Leg 176 was devoted to the deepening and logging of Hole 735B atop a shallow bank along a transverse ridge at the Atlantis II Fracture Zone, Southwest Indian Ridge. There, 500 m of gabbro was cored 10 yr ago during Leg 118. During Leg 176, the hole was deepened more than 1000 m to a total depth of 1508 m before weather-related drill-string failure terminated coring. Fishing cleared some of the pipe from the hole, and logging was conducted in the upper 595 m. More igneous rock was recovered (some 865.99 m, representing 86.3% recovery) than on any previous Ocean Drilling Program (ODP) or Deep Sea Drilling Project (DSDP) leg. The combined results of drilling during Legs 118 and 176 make Hole 735B one of the most important accomplishments in the history of scientific ocean drilling. For the first time, a significant proportion of an all but inaccessible layer of the Earth's crust has been sampled in situ. We are now able to describe the architecture and outline the magmatic, structural, and metamorphic history of a block of the lower ocean crust, which formed at an ultra-slow-spreading ridge 11 m.y. ago.
Because of its inaccessibility, the nature of the lower ocean crust has been one of the longest standing questions in marine geology and has largely been inferred from remote sensing and analogy to on-land sections of fossil ocean crust (ophiolites). From the earliest seismic studies, the ocean crust has been believed to have an uncomplicated and uniform layered seismic structure (Hill, 1957; Raitt, 1963; Christensen and Salisbury, 1975). This structure was long equated to a simple layer-cake sequence, known as the Penrose model (Penrose Conference Participants, 1972; Coleman, 1977), of sediment, pillow basalt and diabase, and a thick gabbro section overlying the Earth's mantle, with the igneous crust/mantle boundary at or near the Moho. Accretion of the lower crust, in this model, resulted from crystallization from some form of near-steady-state magma chamber, or crystal mush zone, where magmas accumulate beneath a sequence of sheeted dikes and pillow lavas over the ascending mantle. Internally, a simple lower crustal stratigraphy was believed to exist, consisting of primitive layered gabbros overlain by more evolved isotropic gabbros, all of which were equated to seismic Layer 3.
Recent studies, however, indicate that the ocean crust has a more complex, three-dimensional structure that is dependent on magma supply and spreading rate and that large steady-state magma chambers are absent (e.g., Whitehead et al., 1984; Detrick et al., 1990; Bloomer and Meyer, 1992; Sinton and Detrick, 1992; Carbotte and Macdonald, 1992; Nicolas et al., 1996). This is particularly the case for slow-spreading ridges. Provision is now made for thinner crust near large transform faults or where half-spreading rates are significantly below 10 mm/yr (e.g., Reid and Jackson, 1981; Bown and White, 1994; Mutter and Detrick, 1984; Dick, 1989). Compilations of dredge results and seismic data show that a continuous gabbroic layer does not exist at slow-spreading ridges (Whitehead et al., 1984; Mutter et al., 1985; McCarthy et al., 1988; Dick, 1989; Cannat, 1993; Tucholke and Lin, 1994). Earlier drilling at Atlantis Bank and the Mid-Atlantic Ridge Kane Fracture Zone region (Leg 153; Cannat, Karson, Miller, et al., 1995) demonstrated that the internal stratigraphy of the lower ocean crust at slow-spreading ridges is governed as much by the dynamic processes of alteration and tectonism as by igneous processes. Finally, the abundance of serpentinized peridotite in dredge hauls from rift valley and fracture zone walls (Aumento and Loubat, 1971; Thompson and Melson, 1972; Fisher et al., 1986; Dick, 1989; Cannat, 1993) has again raised the possibility that serpentine can be a significant component of seismic Layer 3 as originally suggested by Hess (1962).
With this increasing complexity, in situ observation of the lower ocean crust by drilling is a necessity if the processes of ocean crust accretion and the nature of the ocean crust are ever to be understood. DSDP and ODP have previously sampled in situ ocean crust in a variety of spreading environments to sub-bottom depths as great as 2 km. Although this has confirmed many inferences from ophiolites as to the shallow structure and composition of the ocean crust, it also has produced some unexpected results. Despite the recovery of short sections of lower ocean crust and mantle during several ODP legs, no truly representative section of seismic Layer 3 had ever been obtained in situ, leaving its composition, state of alteration, and internal structure almost entirely a matter of inference.
Hole 735B came close to achieving this goal with a 1508-m section of coarse gabbro drilled in a tectonically exposed lower crustal section on a wave-cut platform flanking the Atlantis II Fracture Zone on the slow-spreading Southwest Indian Ridge. The new results from Hole 735B, together with those from Leg 118, substantially change our perception of the lower ocean crust at slow-spreading ridges. The additional km of drilling documents a systematic variation in igneous petrology, structure, and alteration with depth quite unlike that associated with large magma chambers or even the melt lens now inferred to exist beneath fast-spreading ridges. It provides a first assessment of synkinematic igneous differentiation in which the upper levels of the gabbroic crust are enriched in late differentiated melts by means of tectonic processes, rather than simple gravitationally driven crystallization differentiation. Given the typical 4- to 6-km thickness of seismic Layer 3, this long section does not characterize the lower crust for all ocean basins. However, if the lower crust at ultra-slow-spreading ridges is only ~2 km thick, as predicted, then it is a beginning.
Overall, the sequence of rocks sampled in Hole 735B is unlike that in a Penrose-type ophiolite or in layered intrusions. Some of its attributes, including the lack of well-developed layering, and the presence of small 100- to 500-m intrusions, are similar to structural characteristics from ophiolites believed to have formed in slow-spreading environments, such as the Trinity or Josephine ophiolites. However, several of the major features of the section have not been described from these ophiolites. These include (1) the occurrence of innumerable discrete large and small, often sheared oxide-rich gabbros intruding undeformed olivine gabbro; (2) a striking downward decrease in abundance of oxide-rich gabbros through the section; (3) a baseline igneous stratigraphy of successive small "isotropic" olivine gabbro intrusions, the chemistries of which are less primitive in the lower kilometer of the hole; (4) the apparent synkinematic igneous differentiation of the section, with a concentration of late iron-rich melts near the top, apparently intruded along faults and shear zones; and (5) abundant crystal-plastic and brittle--ductile deformation at the top of the section that decreases downward. Together with the well-known geochemical affinities of many slow-spread ophiolites for the "arc-environment," most are not very good analogs for the Hole 735B section. Although an exact on-land counterpart may not exist, there are similarities to some Ligurian ophiolites, and more careful documentation of lower crustal sections in some ophiolites may reveal more similarities.
The Hole 735B gabbros correspond closely to those dredged and drilled at fracture zones and rift valley walls at slow-spreading ridges, and this, then, would seem to confirm the long-standing inference that the accretionary processes in the lower crust are highly sensitive to spreading rate. Gabbros from the East Pacific Rise drilled during Leg 147 at Hess Deep, and gabbros in ophiolites believed formed at fast-spreading centers (e.g., Oman), have an entirely different pattern of alteration (e.g., Dick et al., 1992; Manning and MacLeod, 1996) and lack the extensive crystal-plastic fabrics of the Hole 735B gabbros. Moreover, the lack of well-defined, planar, and continuous igneous layering at Hole 735B also contrasts with layering in the lower two-thirds of the Oman section.
In basic outline, the gabbroic crust drilled over the two legs consists of five, possibly six, main blocks of relatively primitive olivine gabbro and troctolite, from 200 to 500 m thick, each with its own internal chemical and petrological coherence. Overall, each of these bodies appears to be composed of many smaller magma bodies. These probably represent small intrusions into solidified gabbro, where contacts are sharp, or into crystal mushes, where contacts are diffuse or sutured. Smaller crosscutting bodies of olivine gabbro and troctolite may represent feeder channels for melts migrating toward the surface or feeding intrusions higher in the section. Overall, each of the composite intrusions has more fractionated, oxide-bearing gabbros toward the top, and the most primitive and magnesian rocks, toward the base. Thus, each composite body likely represents a major cycle of magmatic intrusion and in situ differentiation as the melts worked their way upward through the section.
Although the boundaries of each of these principal intrusions are fairly well defined, the relationships of one to another are neither simple nor consistent. The uppermost intrusion is separated from the one underneath by a massive zone of Fe-Ti oxide-rich gabbro along a zone of shear. There is also a small, but abrupt, deformational offset between the lower two. There is a zone of brittle faulting near the contact between the second and third masses, but it is not at the contact. Instead, the contact, which is actually distributed over a zone of several tens of meters, appears to represent a series of crosscutting lithologies and to be intrusive in character overall. The presence of multiple composite intrusions, with large-scale repetitions of magmatic-crystallization sequences, confirms that this portion of the Southwest Indian Ridge was not supplied by a steady-state magma source.
Although about 80% of the Leg 118 and 176 cores are oxide-poor gabbro or olivine gabbro, these are crossed by numerous bodies of Fe-Ti oxide-bearing and Fe-Ti oxide-rich (ilmenite and titanomagnetite) gabbro, gabbronorite, and olivine gabbro. The largest body is a 70-m-thick zone of oxide olivine gabbro between 200 and 270 mbsf. Typically, however, they are only a few centimeters to a few tens of centimeters thick, with more than 600 identified by magnetic susceptibility measurements, using a multisensor track (MST), on the Leg 176 core alone. Many of the narrow oxide-rich zones have associated veins or small dikelets of siliceous diorite, trondhjemite, and granodiorite. The oxide-rich zones (and veins) diminish downhole from some 30% of the core in the upper 500 m, to 12% of the core in the subjacent 1000 m, and to less than 1% of the core in the lowest 300 m of the hole. These oxide gabbros represent the product of extended high-iron differentiation of parental basaltic liquids, and the thick sequences near the top of the hole must reflect complex processes of segregation from a large volume of rock. The uniform depletion of most of the Hole 735B cores of Ti and highly incompatible elements suggests that Fe-Ti rich liquids were expelled from their crystalline matrix by the interactive processes of crystallization and compaction in a dynamic environment. The tendency of oxide gabbros to be associated with zones of crystal-plastic deformation, and the presence of late magmatic oxides locally cross-cutting crystal-plastic fabrics, suggests that the processes of aggregation and transport of the late magmatic liquids were related to deformation and faulting of the gabbros before solidification.
Our structural observations demonstrate that crustal accretion at this ultra-slow-spreading ridge was strongly influenced by localized deformation from magmatic to low-temperature cataclastic conditions. High-temperature metamorphic effects are transitional to magmatic pro-cesses, and some, perhaps many, rocks were deformed and recrystallized while still partly molten. Some of the most striking zones of crystal-plastic deformation, however, apparently formed under the equivalent of granulite facies metamorphic conditions (>800º-1000ºC), when there was little or no melt present. Although 77% of the gabbro contains no macroscopic magmatic or deformation fabric, and long sections of the core may be undeformed, there are numerous magmatic, crystal-plastic, and brittle deformation features, with a clear decrease in intensity vertically downward. Above 500 m the dominant sense of shear is normal, whereas below this there are several zones with numerous reverse-sense shears. A weak, subparallel, crystal-plastic fabric, which may record a transition from magmatic to crystal-plastic deformation, commonly overprints magmatic foliations. Many of the deformed rocks also show a continuum between crystal-plastic and brittle behavior. There are some narrow zones of intense cataclasis and several faults, two of which coincide with Leg 118 vertical seismic profile reflectors.
Overall, metamorphism and alteration of the Hole 735B cores can be portrayed as having occurred in two stages. Initially, alteration occurred under dynamic conditions, ranging from granulite to amphibolite, with only minor greenschist facies, beneath the rift valley and at the beginning of uplift and unroofing. This was followed by a lower temperature sequence, under relatively static conditions, representing block uplift to the summit of the transverse ridge and subsequent cooling to the present day. Whereas hydrothermal alteration to amphibole and sodic plagioclase is locally extensive in shear zones near the top of the section, it is minor below 600 m. Overall, static high-temperature alteration in undeformed gabbros is patchy, rather than pervasive. Extensive intervals (>300 m) have less than 10% background alteration, and the gabbro is generally very fresh. Whereas greenschist facies alteration is minor, likely reflecting rapid cooling from amphibolite-facies conditions during block uplift, lower-temperature alteration is more prominent. This is represented by the local formation of abundant late smectite, carbonate, and zeolite-prehnite veins, with iron oxyhydroxides, particularly in a zone of intense alteration between 500 and 600 mbsf. Deeper in the hole, there are also locally abundant smectite-lined fractures associated with sulfides, indicating formation under non-oxidative, low-temperature conditions.
Rock magnetic measurements show a consistent average stable inclination of ~71º in Hole 735B, with only minor downhole variation. The rocks have been tilted ~20º but have a very stable remanent magnetization and often very sharp blocking temperatures, suggesting relatively rapid acquisition of thermoremanence during cooling. This cooling occurred while the rocks were beneath the rift valley and relatively shortly after they crystallized. They thus constitute an ideal source for marine magnetic anomalies, and a gabbroic layer as thick as our drilled section is probably sufficient to account for the marine magnetic anomaly at Atlantis Bank over Hole 735B. Analysis of structures reoriented magnetically indicates that the dominant foliation dips preferentially toward the axial rift to the north.