Figure F1. Preliminary calculation of Mantle Bouger Anomaly (e.g., Lin et al., 1990) from shipboard gravity measurements taken in 1998 (Kelemen et al., 1998a; Matsumoto et al., 1998; Casey et al., 1998). Both figures use the same range of colors, representing slightly different values. A. Range is from about –35 (red) to +40 (pink) mGal north of the 15°20'N Fracture Zone. B. Range is from about –60 (red) to +45 (pink) Mgal south of the 15°20'N Fracture Zone. These data suggest that the magma-starved region with abundant peridotite outcrops from 14°40' to 15°40'N lies on the periphery of large magmatic segments centered at ~14° and 16°N, with thick igneous crust in the segment centers.

Figure F2. Bathymetry and geology from 14° to 16°N along the Mid-Atlantic Ridge. Depth range is ~5400 (violet) to 1600 (red) m. Sample lithologies are compiled from all known dredging and submersible results. A. North of the 15°20'N Fracture Zone. B. South of the 15°20'N Fracture Zone. C. View from the north of the "megamullion" dive site, where a large low-angle normal fault is exposed on the seafloor. Open circles = mantle peridotite, solid circles = basalt.

Figure F3. Geochemical data for samples from the Mid-Atlantic Ridge. A. Low Na₂O (upper panel = equator–70°N data; lower panel = 10°–20°N data) in basalts. B. High Cr/(Cr + Al) in spinel (lower panel) and shallow axial depth (upper panel) can all be taken to indicate high degrees of partial melting. C. High La/Sm (upper panel = data from the equator to 70°N; lower panel = data from 10° to 20°N). D. High ²⁰⁷Pb/²⁰⁴Pb (upper panel) and high ⁸⁷Sr/⁸⁶Sr (lower panel). C and D are indicative of long-term enrichment of the mantle source in incompatible trace elements. All of these characteristics are observed along the Mid-Atlantic Ridge just south of the 15°20'N Fracture Zone (FZ). Basalt data compiled by Xia et al. (1991, 1992, unpubl. data) and Casey et al. (1992). Spinel data and bathymetry from Bonatti et al. (1992) and Sobolev et al. (1991, 1992a).

Figure F4. A. Maps showing locations of conventional seismic refraction profiles (long white lines) and NOBEL experiments (numbered black lines) in 1997. B–D. Preliminary interpretation of data from the long refraction profile (R. Detrick and J. Collins, pers. comm., 1998); (B) two-dimensional velocity model with contours labeled in kilometers per second, (C) indication of data coverage, which is sparse in the lower crust but sufficient to define large lateral velocity variations (contours labeled in kilometers per second), and (D) traveltime data (circles) with model calculations (shading) for comparison. E. Comparison of two one-dimensional sections through the velocity model with a typical one-dimensional section for oceanic crust at the East Pacific Rise (EPR). MAR = Mid-Atlantic Ridge.

Figure F5. Leg 209 drill site locations.

Figure F6. A. Schematic diagram drawn after Barnouin-Jha et al. (1997) showing results for the upper 50 km in a dynamic model of buoyancy-driven three-dimensional (3-D) mantle flow beneath a slow-spreading ridge. Red = flow vectors in the horizontal plane, yellow = flow vectors in the vertical ridge-axis plane, blue = flow vectors in the vertical ridge-normal plane. This illustrates along-axis flow in the shallow mantle from segment centers to segment ends. Note spacing between upwelling centers is ~400 km and the region of melt generation is almost as long as the ridge segments. B. From Ceuleneer (1991), illustrating ductile flow vectors and shear sense inferred from peridotite fabrics in the mantle section of the Maqsad area, Oman ophiolite. Map area is ~17 km long x 14 km wide. Approximate location of inferred paleoridge axis is shown as a red line. C. Schematic diagram from Jousselin et al. (1998) showing their vision of mantle flow, loosely based on observations from the Oman ophiolite, with a narrow zone of upwelling and a thin region of corner flow feeding a ridge segment that is three times longer than the diameter of the mantle upwelling zone. This model requires extensive subhorizontal ridge-parallel flow of residual mantle peridotite from the segment center to the segment ends. Although this geometry seems somewhat extreme and has not been produced in any 3-D dynamic model to date, it illustrates the type of highly focused solid upwelling that could produce the observed along-axis variation in crustal thickness on the Mid-Atlantic Ridge via 3-D focusing of mantle flow. Dynamic models such as that illustrated in A do not have sufficiently narrow zones of mantle upwelling and cannot reproduce the lengths of observed magmatic segments (~30–100 km). MOHO = Mohorovicic seismic discontinuity.

Figure F7. Downhole lithologic distributions in Hole 1268A are shown as percentages for each section (left panel) and as volume percent recovery. The lithologies are grouped as dunites (rocks with 90% olivine), harzburgites (includes all ultramafic rocks exclusive of dunites, but the bulk are harzburgites in the strict sense), gabbros (includes gabbronorites, microgabbros, and gabbroic rocks of undeterminable mineralogy), and breccias (mixtures of highly altered, probably originally gabbroic material and ultramafic host). TD = total depth.

Figure F8. A. Data on the seismic velocity of the mantle beneath oceanic crust (depths >7 km) from seismic refraction experiments, as a function of oceanic plate age. Data compiled from Beé and Bibee (1989), Bratt and Purdy (1984), Bratt and Solomon (1984), Canales et al. (1998, 2000, 2003), Cary and Chapman (1988a), Collins et al. (1989, 1995), Collins and Brown (1998), Detrick et al. (1993, 1994, 1998), Detrick and Purdy (1980), Duennebier et al. (1987), Fowler (1976, 1978), Fowler and Keen (1979), Ginzburg et al. (1985), Hooft et al. (2000), Lewis and Garmany (1982), Lewis and Snydsman (1979), McClain and Atallah (1986), Minshull et al. (1991), Mithal and Mutter (1989), Morris et al. (1993), Orcutt et al. (1976), Purdy (1983), Purdy and Detrick (1986), Shearer and Orcutt (1986), Spudich and Orcutt (1980), Vera et al. (1990), Waldron et al. (1990), White (1979), White et al. (1992), Whitmarsh and Calvert (1986), Whitmarsh et al. (1982, 1990). B. Calculated P-wave velocity for polycrystalline dunite and pyroxenite at 25°C as a function of pressure, with 1 error estimates, from Christensen and Mooney (1995). Based on these estimates, harzburgite with 20% orthopyroxene at the base of oceanic crust would have a P-wave velocity = ~8.2 ± 0.1 km/s at 25°C, or ~ 8.1 km/s at 200°C.

Figure F9. A. Close-up photograph of hand specimen of a typical example of completely serpentinized harzburgite (interval 209-1268A-19R-1, 50–62 cm). B. X-ray diffraction spectrum of that sample shows that it consists almost entirely of lizardite (chrysotile[?]) and magnetite.

Figure F10. A. Close-up photograph of partially talc-altered serpentinite (interval 209-1268A-23R-2, 15–28 cm) with remnants of serpentine pseudomorphs after orthopyroxene. Arrows indicate approximate position of thin section photographs. B, C. Photomicrograph showing pervasive talc alteration (Sample 209-1268A-23R-2, 18–21 cm) (cross polarized light: blue filter; field of view - 2.75 mm). (B) Talc alteration has completely overprinted the matrix texture of the serpentinized harzburgite. A bastite pseudomorph after orthopyroxene (left) is being replaced by veinlets of massive talc (image 1268A_040). (C) Talc replacing orthopyroxene (to the left of prominent spinel crystal in center of image) and former serpentine-chrysotile veins (left). Talc has completely erased the mesh texture of the serpentinized harzburgite (image 1268A_039).

Figure F11. Primitive glasses (Mg# > 60%) from the 14°–16°N region of the Mid-Atlantic Ridge (Melson et al., 1977; C. Xia et al., unpubl. data) extend to >52 wt% SiO₂ and are among the most SiO₂-rich primitive glasses that have been recovered from the mid-ocean ridges. Mid-ocean-ridge glass data, including data from the Melson et al. catalog, were downloaded in April 2003 from PetDB (online at petdb.ldeo.columbia.edu/petdb/).

Figure F12. Thin section photomicrograph showing the core of an orthopyroxene crystal pseudomorphed by talc and pyrite. The core is surrounded by a corona of pleochroic serpentine. Pseudomorphic talc and pleochroic serpentine are replaced by late, nonpseudomorphic talc (bottom) (Sample 209-1268A-13R-2, 3–6 cm) (cross polarized light: blue filter; field of view = 2.75 mm; image 1268A_016).

Figure F13. A. Anhydrous major element compositions of hydrothermally altered peridotite samples from Site 1268, plotted in terms of mole fraction of SiO₂ vs. mole fraction of MgO + FeO, compared to anhydrous compositions of end-member minerals (speciation is SiO₂, TiO₂, Al₂O₃, Cr₂O₃, FeO, MnO, MgO, CaO, Na₂O, K₂O, P₂O₅, with all Fe as FeO). B. Expanded plot of A, with weight percent pyroxene proportions added. Talc-altered peridotites approach the composition of end-member talc. Even serpentinized peridotites have ~25 wt% normative orthopyroxene, indicative of substantial addition of SiO₂ and/or loss of MgO + FeO. Slight divergence of the peridotite compositions from the MgO + FeO to SiO₂ mixing line is due to the minor presence of other oxides, mainly in the form of a few percent Cr-Al spinel in most peridotite samples.

Figure F14. A–D. Anhydrous major element compositions of hydrothermally altered gabbronorite samples from Site 1268, plotted in terms of oxide mole fractions (speciation as for Fig. F13). Calcic amphibole end-members are tremolite/actinolite, hornblende, edenite, pargasite, and tschermakite. Chlorite end-member is clinochlore. Bold triangles in panels B, C, and D enclose possible gabbronorite assemblages with calcic plagioclase. Light gray arrows show trend of increasing alteration and metasomatism for Site 1268 gabbronorites. The freshest samples have compositions similar to igneous gabbronorites, whereas the most altered samples approach the composition of talc-chlorite-quartz mixtures, in accord with the observed alteration assemblage. Cpx = clinopyroxene, opx = orthopyroxene, serp = serpentinite.

Figure F15. Close-up photograph of pegmatitic textured gabbro (interval 209-1268A-21R-1, 10–21 cm).

Figure F16. Magnetic inclination as a function of the amount of rotation around a horizontal axis, for rocks that initially are normally magnetized with a magnetic inclination of 28° and a declination of 0°. Inclinations of 15°, the average value in gabbronorite from Hole 1268A, and 36°, the average value in talc-altered peridotite from Hole 1268A, are shown for reference. Small rotations around an east-west axis could produce inclinations different from 28°. However, this would require opposite senses of rotation for the gabbronorites and the peridotites. Geological reasoning suggests that tectonic rotations in this area are likely to be counterclockwise around a ridge-parallel, nearly horizontal axis striking 020°. Large rotations around a horizontal axis striking 020° could produce the observed 15° inclination in gabbronorites, whereas smaller counterclockwise rotations could produce the observed 36° inclination in peridotites. We hypothesize the gabbronorites acquired their remanent magnetization at somewhat higher temperature than the peridotite during slow cooling of the rocks as they were uplifted toward the seafloor. Peridotites acquired their remanent magnetization at a later time, during magnetite growth associated with hydrothermal alteration and serpentinization at ~300°C. Rotation began before serpentinization of the peridotites, so the peridotites record only part of the tectonic rotation. CW = clockwise, CCW = counterclockwise.

