In 1987, the objective of drilling a fracture zone in the Indian Ocean was largely exploratory. Over 19 years previously, the better part of 17 legs of the Deep Sea Drilling Project and the Ocean Drilling Program had been formally devoted to drilling ocean crust produced at spreading ridges but always starting at the top in basalts. This produced a great deal of information about the chemical stratigraphy of basalts at spreading ridges, their physical and magnetic properties, and their alteration. One of the main things learned, however, was just how difficult it is to core into hard, fractured, and fragmental submarine lavas from a heaving platform using equipment originally designed for drilling in sedimentary formations from stationary rigs while searching for oil. At the outset of Leg 118, several holes had been drilled to depths of >500 m into basaltic basement at various locations in the ocean basins, but only one of them had not yet come to grief because of equipment failure or the instability of the rock. Eventually even that one, Hole 504B near the Costa Rica Rift, had to be abandoned after penetrating >1800 m of basalt pillows and dikes but before truly reaching the lower ocean crust.
The majority of the ocean crust, comprising seismic Layer 3, was presumed to be gabbro overlying mantle peridotite. Such rocks, therefore, had never yet been touched except in genuinely serendipitous fashion by the drill. In many ways, abyssal gabbro contains the keys to understanding accretion of ocean crust at spreading ridges, the mechanism of igneous differentiation influencing all abyssal tholeiites, and the deep exchanges between rocks and circulating fluids that supply high-temperature hydrothermal vents. