Site 1275 is located along the track of Shinkai 6500 Dive 422 on the summit of a topographic dome, known informally as Mt. Mike, ~10 mi west of the Mid-Atlantic rift valley at ~15°44'N. This dome has been interpreted on the basis of bathymetric data as an oceanic core complex, or "megamullion" (Escartin and Cannat, 1999; Escartin et al., in press; Fujiwara et al., 2003; MacLeod et al., 2002). Dive 422 recovered seven samples of gabbro from ~2290 to ~1750 mbsl, mostly from steep blocky outcrops along the southern flank of the dome, and then two samples of peridotite and one sample of gabbro from nearly flat, striated outcrops and overlying talus across the summit of the dome at ~1650–1540 mbsl. In 2001, the James Clark Ross made a survey of this dome using the BRIDGE portable wireline rock drill and dredging to sample (Escartin et al., in press; MacLeod et al., 2002). In addition to gabbro and peridotite, short (~1 m) drill cores from the summit of the dome recovered abundant diabase.

Holes 1275A and 1275B were located in an area with horizontal, striated outcrop surfaces extending for ~100 m², ~300 m north of the top of the dome and 90 m north of the northernmost Shinkai 6500 sample, 422-R010 (described as highly altered plagioclase-bearing dunite and peridotite, recovered from talus resting on a flat outcrop surface). In an attempt to recover as much material as possible from the upper 5 m of the outcrop, we retrieved the core barrel after 5 m of drilling. Core 209-1275A-1R was composed of a small amount of diabase and carbonate-cemented diabase breccia. Recovering this core involved lifting the drill bit above seafloor, so we immediately spudded Hole 1275B a few meters away. Diabase composed 100% of the rocks from Cores 209-1275B-1R and 2R and >50% of Core 3R, extending to 18 mbsf. Below this depth, core was dominated by gabbroic rocks intruded by a few diabase sills or dikes, except within an interval of troctolite or impregnated dunite from ~30 to ~30–35 mbsf. Given the ultramafic lithology in Shinkai 6500 sample 422-R010, we were surprised that the first ultramafic rocks in Hole 1275B were not recovered until the hole reached a depth of 30 mbsf.

After 109 m of drilling in Hole 1275B, it was necessary to change the bit. We marked Hole 1275B for possible reentry but also used the opportunity to move to a position directly above sample 423-R010 to test the resistivity-at-the-bit with coring (RAB-C) coring equipment. The RAB-C test in Hole 1275C recovered very little core, but this included two small pieces of troctolite. Thus, we elected to spud Hole 1275D just a few meters away from Hole 1275C to determine the extent of lateral variability of the lithologies at this site and to provide a complementary rotary core barrel (RCB) hole to fully evaluate the coring potential of the RAB-C equipment. The first three cores in Hole 1275D recovered almost 8 m of core over ~17.6 m of penetration, the best recovery for the first three cores in any hole drilled during Leg 209. By comparison, Hole 1270C, drilled just a few meters away with the RAB-C bit, recovered just a few tens of centimeters of core over ~21 m of penetration, which was the worst recovery for the first three cores in any hole during Leg 209.

In striking contrast to Hole 1275B, troctolite, or impregnated dunite, was recovered in the first 11 cores of Hole 1275D, extending to ~56 mbsf, though diabase was also present in the first core. Below 56 mbsf, the hole passed into gabbroic rocks intruded by diabase sills or dikes, with a second interval of troctolite between ~80 and 95 mbsf. Thus, although Hole 1275D was only ~90 m south of Hole 1275B, the lithostratigraphy is quite different in the two holes. Correlation of rock units from the two holes is still in an initial stage. However, the simplest interpretation seems to be that the lower troctolite in Hole 1275D, flanked by gabbroic rocks, is correlative to the first (and only) troctolite in Hole 1275B, which is also surrounded by gabbroic rocks. If so and assuming a roughly planar upper contact between troctolite and overlying gabbro, the apparent dip of this contact is ~30° to the south in the plane formed by the two drill holes (Fig. F50). Intriguingly, this dip is parallel to the south slope of the dome along the track of the Shinkai 6500 Dive 422 from sample 422-R003 to 422-R008.