Figure F17. Magnetic inclination as a function of the amount of rotation around an axis with an azimuth of 020° for rocks which initially are normally magnetized with a magnetic inclination of 28° and a declination of 0°, showing the effect of different plunges of the rotation axis.

Figure F18. Spherical projections illustrating the effects of rotation of a remanent magnetic vector initially striking 360° with an inclination of 28° around axes striking 360°–040° and plunging –10° to +10°. The initial position of the vector is indicated with a black cross. The rotation axis is shown as a square. Closed symbols = projection points in the lower hemisphere, open symbols = projection points in the upper hemisphere. Small circles = projection points of the rotating magnetic vector plotted at 20° rotation intervals, large circles = magnetic inclinations of 28° (blue = expected at this latitude), 15° (black = average observed in gabbroic rocks), 28°, and 36° (dashed = average observed in talc-altered peridotites). Effects on the inclination are relatively large and are sufficient to account for the observed inclinations in gabbronorite and peridotite in Hole 1268A. Also, in some cases, the effects on the azimuth of the remanent magnetization vector are large. For example, in the most extreme case illustrated, the azimuth changes by almost 90° of counterclockwise rotation around an axis striking 040° and plunging 10° to the south (lower left). CW = clockwise, CCW = counterclockwise.

Figure F19. Close-up photograph of highly (>15 vol%) vesicular basalt (interval 209-1269B-1R-1, 39.5–47 cm).

Figure F20. Proportions of lithogies recovered from Holes 1270A, 1270B, 1270C, and 1270D. TD = total depth.

Figure F21. Photomicrographs depicting the variations in deformation of the oxide gabbronorites ranging from relatively undeformed to strongly deformed. A–C. This oxide gabbronorite has an undeformed igneous texture with lath-shaped euhedral plagioclase, subhedral to euhedral clinopyroxenes, subhedral orthopyroxene, and interstitial oxides (Sample 209-1270B-10M-1, 117–120 cm). (A) The red box indicates the position of the magnified image in B (field of view [FOV] = 11 mm; plane-polarized light: blue + dark gray filters; image 1270B_033). (B) Close up of A showing that although the texture is magmatic the rock has experienced some deformation as indicated by kinked plagioclase and release of strain associated with the polysynthetic twinning (cross polarized light [XPL]: blue + dark gray filters; FOV = 2.75 mm; image 1270B_034). (C) Preservation of an optically zoned core (outlined in red) in a slightly deformed plagioclase (XPL: blue filter; FOV = 1.4 mm; image 1270B_013). D. Ribbon-textured plagioclase surrounded by plagioclase neoblasts in a strongly deformed oxide gabbronorite (Sample 209-1270B-1R-1, 90–93 cm) (XPL: blue filter; FOV = 2.75 mm; image 1270B_036). E, F. Wide view and detail of a stretched orthopyroxene grain (Sample 209-1270B-4M-1, 108–111 cm) (XPL: blue filter). (E) The stretched orthopyroxene grain is outlined in red. It sits in a matrix of plagioclase neoblasts and has neoblast of orthopyroxene in the middle of the grain where it has failed. The yellow box shows the location of a detailed view of the central boudinaged part of this grain (FOV = 11 mm; image 1270B_037). (F) Detail of a boudinaged orthopyroxene crystal with neoblasts of orthopyroxene and plagioclase (FOV = 2.75 mm; image 1270B_038).

Figure F22. Anhydrous major element compositions of gabbronorite samples from Hole 1270B, plotted in terms of oxide mole fractions (mole fractions calculated for SiO₂, TiO₂, Al₂O₃, Cr₂O₃, FeO, MnO, MgO, CaO, Na₂O, K₂O, P₂O₅, with all Fe as FeO). Calcic amphibole end-members are tremolite/actinolite, hornblende, edenite, pargasite, and tschermakite. Chlorite end-member is clinochlore. The bold triangle encloses possible pyroxene gabbro and gabbronorite assemblages with calcic plagioclase. All of the Hole 1270B gabbronorites lie outside this triangle, indicating that they must have contained substantial proportions of igneous olivine (dashed lines) and/or Fe-Ti oxides (thin lines). Olivine could have been consumed by subsolidus or near-solidus oxidation reactions that produced pyroxene + magnetite. However, the high Ti contents of all Hole 1270B gabbronorites require substantial proportions of normative ilmenite and/or magnetite, so it is very likely that these rocks all contain igneous Fe-Ti oxides.

Figure F23. Glasses from the 14° to 16°N region of the Mid-Atlantic Ridge (Melson et al., 1977; C. Xia et al., unpubl. data) have a minimum Mg# = ~50%. Glasses with Mg#s < 50% are rare along the Mid-Atlantic Ridge. Mid-ocean-ridge glass data, including data from the Melson et al., Smithsonian catalog, were downloaded in April 2003 from PetDB (online at petdb.ldeo.columbia.edu/petdb/).

Figure F24. Photomicrograph of zircon crystals (Sample 207-1270D-4R-1 [Piece 7, 33–37 cm]) (cross polarized light: blue filter; field of view = 0.7 mm; image 1270D_002).

