Site 1271 is located along the track of Faranaut 15°N Dive 7 on the western flank of the Mid-Atlantic rift valley. The dive, using the Nautile submersible, recovered basaltic samples at water depths >3700 m and three samples of peridotite and two samples of gabbro higher on the rift valley walls. Hole 1271A was initiated on a smooth, sedimented slope uphill from gabbro sample FR07-10. After a relatively promising beginning, relatively high penetration rates and poor recovery from 13 to 28 mbsf suggested that we had encountered a near-surface fault zone, similar to those encountered at Site 1270. We continued to drill to 44.8 mbsf, but the hole was unstable and ultimately had to be abandoned. Recovery averaged ~13%. We then moved 74 m southwest and spudded Hole 1271B into a similar sedimented slope. Although drilling conditions were far from ideal, we were able to drill to a depth of 104 mbsf with 15% recovery. At this point, because of poor recovery combined with great lithologic diversity, we decided to log the hole, anticipating that density and resistivity contrasts in downhole data would help us to map the distribution of gabbroic intrusions and peridotite host rocks. However, even after the bit was released, it was very difficult to raise the drill string. Apparently, the hole collapsed as the pipe was withdrawn. The logging string encountered obstructions just a few meters below the base of the open pipe, ending operations at Site 1271.
The igneous and residual mantle protoliths of recovered core were dunite, harzburgite, troctolite, gabbro, minor gabbronorite, and a small but significant amount of chromitite in three 1- to 3-cm-thick lenses (Fig. F31). Gabbroic rocks are intrusive into the peridotite in Holes 1271A and 1271B (Fig. F32). The process of igneous intrusion was probably variable; brown amphibole gabbros apparently were injected into sharp-sided dikes, whereas olivine gabbros, troctolites, and impregnated peridotites apparently crystallized from melt migrating along peridotite grain boundaries.
Among igneous and mantle rocks, dunite is abundant in Holes 1271A (98%) and 1271B (56%), whereas harzburgite composes <1% of recovered samples from Hole 1271A and only 9% of the rocks from Hole 1271B. Brown amphibole gabbro composes <2% of the igneous and mantle material in Hole 1271A and 21% in Hole 1271B, whereas olivine gabbro and troctolite compose 14% of igneous and mantle rocks recovered from Hole 1271B. Weighting the proportions of rocks in Holes 1271A and 1271B by the length of each hole yields overall proportions of 68% dunite, 6.5% harzburgite, 15% brown amphibole gabbro, and 10% olivine gabbro and troctolite. Although the ratio of dunite to harzburgite at Site 1271 is much higher than that observed at Sites 1268 and 1270, the ratio of peridotite (75%) to gabbroic rocks (25%) at Site 1271 is remarkably similar to the proportions in Hole 1268 (74% peridotite/26% gabbroic rocks), in Hole 1270 (68% peridotite/32% gabbroic rocks), and recovered from dives and dredging along the Mid-Atlantic Ridge from 14° to 16°N (75.5% peridotite/24.5% gabbroic rocks).
With some notable exceptions, peridotites from Site 1271 are 100% altered, mainly to serpentine and brucite. Many peridotites—both harzburgites and dunites—contain 1%–15% interstitial gabbroic material interpreted as the crystallization products of melt migrating through the peridotite along grain boundaries. However, despite this "impregnation," geochemical analyses of all but one of the Site 1271 peridotites show low concentrations of nominally immobile incompatible elements such as Al, Sc, and V. For example, the Al2O3 concentrations in all but one of the Site 1271 peridotites are <1 wt% (average = 0.9 wt%), whereas Al2O3 concentrations in peridotites from Site 920 on the Mid-Atlantic Ridge at 23°20'N range from 1 to 2 wt% (Casey, 1997). Thus, despite the presence of gabbroic impregnations, the average Al2O3 concentration in 8 peridotites from Site 1271 is low, though not as low as the average of 0.6 wt% Al2O3 in 7 peridotites from Site 1270 and 20 peridotites from Site 1268. If Al2O3 concentrations have not been modified by hydrothermal metasomatism, then the peridotites from Sites 1268, 1270, and 1271 are among the most depleted residual mantle peridotites yet obtained from the mid-ocean ridges.
Molar Mg# of the peridotites from Site 1271 ranges from 85% to 89% in five dunites and from 90% to 92% in three harzburgites. Most of the Site 1271 dunite Mg#s are lower than Mg#s in most harzburgites from Sites 1268, 1270, and 1271 and below the normal range of whole-rock Mg# in residual mantle peridotites from mid-ocean ridges (~89%–92%; average = ~90%–91%) (e.g., Dick, 1989). The low Mg# in dunite could be due to metasomatic changes in Fe/Mg during the hydrothermal metasomatism, olivine crystal fractionation, or reaction of migrating melt with residual mantle peridotites.
Figure F33 shows that two dunites have bulk compositions very close to the composition of pure olivine. Other dunites lie along mixing lines between olivine and pyroxene, as do two of the harzburgites. Still other dunites, including a sample of impregnated dunite with 15% interstitial gabbroic material (probably composed of plagioclase and pyroxene ± hornblende, prior to alteration), show significant compositional shifts away from olivine-pyroxene mixing lines toward mixtures including plagioclase ± Cr-Al spinel ± igneous amphibole. Because of the compositional shifts resulting from the presence of interstitial gabbroic material in most peridotite samples from Site 1271, it is difficult to discern if the peridotites were also modified by subsolidus metasomatism. Nonetheless, there is no compelling evidence for SiO2 gain or MgO loss from these samples, so we tentatively rule out metasomatic changes as the cause of the low Mg#s in Site 1271 dunites.
We now wish to distinguish between olivine crystal fractionation and melt-rock reaction as explanations for the composition of Site 1271 dunites. Figure F34 shows that Site 1271 dunites have Ni contents in olivine that are comparable to those in harzburgites from Sites 1268, 1270, and 1271, despite the fact that Site 1271 dunites have lower Mg#s than the harzburgites. Similar relationships are seen in dunites from the crust–mantle transition zone in the Oman ophiolite (Godard et al., 2000; Koga et al., 2001; Korenaga and Kelemen, 1997). Nearly constant concentrations of compatible elements accompanied by decreasing Mg# are the hallmark of dunites formed by reaction between relatively low Mg#, migrating melt, and residual mantle olivine (e.g., DePaolo, 1981; Kelemen, 1986; Kelemen et al., 1998a; Navon and Stolper, 1987). In contrast, olivine crystal fractionation leads to a steep decline in olivine Ni contents with decreasing Mg#, as seen in Figure F34. Therefore, we conclude that the Site 1271 dunites are the product of reaction between residual peridotite and migrating melts with relatively low Mg#. The low Mg# of the migrating melts probably indicates that they were undergoing gradual conductive cooling, with decreasing magma mass due to reaction and crystallization near the base of the thermal boundary layer beneath the Mid-Atlantic Ridge.
To our knowledge, the three chromitites recovered at Site 1271 are the first to be sampled from the Mid-Atlantic Ridge. In fact, Site 1271 is only the second chromitite locality that has been found along any mid-ocean ridge, the first being near the East Pacific Rise at Hess Deep at ODP Site 895 (Arai and Matsukage, 1996). The recovery of chromitites at Site 1271 is an important result. The chromitite recovered from Hess Deep is a single, elongate, discontinuous train of chromite grains <1 cm wide. It resembles chromite "trails" observed in dunites worldwide almost as much as it resembles massive chromitites in ophiolites. In contrast, the Site 1271 chromitites are rounded rather than elongate, have sharp contacts with surrounding peridotite, and are clearly massive (Fig. F35).
It has been proposed that the formation of chromitite occurs only during subduction-related arc magmatism because the formation process requires hydrous magmas (Matveev and Ballhaus, 2002). Based on our results, confirming the Hess Deep observation, it is clear that chromitites are not restricted to arc magmatism. However, there is abundant high-temperature amphibole in core from Site 1271, some of which could be igneous. If the presence of amphibole is genetically related to the formation of chromitite at Site 1271, then Matveev and Ballhause (2002) may be correct in inferring that hydrous magmas are important in chromitite genesis.
Based on the observation of unusually high molar Cr/(Cr + Al), or Cr#, in spinels from harzburgite and dunite dredged from the Mid-Atlantic Ridge in the 14° to 16°N region (Bonatti et al., 1992; Dick and Kelemen, 1992; Sobolev et al., 1992b), together with the general observation that spinels in chromitites have higher Cr# than spinels in residual mantle peridotites (e.g., Dick and Bullen, 1984), we anticipate that the Cr#s in the Site 1271 chromitites will be the highest yet observed in spinel from mid-ocean ridges. High Cr#s would provide insight into the processes that form chromitite. In addition, such a result would change the interpretation of ophiolite provenance. It has been proposed that spinel Cr#s >60 are only found in mantle peridotites from subduction-related settings (e.g., Dick and Bullen, 1984), and high Cr#s have been cited as evidence that most ophiolites with high Cr# in mantle spinels do not form at normal mid-ocean ridges.