Proportions of Igneous Rocks

The igneous and (possibly) residual mantle protoliths of recovered core at Site 1275, weighting the proportions of recovered lithologies by the depths of each hole, were 14% troctolite or impregnated dunite, 74% gabbroic rocks (gabbro, gabbronorite, oxide gabbro, oxide gabbronorite, and olivine gabbro), 10% diabase, and 2% granophyre (Fig. F50). The large proportion of gabbroic rocks and the presence of diabase were no surprise, given past submersible sampling and more recent dredge and BRIDGE drilling results from Mt. Mike. Obviously, Site 1275 is very different from all previous sites drilled during this leg. Drilling at Sites 1274 and the lower portion of Site 1272 recovered just a few percent gabbroic rocks, and at Sites 1268, 1270, and 1271 we recovered ~75% peridotite and ~25% gabbro. Considering all these sites together, without weighting for the different depths of drilling, we can tentatively infer that the crust from 14°39' to 15°44' is composed of 20%–40% variably altered gabbroic rocks and 80%–60% partially serpentinized peridotite. As discussed in "Site 1268," it is possible that these proportions of gabbro to peridotite extend to depths below the Moho.

Petrogenesis of Troctolites or Impregnated Dunites

As mentioned above, a lithology termed "troctolite" in lithologic logs and site chapters was recovered in Holes 1275B, 1275C, and 1275D. Several meters of troctolite were recovered from between ~30 and 35 mbsf in Hole 1275B, and tens of meters of troctolite were recovered from the upper 95 m of Hole 127D (Fig. F51). The rocks are highly altered, but relict primary phases remain and the alteration was static so that pseudomorphic assemblages allow unambiguous determination of the primary mineralogy. The protolith of this rock type was composed of 50% to almost 100% olivine with 1%–3% chromian spinel surrounded by poikilitic orthopyroxene and clinopyroxene, and by interstitial plagioclase. Minor brown hornblende in these rocks could be igneous but is probably of metamorphic origin. Olivine crystals are generally rounded and locally embayed, but no reaction rims are present between any of the primary phases.

Because the rock is coarse grained and the proportions of the minerals vary on a centimeter scale, different parts of the troctolite could be classified as dunite, harzburgite, wehrlite, troctolite, or lherzolite. Nonetheless, the different rock types grade into one another and are spatially continuous. Also, chemical analyses reveal that they are relatively homogeneous from one part of the core to another when sample sizes are a few cubic centimeters. For simplicity, throughout this report we refer to all these rocks as troctolites.

The troctolite is cut by numerous centimeter-scale dikes and veins of gabbroic to felsic material. These are almost all 100% altered to talc + amphibole assemblages. However, the presence of pyroxene oikocrysts that cross contacts between troctolite and intruding gabbroic rock suggests that the two rock types are genetically related. Although the gabbroic dikes and veins that cut the troctolite may be coarser than most gabbroic rocks elsewhere in Holes 1275B and 1275D, the two are probably related. Abundant magnetite is observed in most gabbroic rocks from Site 1275, including the dikes and veins cutting troctolite. We have observed qualitatively in ophiolites that grain size in gabbro increases near intrusive contacts with peridotite (P.B. Kelemen, H.J.B. Dick, and J.F. Casey, pers. comm., 2003). Thus, we interpret the troctolite lenses at 30–35 mbsf in Hole 1275B and 80–95 mbsf in Hole 1275D as rafts of older rock included within a gabbroic pluton and the thick troctolite unit that forms the upper 56 m of Hole 1275D as another raft, cut by the fault that forms the top of the dome, or as a part of the roof of the pluton.

The troctolite is of great interest for two main reasons. First, the apparent equilibrium between igneous olivine, pyroxenes, plagioclase, and spinel places important constraints on the pressure at which this assemblage crystallized. Second, some of the olivine in the troctolite may be relict, residual mantle olivine and thus the presence of the troctolite may indicate that the gabbroic rocks from Site 1275 were intruded into mantle peridotite. We discuss each of these topics in turn in the following paragraphs.

First, we discuss constraints on equilibrium between basaltic liquid, olivine, orthopyroxene, clinopyroxene, spinel, and plagioclase. As is well known to igneous petrologists, the five solid phases in this assemblage are related by reactions such as

which consumes olivine plus the anorthite component in plagioclase, stable at relatively low pressure, to produce orthopyroxene, clinopyroxene, and spinel at higher pressures. Because of constrained Fe/Mg exchange between all these phases, this assemblages is nearly univariant in Ca-Mg-Al-Si-Fe. However, because albite (NaAlSi₃O₈) is an important component in plagioclase and Cr spinel (e.g., MgCr₂O₄) is an important part of spinel solid solutions in these rocks, there is actually a broad range of pressure and temperature conditions over which this assemblage can be stable. Because this assemblage also contains several geothermometers (e.g., pyroxene solvus and Cr/Al exchange between orthopyroxene and spinel) and because of recent improvements in understanding of spinel solid solutions, during postcruise research it will be possible to constrain the pressure and temperature at which all these solid phases last equilibrated.

Meanwhile, we can infer likely pressures at which basaltic liquid could have been in equilibrium with olivine, orthopyroxene, clinopyroxene, spinel, and plagioclase [N12]. The troctolites have whole-rock Mg#s of 86%–89%, similar to and slightly lower than typical values for oceanic residual mantle peridotites. This is fortunate, because an extensive body of experimental work has been devoted to determining the relationships between liquid compositions, pressures, and temperatures for equilibrium between melts and the mineral assemblages in mantle peridotites. The presence of plagioclase in this assemblage indicates that the pressure was less than ~1.2 GPa (e.g., Green and Hibberson, 1970). Conversely, it is well known that the minimum pressure for orthopyroxene saturation in primitive melts is a function of silica content in the melt (e.g., Elthon and Scarfe, 1984; O'Hara, 1965; Stolper, 1980). Thus, primitive tholeiitic basalts with ~49 wt% SiO₂ are last saturated in orthopyroxene at a pressure of ~1 GPa (e.g., Baker and Stolper, 1994), whereas primitive andesites with ~54 wt% SiO₂ can be saturated in orthopyroxene at ~ 0.5 GPa (e.g., Jaques and Green, 1980) and boninitic melts with ~58 wt% SiO₂ can crystallize orthopyroxene at 0.1 MPa (e.g., Crawford, 1989). In this context, the relatively SiO₂-rich nature of mid-ocean-ridge basalts from the 14° to 16°N region along the Mid-Atlantic Ridge, first noted in "Site 1268," is significant.

Kinzler and Grove (1992, 1993) developed a relatively simple algorithm that relates the composition of melts saturated in olivine, orthopyroxene, clinopyroxene, and plagioclase (±spinel) to pressure and temperature. Although their main goal was to calculate melt compositions given a source peridotite composition, pressure, and temperature, their method can be inverted to estimate pressure and temperature from a melt composition. Figure F52 shows predicted pressures vs. experimental pressures for the suite of olivine + orthopyroxene + clinopyroxene + plagioclase saturated liquid compositions that was used to calibrate the Kinzler and Grove expressions. The agreement between predicted and experimental values is good, with an uncertainty of approximately ±0.2 GPa, similar to the uncertainty in pressure calibrations for the piston cylinder apparatus used in the experiments.

Accordingly, we have taken all published compositions of basaltic glasses from 14° to 16°N along the Mid-Atlantic Ridge and calculated the pressures and temperatures at which each could be saturated in olivine, orthopyroxene, clinopyroxene, and plagioclase. The results are illustrated in Figure F53. Estimates for primitive glass compositions (molar Mg# = 60%–73%) form a tight cluster, averaging 0.54 GPa (2 = ±0.14 GPa) and 1220°C (2 = ±16°C), though a few glasses yield higher temperature and pressure values. Because the data are tightly clustered, the standard errors of the mean for the average pressure and temperature are very small (~1 MPa and 0.1°C). It should be noted that these calculations are based on the assumption that the MORB glasses could be saturated in olivine, orthopyroxene, clinopyroxene, and plagioclase at some pressure and temperature and does not prove that they were in equilibrium with all these phases at any time. Nevertheless, the more primitive basalt glass compositions are close to Fe/Mg exchange equilibrium with the minerals in the troctolite, whose composition can be inferred from the whole-rock Mg#. Because some of the glasses are very primitive, we conclude that they probably were close to saturation in a mantle phase assemblage.

Thus, we believe that the data in Figure F53 indicate that the Site 1275 troctolites probably formed in equilibrium with a melt at 0.40–0.68 GPa. Because the orthopyroxene, clinopyroxene, and plagioclase in these rocks clearly crystallized from a melt, it follows that conductive cooling and crystallization of rising magma began at a depth of 12–20 km beneath the Mid-Atlantic Ridge, at least in this region. This result is consistent with thermal models for slow-spreading ridges, which predict that the thermal boundary layer extends to depths of ~20 km beneath the Mid-Atlantic Ridge (Braun et al., 2000; Reid and Jackson, 1981; Sleep, 1975). This result is also consistent with fractional crystallization models, which require pressures of 0.4–0.6 GPa to account for observed variation in the composition of genetically related basalts from the Mid-Atlantic Ridge in general (Grove et al., 1992; Meurer et al., 2001; Michael and Chase, 1987) and from the 14° to 16°N region in particular (C. Xia et al., unpubl. data). Finally, in keeping with the observation that residual mantle peridotites are extensively exposed in outcrop in the 14° to 16°N region, this result reminds us that rocks that were incorporated into the plate at great depth have been tectonically uplifted and unroofed on the seafloor. And finally, since the troctolites lack evidence for viscous deformation, we can infer that at depths <12–20 km the tectonic exhumation process involves localized deformation at a temperature less than ~1220°C.

More evolved glass compositions (molar Mg# = 50%–60%) from basalts sampled between 14° and 16°N yield estimated pressures ranging from ~0.6 to ~0.3 GPa and temperatures ranging from 1220° to 1175°C. As can be seen in Figure F53, the pressure-temperature estimates for the evolved glasses form a tight cluster at significantly lower pressures and temperatures compared to the cluster for pressure-temperature estimates for more primitive glasses. Although further analysis is necessary, this suggests that crystallization of melt in the thermal boundary layer beneath the Mid-Atlantic Ridge is a polybaric process in which increasingly evolved melts crystallize at increasingly shallow depths.

We turn now to the question of whether the olivine in the troctolites is igneous or is relict, residual olivine from mantle peridotite. If the olivine is igneous, the troctolites would be the most primitive igneous cumulates yet recovered from a mid-ocean ridge. If the olivine is derived from mantle peridotite, then the troctolites should be viewed as impregnated dunites, similar to those sampled at Site 1271 and to impregnated peridotites dredged from the mid-ocean ridges and common in some ophiolites. In thin section, some spatially isolated olivine crystals appear to be optically continuous with other nearby olivine grains, though they are now separated by intervening pyroxene and plagioclase. Thus, some rocks may preserve an olivine lattice-preferred orientation, despite the undeformed nature of the pyroxene, plagioclase, and alteration assemblages. If so, this would be consistent with a polygenetic origin for the rocks, composed of older, previously deformed olivine grains that have been partially disaggregated and incorporated in a matrix of younger igneous minerals.

In "Site 1271" we argued that covariation of olivine Mg# and Ni contents can be used to distinguish some impregnated dunites from cumulate igneous dunites. Whereas olivine Mg#s are similar to typical values for mantle peridotites, olivine cumulates cannot be distinguished from mantle dunites. However, where the Mg#s are lower, in the range of 88%–85%, then high Ni contents in olivine at a given Mg# are indicative of an origin via melt/rock reaction and impregnation (e.g., Godard et al., 2000; Kelemen, 1986, 1990), whereas lower Ni contents at the same Mg# are consistent with an origin via fractional crystallization (e.g., Hart and Davis, 1978). Because the Site 1275 troctolites are highly altered (some contain up to 9 wt% CO₂), it is difficult to be sure about the composition of the remaining olivine. However, we have inferred the olivine compositions, based on the bulk compositions of samples with 0.2–3 wt% CO₂. To do this, we calculated CIPW norms for the anhydrous bulk composition. These yielded normative proportions ranging from 57 to 78 wt% olivine (with four of six compositions between 70 and 72 wt%), 11 to 30 wt% pyroxene (four of six between 11 and 15 wt%), and 4 to 12 wt% plagioclase (four of six from 9 to 12 wt%). 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 calculated Ni concentrations in olivine are plotted in Figure F54, together with similar estimates for Ni in olivine from mantle harzburgites and dunites recovered at other sites during Leg 209 and with published olivine compositions from dunites, impregnated dunites, and wehrlites in the mantle transition zone of the Oman ophiolite.

It is apparent that the estimated Ni contents of olivine in the Site 1275 troctolites are higher than those predicted for crystal fractionation of olivine. This is consistent with the hypothesis that the troctolites are the products of melt-rock reaction, in which Fe/Mg exchange with migrating melt lowered olivine Mg#s but the original high Ni concentrations were retained in the olivine. Therefore, we tentatively infer that the Site 1275 troctolites are impregnated dunites, containing relict, residual olivine derived from mantle peridotites plus pyroxenes and plagioclase that crystallized from a cooling melt migrating along olivine grain boundaries.

Petrogenesis of Gabbroic Rocks and Diabase

Gabbroic rocks, including gabbro, gabbronorite, oxide gabbro, oxide gabbronorite, and olivine gabbro, compose ~74% of the core from Site 1275. Diabase composes an additional 10% of the core, particularly in a thick interval at the top of Hole 1275B. The gabbroic rocks are characterized by large variations in mineral proportions and textures on short length scales (Fig. F55), such that the longest interval with homogeneous grain size and mineral proportions is only a few meters. Contacts between contrasting lithologies are generally sharp and planar, though "crenelated" contacts are also observed in dozens of intervals (Fig. F56). Diabase contacts with gabbroic rocks are very sharp. However, there is a gradation in grain size between coarse-grained diabase and fine-grained gabbroic rocks. Thus, the range of grain sizes appears to represent a continuum and we tentatively infer that the gabbroic sections were constructed by multiple intrusions that continued over a range of temperatures and cooling rates. Although steep contacts are locally observed, the great majority of contacts between contrasting gabbro and diabase units have dips <30°, so that—unless the section has undergone tectonic rotations of >30°—it is probable that most of the intrusions were subhorizontal sills.

Diabase and fine-grained gabbros have compositions that resemble basaltic liquid compositions. If they are liquid compositions, the melt was relatively primitive, with Mg# = ~60%. This is striking, because the coarser gabbroic rocks from Site 1275 are probably cumulates formed by crystallization of substantially more evolved melts. We interpret these rocks as cumulates because they have low incompatible element abundances (e.g., low Zr) compared to evolved MORB at the same, or even higher, Mg#. However, there is little sign of the complementary evolved melts that formed these cumulates, either in core from Site 1275 or among erupted lavas from the 14°–16°N region.

Granophric dikes and veins are a minor but ubiquitous part of the gabbroic core. These might conceivably represent melts in equilibrium with the evolved cumulate gabbros, but if so there must be a much larger volume of lava or intrusive rocks with the same composition somewhere else. The granophyres have sharp to completely gradational contacts with host gabbros and probably represent small amounts of highly evolved melt that were extracted from the gabbros in their last stages of crystallization and then ascended to intrude other gabbros higher in the section.

Granophyres locally show complex mingling relationships with host gabbros, and even host diabase. Similarly, some contacts between contrasting gabbroic units also show complex mingling and some fine-grained gabbroic units contain coarse pyroxene crystals that are tentatively interpreted as xenocrysts derived locally from coarser gabbroic host rocks. Thus, the multiple intrusions that constructed the Site 1275 gabbro section occurred while earlier crystallized gabbroic rocks were still partially molten. This process of repeated intrusion led to mingling and, probably, to magma mixing.

Olivine gabbros were only observed in lower portions of Hole 1275D. Unfortunately, these samples were recovered after the shipboard geochemistry laboratory closed for transit to Bermuda. Thus, it is not certain whether these rocks are primitive cumulates, liquidlike compositions, or simply evolved gabbros including olivine.

Aside from the olivine gabbros, the diabases, and one or two fine-grained gabbros with high Mg#s, the gabbroic rocks from Site 1275 that we analyzed have evolved compositions with low Mg# and high incompatible element concentrations. Such evolved compositions appear to be typical of gabbroic intrusions into peridotite that have previously been sampled by dredging and submersibles along the mid-ocean ridges (e.g., Cannat, 1993; Cannat et al., 1992). The common observation of evolved gabbroic intrusions into mantle peridotite poses an interesting conundrum. Evolved lavas with compositions that could crystallize such iron- and titanium-rich gabbroic cumulates have not been sampled in the 14° to 16°N area. By the same token, basalt compositions from the Mid-Atlantic Ridge in this area record fractional crystallization of a relatively primitive cumulate assemblage.

In fact, this problem is more general. As shown in Figures F57 and F58, truly primitive cumulates are rare among oceanic samples, even in the thick gabbroic sequences at ODP Sites 735, 921, and 923 that have been interpreted as parts of a lower crustal gabbro section beneath sheeted dikes in "typical" layered oceanic crust (e.g., Cannat et al., 1995; Dick et al., 2002; Natland and Dick, 2002). Because average mid-ocean-ridge basalts have about twice as much zirconium and other incompatible elements as primitive mid-ocean-ridge basalts, the average lavas record fractional crystallization of ~50% mafic cumulates, of which at least one-half must have Mg#s > 70%. In contrast, the average Mg# of rocks recovered from Sites 735, 921, and 923 is 67%–71%. It is clear that a substantial proportion of more primitive cumulates has formed beneath the mid-ocean ridges. However, these primitive cumulates have rarely been sampled.

During Leg 209, only the gabbronorites from Site 1268 have the high Mg# and low incompatible element concentrations required for the primitive cumulates complementary to fractionation of typical mid-ocean-ridge basalts. This is intriguing because, as for the troctolites at Site 1275, the Site 1268 gabbronorites include igneous orthopyroxene. Mid-ocean-ridge basalts are saturated in orthopyroxene only at subcrustal pressures (see discussion of troctolite genesis, above). As we outlined in "Site 1268," if the Site 1268 gabbronorites formed from primitive basalts, they must have crystallized at relatively high pressure (perhaps ~0.3–0.4 GPa) within an upper mantle thermal boundary layer beneath the Moho. However, we caution readers that the Site 1268 gabbronorites are highly altered, and alteration may have modified the bulk rock Mg#, so that full evaluation of this hypothesis awaits shore-based analysis of relict clinopyroxene crystals.

The depth of crystallization of the relatively evolved gabbroic rocks at Site 1275 is not well constrained. If the earliest formed gabbroic rocks are temporally related to the melts that crystallized impregnations of plagioclase and pyroxene in the troctolites, then they must have crystallized at depths of 12–20 km, driven by slow conductive cooling near the base of the thermal boundary layer, with initial magmatic temperatures of ~1200°C. The aphanitic texture of the diabase intrusions, on the other hand, suggests that they cooled relatively quickly. Since some diabases show mingling with hornblende-bearing granophyre, these diabases must have intruded while the host rocks were still above the fluid-saturated granite/trondjhemite solidus at ~700°C. Other massive diabases, particularly those at the top of Hole 1275B, may have intruded even colder wallrocks, perhaps at shallow depths. Again, we infer that the igneous construction of the gabbroic section took place via multiple intrusions, with wallrock temperatures varying over time from ~1200°C to <700°C. Probably, this process also took place over a range of depths, beginning at 12–20 km and ending near the seafloor. Alternatively, if the crystallization of the gabbroic rocks at Site 1275 was not temporally related to the impregnation of the troctolites, then crystallization of the gabbroic section could have taken place over a more restricted range of temperature and pressure. We hope that oxide-silicate thermobarometry on oxide gabbro and gabbronorite samples from Site 1275 will help to distinguish between these different possibilities.

Hydrothermal Alteration, Metamorphism, and Metasomatism

Metasomatic Changes in Troctolite

As for carbonate-altered peridotites from previous sites on this leg, some troctolites from Site 1275 include abundant carbonate veins. CO₂ contents in these rocks range up to 9 wt% and are correlated with CaO contents. As was proposed in "Site 1274" in the discussion of metasomatism of altered peridotites, it could be that the troctolites have been modified by metasomatic addition of Ca as well as CO₂. However, this is unlikely. The troctolites contain substantial Al₂O₃, abundant plagioclase and plagioclase pseudomorphs and only minor amounts of spinel. Thus, most of the Al₂O₃ in these rocks must have originally been in plagioclase. If so, CaO is required to form the anorthite component in the plagioclase. CIPW norms for the anhydrous, CO₂-free compositions of the troctolites include very little clinopyroxene, and some have trace amounts of normative corundum. Thus, most or all of the calcium currently present in the troctolites is required to balance the Al₂O₃ in plagioclase. As a result, we tentatively conclude that CO₂ was added to the troctolites during hydrothermal metasomatism but that calcium was relatively immobile during this process.

Mainly Static Alteration

Alteration intensity in Holes 1275B and 1275D decreases with increasing depth. Cataclastically deformed gabbroic rocks and diabase and a talc amphibole schist with an ultramafic protolith are abundant in the first five cores of Hole 1275D, and these show 50%–100% amphibolite to greenschist facies alteration. The cataclastic deformation is probably related to the fault(s) that formed the top of the bathymetric dome at Site 1275. However, below this fault-related cataclastic deformation, hydrothermal alteration of the troctolites and gabbroic rocks at Site 1275 took place almost entirely under static conditions along microcracks that lack a systematic preferred orientation and possibly formed during cooling. This static alteration is very different from what has been observed in the thick gabbroic sections at ODP Sites 735, 921, and 923, where alteration intensity is highest in zones of penetrative crystal-plastic deformation that occur throughout the section.

Structural Geology

Crystal-plastic deformation is rare in samples from Site 1275. Only a few gabbros show evidence for dynamic recrystallization of plagioclase. Evidence of minor strain, in the form of bent plagioclase with undulatory extinction and deformation twinning, is present throughout Holes 1275B and 1275D, but this reflects minor high-temperature deformation of the type that may be common during igneous compaction of gabbroic intrusions.