Figure F25. Anhydrous major element compositions of hydrothermally altered peridotite samples from Sites 1268 and 1270, plotted in terms of mole fraction of SiO₂ vs. mole fraction of MgO + FeO, compared to anhydrous compositions of end-member minerals (speciation as for Fig. F22). Tick marks along the vector connecting olivine and pyroxene are for variations in normative orthopyroxene content in terms of weight percent orthopyroxene/(orthopyroxene + olivine), assuming olivine Mg# = orthopyroxene Mg#. All but two of the serpentinized peridotites from these sites have >30 wt% normative orthopyroxene, indicative of substantial addition of SiO₂ and/or loss of MgO + FeO. Slight divergence of the peridotite compositions from the MgO + FeO to SiO₂ mixing line is due to the minor presence of other oxides, mainly in the form of a few percent Cr-Al spinel in most peridotite samples.

Figure F26. Photograph of Shinkai 6500 dive sample 425-R007, a peridotite mylonite ~45 cm long, recovered from a planar, striated outcrop surface dipping ~20° to the west, just a few tens of meters from the site of Hole 1270A. Textures in this sample record localized mylonitic deformation of peridotite at ~600°C, strikingly different from textures in peridotites from Holes 1270A, 1270C, and 1270D, which record protogranular to mildly porphyroclastic deformation at temperatures >1000°C, except within a few millimeters of gabbro-hosted, mylonitic shear zones.

Figure F27. A, B. Data from Hole 1270B gabbronorites, comparing the intensity of crystal-plastic deformation (estimated during visual core description, supplemented with thin section observation) to the proportion of magnetite estimated from magnetic susceptibility measurements on large individual core pieces. Deformation intensities vary on a small scale, due to the presence of anastomosing shear zones around less deformed gabbroic rocks, so the deformation intensity data in both A and B are smoothed using a 10-piece running average, weighted by piece length. The estimated proportion of magnetite also shows large variations on short length scales, as can be seen in A. Thus, in B, the magnetite proportions are presented as a 5-point running median. Magnetite proportions have not been estimated for every core piece and the spacing of estimates downhole is not uniform, but the data are so sparse that there is no way to weight the observations by length. For this reason, the comparison illustrated here is only preliminary. However, this provisional comparison supports the hypothesis that deformation intensity and magnetite proportions are correlated within some intervals of the core, as has been proposed for ODP Holes 735B and 1105A (e.g., Dick et al., 1991, 2000, 2002; Natland and Dick, 2002; Natland et al., 1991; Niu et al., 2002; Ozawa et al., 1991; Pettigrew, Casey, Miller, et al., 1999).

Figure F28. Schematic illustration of possible dips of high-temperature crystal-plastic foliations in gabbroic rocks from Hole 1270B, and peridotites from Holes 1270C and 1270D. Gabbroic rocks record dominantly negative magnetic inclinations, suggesting that these rocks may be reversely magnetized, consistent with—but not required by—their position ~13–18 km east of the rift axis. Based on this assumption, rotation of crystal-plastic foliations around a vertical axis to reorient core so that remanent magnetizations have a constant azimuth planes that dipped east at ~45° when the remanent magnetization vector pointed south. A. If both gabbroic rocks and peridotites are reversely magnetized, the crystal-plastic foliations in peridotites dipped west at ~45° when the remanent magnetization vector pointed south, at a 90° angle to the fabrics in gabbroic rocks. B. If, instead, both gabbroic rocks in Hole 1270B and peridotites in Holes 1270C and 1270D are normally magnetized, high-temperature crystal-plastic foliations in gabbroic rocks from Hole 1270B dipped east at ~45° when the remanent magnetization vector pointed north and mylonite zones in peridotites from Holes 1270C and 1270D dipped west at ~45° when the remanent magnetization vector pointed north. C. It is assumed that the peridotites were magnetized later than the gabbroic rocks and record normal polarity. With this assumption, crystal-plastic fabrics in Holes 1270B, 1270C, and 1270D are parallel and dipped ~45° to the east when the remanent magnetization vector in Hole 1270B pointed south and the remanent magnetization vector in Holes 1270C and 1270D pointed north.

Figure F29. Magnetic inclination as a function of the amount of rotation around a horizontal axis for rocks with a magnetic inclination of ±28° and a declination of 0°. Inclinations of –14°, the average value in gabbro and gabbronorite from Hole 1270B, and –3°, the average value in peridotite from Holes 1270C and 1270D, are shown for reference. Relatively small rotations around a horizontal east-west axis could produce the inclinations in the gabbroic rocks and the peridotites. However, this is problematic if gabbroic rocks are reversely magnetized and peridotites are normally magnetized, as in Figure F28C. If all samples are normally magnetized, as in Figure F28B, then clockwise rotations of 42° and 31° around an east-west axis could produce the inclinations in the gabbroic rocks and the peridotites, respectively. However, there is little evidence for east-west—striking, north-dipping faults in bathymetric data from the region around Site 1270 (e.g., Fujiwara et al., 2003). Instead, geological reasoning suggests that tectonic rotations in this area are likely to be clockwise around a ridge-parallel, nearly horizontal axis striking 020°. Large rotations around a horizontal axis striking 020° could produce the observed –14° inclination in gabbroic rocks, whereas somewhat smaller clockwise rotations could produce the observed –3° inclination in peridotites. For the scenario in Figure F28C, in which gabbroic rocks are reversely magnetized (dashed curves) while peridotites are normally magnetized (solid curves), this requires ~45°–50° of clockwise rotation after both rock types acquired their remanent magnetization. For the scenario in Figure F28B, in which both gabbroic rocks and peridotites are normally magnetized (solid curves), we hypothesize the gabbronorites acquired their remanent magnetization at the magnetite blocking temperature (~500°–570°C) during slow cooling of the rocks as they were uplifted toward the seafloor. Peridotites acquired their remanent magnetization at a later time, during magnetite growth associated with hydrothermal alteration and serpentinization at ~300°C. Rotation began before serpentinization of the peridotites, so the peridotites record only part of the tectonic rotation.