There appear to be two types of gabbroic rocks in core from Site 1271. Brown amphibole–bearing metagabbros (termed "BAG") with relicts of igneous plagioclase (and locally some relict clinopyroxene) always contain ~50% plagioclase and 50% mafic minerals in their igneous protolith (Fig. F36). Contacts of BAG with peridotite were not recovered, but its consistent phase proportions suggest that the BAG forms dikes with sharp contacts in peridotite. In contrast, olivine gabbros and troctolites (>15% plagioclase, pyroxene, amphibole, and alteration products derived from these phases) are gradational into host peridotites with 1%–15% interstitial gabbroic material. As a result, we interpret much of the olivine in olivine gabbros and troctolites as xenocrysts, derived from a partially disaggregated residual mantle protolith.
The origin of amphibole in the BAG is uncertain. Some of the amphibole is tremolite-actinolite, clearly metamorphic in origin, but other crystals appear to be idiomorphic brown hornblende and could be igneous (Figs. F36, F37). We analyzed a fragment of this amphibole obtained from a coarse-grained sample of the BAG. Its chemistry indicates that the amphibole is a hornblende solid solution composed of ~71 mol% tremolite-actinolite and ~29 mol% pargasite. In detail, the chemical analysis combined with stoichiometric constraints suggests that there is also ferric iron in a hastingsite component. If this amphibole formed by subsolidus reaction of igneous minerals with H2O, it could have formed from olivine, pyroxene, and plagioclase via reactions such as
14 H2O + 6 (Mg, Fe)2SiO4 + 15 (Mg, Fe)2Si2O6 + 24 Ca(Mg, Fe)Si2O6 + 4 CaAl2Si2O8 + 4 NaAlSi3O8 =
4 NaCa2(Mg, Fe)4AlSi6Al2O22(OH)2 + 10 Ca2(Mg, Fe)5Si8O22(OH)2.
Alternatively, this type of amphibole could be igneous. However, the large proportion of tremolite-actinolite component in the solid solution suggests a relatively low, near-solidus temperature of crystallization.
A fragment of altered white material believed to be relict plagioclase and alteration products replacing plagioclase were analyzed separately, yielding a composition close to that of plagioclase with 81 mol% anorthite. However, the composition includes appreciable Mg, Fe, and Ti and 0.5 wt% K2O, so the sample probably incorporates minerals other than plagioclase and plagioclase alteration products.
One BAG was analyzed for major and trace element contents. It has Mg# = 84, higher than that in any of the gabbroic rocks from Sites 1268 and 1270 (maximum = 83% at Site 1268). This Mg# is indicative of crystallization from a relatively primitive melt, close to Fe/Mg equilibrium with residual mantle peridotite. At the same time, the BAG has higher incompatible element concentrations than any gabbroic rocks from the previous sites. For example, the BAG has Y = 48 ppm and Zr = 123 ppm, compared with a maximum of Y = 17 ppm and Zr = 48 ppm in gabbroic rocks from Sites 1268 and 1270. Although the high Y and Zr contents may be due, in part, to the (possible) presence of igneous hornblende in the BAG, it is likely that the parental melt also had high Y and Zr contents. The combination of high Mg# together with high incompatible element concentrations in magmas is the signature of decreasing melt mass resulting from partial crystallization combined with buffering of the Mg# at high values from reaction of melt with residual mantle peridotite (e.g., DePaolo, 1981; Kelemen, 1986). Thus, we infer that the melt parental to the BAG was derived from a primary basalt but was modified by conductive cooling, partial crystallization, and reaction with peridotite wallrocks.
We also analyzed an olivine gabbro from Hole 1271B. It has Mg# = 67, rather low for a rock that is thought to include xenocrysts of mantle olivine. If our interpretation of these rocks as hybrids is correct, then this sample must have formed with a relatively high melt/rock ratio.
Many peridotites from Site 1271 are completely altered to serpentine. Unlike the SiO2-rich metaperidotites from Sites 1268 and 1270, the compositions of metaperidotites from Site 1271 are similar to the compositions of fresh dunites and harzburgites and thus do not require large metasomatic increases in Si/(Mg + Fe) during alteration (Fig. F33). Brucite is an important accessory phase in the background alteration of many peridotites from Site 1271. The presence of brucite probably reflects the olivine-rich, low Si/(Mg + Fe) dunite protoliths for most Site 1271 peridotites (see Fig. F33). Brucite is not stable in bulk compositions with molar Si/(Mg + Fe) greater than that in serpentine (0.67). Most peridotites from Sites 1268 and 1270 have Si/(Mg + Fe) greater than 0.67, consistent with observed low-temperature alteration to serpentine + talc rather than serpentine + brucite. Lack of pyroxene in dunite protoliths may have allowed fluid compositions to reach low aqueous Si activities and relatively high pH compared to fluid compositions during hydrothermal alteration of harzburgites at Sites 1268 and 1270. These fluid compositions favor formation of brucite + serpentine.