Brittle deformation, in the form of cataclastic breccias and semibrittle schists, is common in the upper 15 m of Hole 1275B and the upper 50 m of Hole 1275D. The difference in depth of cataclastic deformation in the two holes may be related to the greater thickness of relatively weak, altered troctolite in Hole 1275D compared to Hole 1275B. Cataclastic breccias are commonly (always?) developed from gabbroic protoliths, whereas semibrittle schists have mainly ultramafic protoliths. The intensity of brittle deformation decreases markedly downhole in Holes 1275B and 1275D. This is consistent with the hypothesis that the brittle deformation is mainly related to the fault zone that likely forms the top of the topographic dome at Mt. Mike.

Two fascinating observations are that there are fine-grained diabase clasts—and perhaps even chilled diabase material within the matrix—in some cataclastic breccias and that diabase intrusions form deformed lenses, possibly preserving chilled margins, within semibrittle schists. These observations— whose interpretations are still uncertain—may be consistent with the presence of a thick and largely undeformed diabase intrusion in the top 18 m of Hole 1275B. We suggest that diabase sills intruded along the fault zone at the top of Mt. Mike, penecontemporaneously with continued slip on the fault.

Metamorphic veins in core from Site 1275 do not show any preferred orientation and thus may have followed cracks that formed as a consequence of nearly hydrostatic stresses during cooling of the igneous rocks.

Magmatic banding or layering in gabbroic rocks is defined by sharp contacts between lithologies with different proportions of minerals and/or different grain sizes and by plagioclase shape fabrics. This magmatic banding is subhorizontal. Contacts between granophyres and gabbroic host rocks are irregular, but a significant subset have steep dips, >70°. Since the statistical probability of sampling steeply dipping features in a vertical drill hole is very small compared with the probability of intersecting equally abundant horizontal features (i.e., features with the same width and density), we can infer that near-vertical granophyre veins are much more abundant within the Site 1275 gabbroic rocks than subhorizontal granophyre veins. Because the evolved melt that formed the granophyre was much less dense than the host gabbros, one possible interpretation is that the steeply dipping granophyres crosscut the gabbros as near-vertical dikes and veins (as, for example, in the "transgressive granophyres" of the Skaergaard intrusion [Wager and Brown, 1997, figs. 101 and 105 and accompanying text; Hirschmann et al., 1992]) and thus that the section has not undergone large rotations around near-horizontal axes since the formation of the granophyres. If so, this places an important constraint on the direction and magnitude of possible tectonic rotations of the gabbroic section at Site 1275. Alternatively, the majority of the tranophyres with steep dips could have formed as subhorizontal "segregation sheets" (e.g., review in Boudreau and Philpotts, 2002). In this case, one would infer that the section could have undergone almost 90° of rotation around a nearly horizontal axis.

Five of six diabase/gabbro contacts are subhorizontal. If the gabbroic rocks have not undergone substantial tectonic rotations around a horizontal axis, this indicates that the diabases were intruded as sills emplaced in gabbro. In this case, many of the other subhorizontal magmatic layers could also be sills, intruded into preexisting gabbroic rocks while the growing plutonic complex was at relatively high temperatures. These observations and inferences are consistent with the idea that gabbroic sections beneath some mid-ocean ridges, and perhaps some layered intrusions, form via multiple injection of "sheeted sills" (Bédard et al., 1988; Bernstein et al., 1992, 1996; Browning, 1984; Browning et al., 1989; Kelemen and Aharonov, 1998; Kelemen et al., 1997b; Korenaga and Kelemen, 1997, 1998). Alternatively, if the steeply dipping granophyres at Site 1275 were emplaced as subhorizontal segregation sheets, then most of the contacts between diabase and gabbro and between different gabbro units probably formed with a near-vertical orientation. Whereas dominantly vertical contacts between different gabbroic units are rare in large intrusions, such structures are observed in specific plutons such as the large, shallow-level Lilloise intrusion, intruding Tertiary flood basalts in East Greenland (Chambers and Brown, 1995). And, finally, we emphasize to readers that in presenting two alternative interpretations of these structures, we do not mean to imply that this is an exhaustive list of possibilities. Instead, we simply present two ideas which seem plausible based on our preliminary data.

Paleomagnetic Data and Tectonics

Paleomagnetic data were collected on half cores and individual discrete samples. Using these data, we rotated the measured orientations of foliations, faults, veins, and dikes in individual core pieces around a vertical axis, thereby restoring core pieces to an orientation with a common azimuth for the remanent magnetization vector.