Figure F30. Magnetic inclination as a function of the amount of rotation around an axis with an azimuth of 020° for rocks which initially have a magnetic inclination of ±28° and a declination of 0°, showing the effect of different plunges of the rotation axis. Solid curves = normally polarized rocks, dashed curves = reversely polarized rocks.

Figure F31. Lithostratigraphic summary for Holes 1271A and 1271B. TD = total depth.

Figure F32. Close-up photograph of network of gabbro intrusions in dunite (interval 209-1271A-1R-1, [Piece 9, 37–51 cm]). A large crystal of clinopyroxene (now amphibole) is included in the dunite at 46–51 cm.

Figure F33. Anhydrous major element compositions of hydrothermally altered peridotite samples from Sites 1268, 1270, and 1271, plotted in terms of mole fraction of SiO₂ vs. mole fraction of MgO + FeO + CaO, compared to anhydrous compositions of end-member minerals (speciation as for Fig. F22). Tick marks along the vector connecting olivine and pyroxene indicate variations in normative orthopyroxene content, in terms of weight percent orthopyroxene/(orthopyroxene + olivine), assuming olivine Mg# = orthopyroxene Mg#. Most serpentinized dunites and harzburgites from Site 1271 have <30 wt% normative orthopyroxene and so do not require metasomatic increases in Si/(Mg + Fe) during alteration. Substantial divergence of the Site 1271 peridotite compositions from the MgO + FeO to SiO₂ mixing line is due to the presence of interstitial gabbroic material, originally including plagioclase and/or igneous amphibole, especially in the impregnated dunite sample. Site 1270 dunites also include relatively abundant Cr-Al spinel.

Figure F34. Molar Mg# vs. Ni concentration for peridotite samples from Sites 1270 and 1271. Whole-rock Mg#s are assumed to be close to original olivine Mg#s in all samples. The whole-rock Ni content in dunites is a good approximation for the Ni concentration in olivine in these samples. Ni concentrations in olivine in the harzburgites were estimated using an olivine/orthopyroxene Ni distribution coefficient of 4 at ~1250°C (Kelemen et al., 1998a), appropriate for igneous conditions beneath a mid-ocean ridge, and bounds of 10 and 30 wt% on the modal proportion of orthopyroxene. The Ni concentration in olivine in a sample of impregnated dunite was calculated in the same way. The olivine crystallization curve shows the composition of olivine in equilibrium with liquid during fractional crystallization of olivine from primitive mid-ocean-ridge basalt (MORB). The initial liquid composition was an estimated 10% melt of the MORB source, assuming polybaric incremental melting with an average pressure of 1 GPa (Kinzler and Grove, 1992, 1993). We used an olivine/liquid Fe/Mg K_d of 0.3 (Roeder and Emslie, 1970) and the dependence of the olivine/liquid Ni distribution coefficient on MgO content of the liquid determined by Hart and Davis (1978). Light gray symbols in the background are olivine compositions from the crust–mantle transition zone (MTZ) in the Samail and Wadi Tayin massifs of the Oman ophiolite (Godard et al., 2000; Koga et al., 2001). The Site 1271 dunites with the lowest Mg# have olivine Ni concentrations comparable to Ni in olivine in the most depleted residual harzburgites from Sites 1270 and 1271. These data are consistent with formation of the dunites by reaction between relatively low Mg# melt and residual mantle olivine and inconsistent with formation of the dunites as olivine cumulates during fractional crystallization of olivine from primitive MORB. Opx = orthopyroxene, harz = harzburgite.

Figure F35. Close-up photograph of chromitite pod in dunite. The matrix surrounding chromite in the chromitite pod consists of chlorite (possibly replacing plagioclase) and amphibole. The host dunite intruded by the chromitite is also rich in chromian spinel (interval 209-1271B-4R-2 [Piece 2, 38–48 cm]).

Figure F36. Close-up photograph of brown amphibole gabbro (interval 209-1271B-6R-1, 21–27 cm).