Spinel in many metaperidotites from Site 1271 is surrounded by chlorite. The chlorite could be a hydrothermal alteration product of plagioclase, with the implication that crystallization of melt, migrating along peridotite grain boundaries, formed plagioclase halos around spinel crystals. However, chlorite around spinel may also have formed via subsolidus reactions involving olivine, pyroxene, spinel, and fluid, such as
Similar chlorite rims around spinel are common in amphibolite-grade metaperidotites worldwide, even in the absence of plagioclase.
A strikingly large proportion of impregnated peridotites from Site 1271 with 1%–15% interstitial gabbroic material contains fresh olivine and pyroxene. In this way, they resemble the relatively fresh peridotite along contacts with gabbroic veins in cores from Holes 1270C and 1270D. The presence of relict clinopyroxene and plagioclase may have buffered fluid compositions to low pH, inhibiting the formation of serpentine from adjacent olivine and orthopyroxene.
Interstitial gabbroic material in impregnated peridotites, olivine gabbros, and troctolites from Site 1271 includes a large proportion of high-temperature hornblendic amphibole. As for the amphibole in the brown amphibole gabbros, it is not clear whether any of the amphibole in these other lithologies is igneous. Sharp contacts between olivine and high-temperature amphibole in some samples suggest that at some point the two phases were in equilibrium. Sharp contacts between olivine and plagioclase in peridotites, olivine gabbros, and troctolites likewise suggest that olivine and plagioclase were in equilibrium, probably at igneous temperatures. However, spinel + chlorite(?) symplectites separate olivine from amphibole and olivine from plagioclase in other parts of the core.
In addition to the presence of high-temperature metamorphic or igneous amphibole, gabbroic rocks at Site 1271 underwent secondary alteration to tremolite-actinolite, talc, chlorite, sericite, metamorphic plagioclase, and, locally, quartz. In the brown amphibole gabbros, rutile intergrown with quartz appears to replace Fe-Ti oxides. The presence of rutile rather than sphene suggests that this replacement occurred under amphibolite rather than greenschist facies conditions.
Enstatite in orthopyroxene-poor harzburgites and orthopyroxene-bearing dunites from Site 1271 has transitional protogranular to porphyroclastic textures. In some particularly intriguing examples, elongate porphyroclastic orthopyroxene crystals showing incipient recrystallization are rimmed by undeformed vermicular spinels. This texture suggests that high-temperature exsolution of spinel or formation of spinel during reaction of migrating melt with orthopyroxene postdated deformation in these rocks.
Lower-temperature, more localized deformation at Site 1271 occurred preferentially in gabbroic veins and surrounding peridotite wallrocks, as observed at Sites 1268 and 1270. Shear zones in Hole 1271B show a well-developed progression from granulite facies recrystallization of dunite and gabbroic rocks into porphyroclastic mylonites, through amphibolite facies replacement of deformed minerals with amphibole that was itself deformed, to formation of greenschist facies, chlorite-amphibole, and serpentine schists. High-temperature shear zones, although locally crosscut by later features, apparently continued to be the primary locus of deformation throughout most of the cooling history of the Site 1271 core to temperatures <300°C.
Fault gouge and fault breccia zones sampled at several depths appear to be parts of major faults with substantial brittle offsets.
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 the case of Site 1271, split cores and discrete samples all have positive inclinations, indicating that the rocks are normally polarized. The average remanent inclination for 15 discrete samples is 25° (95% CI = +11°/–13°), and the mean inclination for archive-half cores is 29° (95% CI = +3°/–4°). These inclinations are statistically indistinguishable from the expected inclination of 28°, which is consistent with the hypothesis that no significant block rotation has affected the section sampled at Site 1271. This hypothesis is quite different from the tectonic history of Sites 1268 and 1270, where remanent magnetic inclinations also require substantial rotation of the section. Alternatively, large rotations may have coincidentally produced an inclination magnetization. For example, a 70° counterclockwise rotation about a horizontal axis along 020° would restore the inclination to 28°, and so would also be consistent with the inclination data.