In Hole 1275B, half-core measurements and discrete samples all have negative stable inclinations, suggesting that the rocks are reversely polarized. Rotation of structural features around a vertical axis into a common remanent magnetization direction yields two important results. First, magmatic layering in the gabbroic rocks forms a relatively tight cluster on a lower hemisphere stereo plot, with an average dip of ~20°. Second, although many granophyre/gabbro contacts also have shallow dips, there is a distinct population with near-vertical dips. These are roughly perpendicular to the average attitude of layering in the gabbros. As explained in the previous section of this summary, these observations could be consistent with two different end-member scenarios. In the first scenario, we interpret the steeply dipping granophyre/gabbro contacts as defining the paleovertical direction during granophyre intrusion. In this view, the nearly horizontal present-day attitude of the gabbro layering probably indicates that these formed as subhorizontal layers. Alternatively, the steeply dipping granophyres could have formed as subhorizontal sheets, in which case the section has undergone large tectonic rotations.

The average remanent inclination for 26 discrete samples of gabbro and diabase from Hole 1275B is –9.5° (95% CI = +3.9°/–3.7°), which agrees well with data on inclination for archive half cores. The average remanent inclination for 34 discrete samples from Hole 1275D is –3.5° (95% CI = +7.0°/–6.7°), also in reasonable agreement with data on inclination for archive half cores. The values for gabbroic rocks are not statistically different.

Discrete sample data for Hole 1275B were all quite similar, but in Hole 1275D there were several samples (particularly troctolites and diabases in the upper 50 m of the hole) that have scattered but positive inclinations. One sample of troctolite contains two nearly antipodal magnetization components. These data suggest that both normal and reversed polarity magnetizations are recorded in Hole 1275D samples. Acquisition of an early reversed polarity magnetization probably occurred via cooling of igneous magnetite through 500°–600°C. Acquisition of a later normal polarity magnetization could have occurred during crystallization of secondary finer-grained metamorphic magnetite associated with serpentinization of olivine in the troctolites and/or during local reheating of fine-grained magnetite in the rocks associated with late intrusion of some of the diabase sills and dikes. Removal of troctolite and diabase samples with positive inclinations from the data set for Hole 1275D yields an average inclination of approximately –12° for the remaining samples, which is even more consistent with the data from Hole 1275B than the average for the entire Hole 1275D data set.

The average inclination of approximately –10° to 12° in gabbroic rocks from Site 1275 is significantly lower than the expected inclination of –28° for rocks with a reversed polarity magnetization. These data suggest that block rotation has affected the section sampled at Site 1275, as found for Sites 1268, 1270, 1272, 1274, and 1275, where remanent magnetic inclinations also require substantial rotation of the section. For Sites 1268 and 1270, we inferred that rotation axes were probably near horizontal and parallel to the normal faults that form steep slopes along the rift valley. At Sites 1272, 1274, and 1275, the choice of a tectonically reasonable rotation axis is less clear.

Decreasing the inclination from –28° to –11° in the gabbroic rocks from Site 1275 requires a minimum of 17° of counterclockwise rotation around a horizontal rotation axis with an azimuth of 270° or larger rotations around axes parallel to the trend of the rift valley (~020°). One option is that the section has indeed been rotated ~17° around a near-horizontal east-west axis. This is consistent with the first structural interpretation discussed above, in which the steeply dipping granophyres formed as near-vertical intrusions into horizontally layered gabbros. However, we cannot rule out the possibility that the steeply dipping granophyres were originally subhorizontal and the section has undergone large rotations, for example, around a nearly horizontal axis parallel to the rift valley walls, as inferred for Sites 1268 and 1270. And again, we emphasize that in listing two alternatives we do not mean to imply that this list is exhaustive. There may be many other interpretations that are consistent with our data.

We are confident that continuing analysis of the structural and paleomagnetic data will place additional constraints on the tectonic history of Mt. Mike. For example, the Hole 1275 data may offer the opportunity for another refinement. There seems to be a significant difference between the magnetic inclinations in gabbroic rocks from ~50 to 140 mbsf, which averages approximately –6°, and magnetic inclinations of approximately –15° in gabbroic rocks from 150 to 209 mbsf (Fig. F59). This may reflect smaller amounts of tectonic rotation in the bottom of Hole 1275D compared to the middle of Hole 1275D and the entire section sampled in Hole 1275B.