Figure F37. Photomicrographs showing that the evolution of brown amphibole gabbro during deformation and metamorphism is complex and involves several stages. A–C. Sample 209-1271B-14R-1, 25–29 cm (cross polarized light [XPL]: blue filter). (A) Ductile deformation. Amphibole porphyroclast with wavy extinction (a(p)) represents a microboudin that has been syntectonically recrystallized to fine-grained amphibole neoblasts (a(n)) along its margins (field of view [FOV] = 5.5 mm; image 1271B_058). (B) Amphibole porphyroclasts (a(p)) have been partially replaced by amphibole neoblasts (a(n)). The amphibole is surrounded by bands of former plagioclase neoblasts that have been statically altered to sericite (sc(n)) (FOV = 5 mm; image 1271B_059). (C) The outline of a "ghost" plagioclase porphyroclast replaced by sericite (sc(p)) can be recognized in this sample. The original polysynthetic twinning of the plagioclase porphyroclasts (white arrows) is pseudomorphed by sericite with different orientations. Former plagioclase neoblasts are pseudomorphically replaced by sericite aggregates (sc(n); red arrows) (FOV = 2.75 mm; image 1271B_060). D, E. Sample 209-1271B-11R-1, 44–46 cm (XPL: blue + dark gray filters; FOV = 1.4 mm). (D) Intergrowth of fibrous amphibole (a2) and chlorite after former high-temperature amphibole. Fibrous amphibole and chlorite are stretched and deformed along shear bands (image 1271B_049). (E) Amphibole-chlorite schists after gabbroic protolith. The secondary fibrous amphibole (a2), sericite (sc), and chlorite (ch) are stretched and banded, defining the schistosity (image 1271B_048). F. Fold of fibrous amphibole (a2) in a amphibole-chlorite schists after gabbro. Chlorite (ch) crystallized in the pressure shadow of the inner part of the fold. The matrix is composed of fine fibrous aggregates of secondary amphibole (a2) and chlorite (a2) (may by replacing former high-temperature amphibole) alternating with dark bands of sericite and hydrogrossular (sc and hgr) (may by replacing former plagioclase) (Sample 209-1271B-11R-1, 55–57 cm) (plane polarized light: blue + dark gray filters; FOV = 5.5 mm; image 1271B_053).

Figure F38. Stratigraphic distribution of lithologic proportions, lithologic variability, and modal orthopyroxene in peridotites are shown along with the graphical depiction of the recovery for each interval. TD = total depth.

Figure F39. Core photographs of alternating bands of dunite (D) and harzburgite (H). Harzburgite at the bottom is enriched in orthopyroxene (opx) (interval 209-1272A-23R-1 [Pieces 2, 3, 4–77 cm]).

Figure F40. A. Comparison of compositions of diabase and miarolitic gabbro from Hole 1272A with compositions of Mid-Atlantic Ridge basalt glasses. The Hole 1272A samples have compositions that lie within the range of observed glass compositions, consistent with the hypothesis that all preserve liquid, rather than cumulate, compositions. B. Like glasses from the 14° to 16°N region, the samples from Site 1272 fall at the high end of the range of SiO₂ and Zr concentration in Mid-Atlantic Ridge glasses at a given Mg#. Glass compositions from the 14° to 16°N region of the Mid-Atlantic Ridge are from Melson et al. (1977) and C. Xia et al. (unpubl. data). Mid-ocean-ridge glass data, including data from the Melson et al. catalog, were downloaded in April 2003 from PetDB (online at petdb.ldeo.columbia.edu/petdb/).

Figure F41. Anhydrous major element compositions of hydrothermally altered peridotite samples from Sites 1268, 1270, 1271, and 1272, plotted in terms of mole fraction of SiO₂ vs. mole fraction of MgO + FeO, compared to anhydrous compositions of end-member minerals (speciation as for Figure F22). Tick marks along the vector connecting olivine and pyroxene are for variations in normative orthopyroxene content, in terms of weight percent orthopyroxene/(orthopyroxene + olivine), assuming olivine Mg# = orthopyroxene Mg#. Most serpentinized dunites and harzburgites from Site 1272 have <25 wt% normative orthopyroxene and so do not require metasomatic increases in Si/(Mg + Fe) during alteration. Substantial divergence of two Site 1272 peridotite compositions from the MgO + FeO to SiO₂ mixing line is due to the presence of carbonate alteration in veins and perhaps also in the serpentinite matrix.

Figure F42. Close-up photographs of examples of semiplastic fault gouges from Hole 1272A. The fault gouges were semicohesive fault breccias with matrix-supported clasts. A. Interval 209-1272A-18R-1, 134–149 cm. B. Interval 209-1272A-18R-2, 1–19 cm. C. Interval 209-1272A-19R-1, 138–150 cm. D. E. Interval 209-1272A-19R-2, 1–18 cm. F. Interval 209-1272A-25R-2, 8–26 cm.

Figure F43. Stratigraphic distribution of lithologic proportions and lithologic variability in Hole 1274A with a graphical depiction of the recovery for each interval. TD = total depth.

Figure F44. Photomicrographs of Sample 209-1274A-8R-1, 15–18 cm (cross polarized light). A. Clinopyroxene-spinel symplectite at the contact between protogranular orthopyroxene and olivine. The symplectite extends into the olivine (field of view = 2.75 mm; image 1274A_049). B. Close up showing that the clinopyroxene is a single crystal with undulatory extinction (field of view = 1.4 mm; image 1274A_050).

Figure F45. Anhydrous major element compositions of hydrothermally altered peridotite samples from Sites 1268, 1270, 1271, 1272, and 1274, plotted in terms of mole fraction of SiO₂ vs. mole fraction of MgO + FeO, compared to anhydrous compositions of end-member minerals (speciation as for Figure F22). Tick marks along the vector connecting olivine and pyroxene are for variations in normative orthopyroxene content, in terms of weight percent pyroxene/(pyroxene + olivine), assuming olivine Mg# = pyroxene Mg#. Partially serpentinized "dunites" from Site 1274 have ~5–13 wt% normative pyroxene. Five of six partially serpentinized harzburgites from Site 1274 have 26–29 wt% normative pyroxene, more than expected for highly depleted residual peridotites, and so may have undergone metasomatic increases in Si/(Mg + Fe) during alteration. Divergence of a Site 1274 fault gouge mud from the MgO + FeO + CaO to SiO₂ mixing line is probably due mechanical mixing of gabbroic and ultramafic material in the fault gouge.

