island arcs and marginal basins of the Scotia Arc region. In: Talwani, M. & Pitman, W. C. (eds) Island Arcs, Deep Sea Trenches and Back-Arc Basins. Washington, DC: American Geophysical Union, pp. 367–377.

Todt, W., Cliff, R. A., Hanser, A. & Hoffmann, A. W. (1984). ²⁰²Pb + ²⁰⁵Pb double spike for lead isotope analysis. Terra Cognita 4, 209.

Turner, S. & Hawkesworth, C. (1998). Using geochemistry to map mantle flow beneath the Lau Basin. Geology 26, 1019–1022.[Abstract/Free Full Text]

Turner, S., Hawkesworth, C., van Calsteren, P., Heath, E., Macdonald, R. & Black, S. (1996). U-series isotopes and destructive plate margin magma genesis in the Lesser Antilles. Earth and Planetary Science Letters 142, 191–207.

Vallier, T. L., Jenner, G. A., Frey, F. A., Gill, J. B., Davis, A. S., Volpe, A. M., Hawkins, J. W., Morris, J. D., Carwood, P. A., Morton, J. L., Scholl, D. W., Rautenschlein, M., White, W. M., Williams, R. W., Stevenson, A. J. & White, L. D. (1991). Subalkaline andesite from Valu Fa Ridge, a back-arc spreading centre in southern Lau Basin: petrogenesis, comparative chemistry, and tectonic implications. Chemical Geology 91, 227–256.

Verwoerd, W. J. (1990). Bouvetøya. In: LeMasurier, W. E. & Thomson, J. W. (eds) Volcanoes of the Antarctic Plate and Southern Oceans. American Geophysical Union, Antarctic Research Series 48, 405–410.

Weaver, B. L., Wood, D. A., Tarney, J. & Joron, J. L. (1987). Geochemistry of ocean island basalts from the South Atlantic: Ascension, Bouvet, St. Helena, Gough and Tristan da Cunha. In: Fitton, J. G. & Upton, B. G. J. (eds) Alkaline Igneous Rocks. Geological Society, London, Special Publication 30, 253–267.

Woodhead, J. D., Eggins, S. M. & Johnson, R. W. (1998). Magma genesis in the New Britain island arc: further insights into melting and mass transfer processes. Journal of Petrology 39, 1644–1668.

CiteULike    Connotea    Del.icio.us    What's this?

This article has been cited by other articles:

				A. Zagorevski, N. Rogers, C.R. van Staal, V. McNicoll, C.J. Lissenberg, and P. Valverde-Vaquero Lower to Middle Ordovician evolution of peri-Laurentian arc and backarc complexes in Iapetus: Constraints from the Annieopsquotch accretionary tract, central Newfoundland Geological Society of America Bulletin, March 1, 2006; 118(3-4): 324 - 342. [Abstract] [Full Text] [PDF]

				J. A. PEARCE Mantle Preconditioning by Melt Extraction during Flow: Theory and Petrogenetic Implications J. Petrology, May 1, 2005; 46(5): 973 - 997. [Abstract] [Full Text] [PDF]

				D. P. West, D. P. WEST JR, R. A. COISH, and P. B. TOMASCAK Tectonic setting and regional correlation of Ordovician metavolcanic rocks of the Casco Bay Group, Maine: evidence from trace element and isotope geochemistry Geological Magazine, March 1, 2004; 141(2): 125 - 140. [Abstract] [Full Text] [PDF]

				J. Kim, R. Coish, M. Evans, and G. Dick Supra-subduction zone extensional magmatism in Vermont and adjacent Quebec: Implications for early Paleozoic Appalachian tectonics Geological Society of America Bulletin, December 1, 2003; 115(12): 1552 - 1569. [Abstract] [Full Text] [PDF]

				S. Fretzdorff, K.M. Haase, P.T. Leat, R.A. Livermore, C.-D. Garbe-Schonberg, J. Fietzke, and P. Stoffers 230Th-238U disequilibrium in East Scotia backarc basalts: Implications for slab contributions Geology, August 1, 2003; 31(8): 693 - 696. [Abstract] [Full Text] [PDF]

				P. T. Leat and R. D. Larter Intra-oceanic subduction systems: introduction Geological Society, London, Special Publications, January 1, 2003; 219(1): 1 - 17. [Abstract] [PDF]

				F. Martinez and B. Taylor Controls on back-arc crustal accretion: insights from the Lau, Manus and Mariana basins Geological Society, London, Special Publications, January 1, 2003; 219(1): 19 - 54. [Abstract] [PDF]

				P. T. Leat, J. L. Smellie, I. L. Millar, and R. D. Larter Magmatism in the South Sandwich arc Geological Society, London, Special Publications, January 1, 2003; 219(1): 285 - 313. [Abstract] [PDF]

				R. Livermore Back-arc spreading and mantle flow in the East Scotia Sea Geological Society, London, Special Publications, January 1, 2003; 219(1): 315 - 331. [Abstract] [PDF]

				D. Harrison, P. T. Leat, P. G. Burnard, G. Turner, S. Fretzdorff, and I. L. Millar Resolving mantle components in oceanic lavas from segment E2 of the East Scotia back-arc ridge, South Sandwich Islands Geological Society, London, Special Publications, January 1, 2003; 219(1): 333 - 344. [Abstract] [PDF]

				S. FRETZDORFF, R. A. LIVERMORE, C. W. DEVEY, P. T. LEAT, and P. STOFFERS Petrogenesis of the Back-arc East Scotia Ridge, South Atlantic Ocean J. Petrology, August 1, 2002; 43(8): 1435 - 1467. [Abstract] [Full Text] [PDF]

				A. C. KERR, R. V. WHITE, and A. D. SAUNDERS LIP Reading: Recognizing Oceanic Plateaux in the Geological Record J. Petrology, July 1, 2000; 41(7): 1041 - 1056. [Abstract] [Full Text] [PDF]

This Article Abstract FREE Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Similar articles in ISI Web of Science Alert me to new issues of the journal Add to My Personal Archive Download to citation manager Search for citing articles in: ISI Web of Science (38) Request Permissions Google Scholar Articles by LEAT, P. T. Articles by PEARCE, J. A. Search for Related Content GeoRef GeoRef Citation Social Bookmarking          What's this?

Online ISSN 1460-2415 - Print ISSN 0022-3530

Copyright © 2009 Oxford University Press

Oxford Journals Oxford University Press

• Site Map
• Privacy Policy
• Frequently Asked Questions

Other Oxford University Press sites: Oxford University Press Oxford Journals China Oxford Journals Japan American National Biography Booksellers' Information Service Children's Fiction and Poetry Children's Reference Corporate & Special Sales Dictionaries Dictionary of National Biography Digital Reference English Language Teaching Higher Education Textbooks Humanities International Education Unit Law Medicine Music Online Products Oxford English Dictionary Reference Rights and Permissions Science School Books Social Sciences Very Short Introductions World's Classics