Figure F46. Weight percent CO₂ vs. weight percent CaO in metaperidotite samples from ODP Legs 153 and 209. Leg 153 data from Cannat, Karson, Miller, et al. (1995) and Casey (1997). The upper diagram shows that most samples with high CaO have high CO₂ and lie along a mixing line toward calcite and aragonite CaCO₂. A few samples with high CaO and low CO₂ are impregnated peridotites, which include igneous plagioclase and clinopyroxene precipitated from melts migrating along grain boundaries. The lower diagram shows more detail for samples with low CO₂ and CaO contents. Although the data are scattered, much of this may be due to analytical uncertainty. Most metaperidotites from Legs 153 and 209 contain appreciable CO₂, mainly in aragonite and calcite. Thus, carbonate alteration accounts for a significant fraction of the whole-rock CaO budget in these rocks. This suggests, but does not prove, that CaO may have been metasomatically added to these rocks along with CO₂ during hydrothermal alteration.

Figure F47. Sr and Ba concentrations vs. weight percent CO₂ in metaperidotite samples from Leg 209. The samples with the highest Sr contents all include substantial proportions of carbonate; thus, carbonate minerals probably are the primary host for Sr in Leg 209 metaperidotites and much of the Sr may have been introduced during hydrothermal metasomatism. A few samples with high Ba and low CO₂ are impregnated peridotites, which include igneous plagioclase and clinopyroxene precipitated from melts migrating along grain boundaries. However, other samples with high Ba also have high CO₂ contents, consistent with the hypothesis that metasomatically introduced Ba in hydrothermal carbonate provides most of the Ba in the whole-rock budget for many Leg 209 metaperidotites.

Figure F48. Percent of hydrothermal alteration in Sites 1272 and 1274 metaperidotites (estimated during visual core description) vs. density and magnetic susceptibility. Variations in these parameters, as well as data on seismic P-wave velocity and porosity, are moderately well correlated.

Figure F49. Close-up photograph of mud interval from the fault gouge zone recovered in Hole 1274A (interval 209-1274A-24R-1, 3–29 cm).

Figure F50. Stratigraphic distribution of lithologic proportions and lithologic variability of Holes 1275B and 1275D plotted at the same vertical scale. Recovery columns are set at 90 m apart on the same scale (which is the approximate distance between holes). Dashed lines connect intervals of troctolite. TD = total depth.

Figure F51. Close-up photograph of troctolite with intergranular gabbroic material cut by altered gabbroic veins (interval 209-1275D-9R-1, 79–109 cm).

Figure F52. Comparison of experimental pressures for melts equilibrated with olivine + orthopyroxene + clinopyroxene + plagioclase ± spinel, with pressures calculated using the method of Kinzler and Grove (1992, 1993). Pressures were calculated for each of four expressions using normative olivine, plagioclase, quartz, and clinopyroxene in the melt compositions. These four pressures were then averaged, as recommended by Kinzlier and Grove. The error bars on the calculated pressures show the range from maximum to minimum pressure calculated from the four expressions. As the figure shows, this method reproduces the experimental data on which it is calibrated with a precision of about ±0.2 GPa.

Figure F53. Estimated conditions for equilibration of MORB glass compositions from 14° to 16°N along the Mid-Atlantic Ridge with olivine + orthopyroxene + clinopyroxene + plagioclase ± spinel. Equilibration pressures were calculated as described in the caption for Figure F52. Equilibration temperatures were calculated using each of the four pressure estimates for each glass composition, and then averaged. For both pressure and temperature (P-T), the error bars for each individual composition show the range from maximum to minimum calculated pressure and temperature. Error bars for the average P-T estimate for primitive glasses, with molar Mg# > 0.6, are for 2 standard deviations from the mean pressure and temperature. The standard error of the mean for the average pressure-temperature estimate for primitive glasses is about 90 times smaller than the standard deviation.

Figure F54. Estimated molar Mg# vs. Ni concentration in olivine in Site 1275 troctolites, compared to similar estimates for peridotite samples from Sites 1270, 1271, 1272, and 1274. Whole-rock Mg#s are assumed to be close to original olivine Mg#s in all samples. We estimated the olivine Ni concentrations in the troctolites based on the bulk compositions of samples with 0.2–3 wt% CO₂. To do this, we calculated CIPW norms for the anhydrous CO₂-free bulk composition. We then estimated the original Ni concentration in olivine, assuming Ni partitioning between olivine and pyroxenes was governed by partitioning at 1250°C, using relationships derived by Kelemen et al. (1998a), and that there is no Ni in plagioclase (or any other phase, including sulfide). The whole-rock Ni content in dunites is a good approximation for the Ni concentration in olivine in these samples. Ni concentrations in olivine in the harzburgites were estimated using an olivine/orthopyroxene Ni distribution from Kelemen et al. (1998a) and bounds of 10 and 30 wt% on the modal proportion of orthopyroxene. The olivine crystallization curve shows the composition of olivine in equilibrium with liquid during fractional crystallization of olivine from primitive mid-ocean-ridge basalt (MORB). The initial liquid composition was an estimated 10% melt of the MORB source, assuming polybaric incremental melting with an average pressure of 1 GPa (Kinzler and Grove, 1992, 1993). We used an olivine/liquid Fe/Mg K_d of 0.3 (Roeder and Emslie, 1970) and the dependence of the olivine/liquid Ni distribution coefficient on MgO content of the liquid determined by Hart and Davis (1978). Light gray symbols in the background are olivine compositions in dunites, impregnated dunites, and wehrlites from the crust–mantle transition zone (MTZ) in the Samail and Wadi Tayin massifs of the Oman ophiolite (Godard et al., 2000; Koga et al., 2001). Like most Site 1271 dunites and some Oman MTZ dunites, the estimated olivine compositions in the Site 1275 troctolites have olivine Ni concentrations comparable to Ni in olivine in the most depleted residual harzburgites from Sites 1270 and 1271. These Mg#s are consistent with formation of the troctolites by reaction between relatively low Mg# melt and residual mantle olivine and inconsistent with formation of the dunites as olivine cumulates during fractional crystallization of olivine from primitive MORB. Readers should keep in mind that the method of estimating olivine composition, using the bulk composition of highly altered troctolites, is uncertain, so this result is preliminary. Opx = orthopyroxene.

Figure F55. Close-up photograph of gabbro with mixed grain sizes showing patches of fine-grained gabbro in a matrix of coarse-grained gabbro (interval 209-1275B-5R-1, 113–120 cm).

Figure F56. Close-up photographs of gabbro with various grain size domains in curviplanar and sharp (crenulate) contacts. A. Interval 209-1275B-13R-1, 75–91 cm. B. Interval 209-1275D-18R-1, 55–74 cm.

Figure F57. Histograms for molar Mg# in gabbroic rocks from sites drilled during this leg, Site 923 in the Kane Fracture Zone (MARK) area (Agar and Lloyd, 1997), and Site 735B along the Southwest Indian Ridge (Natland and Dick, 2002). Cumulate oxide gabbros and oxide gabbronorites dominate the gabbroic rocks recovered during Leg 209, especially at Sites 1270 and 1275. Diabases and miarolitic gabbros from Sites 1275 and 1272 have higher Mg#s than the oxide gabbros and closely approximate liquid compositions (see the "Site 1272" summary and Fig. F58). Site 1275 troctolites are probably hybrid rocks, formed by interaction between migrating melt and residual mantle olivine. Only at Site 1268 were gabbroic rocks with primitive cumulate compositions sampled. These rocks were gabbronorites. However, extensive alteration makes it difficult to be certain that their high Mg# is a primary igneous feature. In general, the Leg 209 gabbro samples, and particularly the gabbroic rocks from Site 1275, are the most evolved suite of plutonic rocks recovered by ODP drilling along a mid-ocean ridge. However, although some primitive gabbroic cumulates were sampled at Sites 921, 923, and 735B, the abundance of evolved gabbros leads to average compositions with intermediate Mg#s at these other sites. Thus, with the possible exception of Site 1268 gabbronorites and of impregnated peridotites from, for example, Sites 1271 and 1275, no plutonic suite recovered by ODP drilling provides an example of the refractory cumulates (80% < Mg# < 90%) required to balance crystal fractionation of MORB.

Figure F58. Plot of molar Mg# vs. ppm Zr in gabbroic rocks recovered during Leg 209. Also shown are the compositional field defined by mid-ocean-ridge basalt (MORB) glasses from the Mid-Atlantic Ridge, the average of Mid-Atlantic Ridge (MAR) MORB glass compositions, the composition of "normal MORB" (N-MORB) (from Hofmann, 1988), and average compositions of gabbros from ODP Sites 735B (Natland and Dick, 2002), and 923 (Agar and Lloyd, 1997). MORB glass data were downloaded in April 2003 from PetDB (online at petdb.ldeo.columbia.edu/petdb/). Because (1) Zr is an incompatible element that is almost completely retained in the melt during fractional crystallization of basalt, (2) primitive MORB glasses have ~50 ppm Zr, and (3) average MORB has ~100 ppm Zr, it follows that average MORB probably records ~50% crystal fractionation over the span of liquid Mg#s from ~70% in mantle-derived melts to 60% in average MORB. Gabbroic cumulates in equilibrium with melts in this compositional range should have high Mg#s. For example, using an Fe/Mg clinopyroxene/liquid K_d of 0.23 (e.g., Sisson and Grove, 1993a, 1993b) and assuming that 90% of the iron in melt is ferrous, we estimate that Mid-Atlantic Ridge glasses with Mg#s from 73% to 50% should crystallize clinopyroxene with Mg#s of ~93%–82%. Thus, there must be a mass of primitive gabbroic rocks with Mg#s substantially >70% and low Zr concentrations that is approximately equal to the mass of erupted volcanics and sheeted dikes. However, as also noted in the text and in the caption to Figure F57, very few gabbroic rocks sampled from the mid-ocean ridges to date have these characteristics. For example, the average gabbroic rocks from Sites 735 and 923 do not have appropriate compositions to be the complementary cumulates required by MORB fractionation. During Leg 209, only the gabbronorites from Site 1268 and perhaps impregnated peridotites from, for example, Sites 1271 and 1275 could represent part of this complementary refractory cumulate reservoir. Diabases and miarolitic gabbros from Sites 1272 and 1275, together with a few fine-grained gabbroic rocks from Site 1275, have compositions very similar to MORB and are probably chilled liquid compositions. Most of the gabbroic rocks from Site 1275, together with more than one-half of those from Hole 1270B, have such low Mg#s that they cannot play a significant role in the main process of MORB crystal fractionation. Instead, they must crystallize from highly evolved, rarely sampled melts.

Figure F59. Histograms of stable remanent inclinations determined for archive-halves from Hole 1275D, illustrating how different depth intervals preserve distinctly different remanence inclinations. The top 50 m of the hole apparently records a mixture of normally and reversely polarized magnetization. The middle of the hole, from 50 to 140 mbsf, preserves a well-defined shallow negative inclination. Below a relatively sharp break at ~140 mbsf, the lower portion of the hole has a significantly higher negative inclination.

Figure F60. Comparison of rotary core barrel (RCB) and resistivity at the bit while coring (RAB-C) in the first three cores for each hole drilled during Leg 209.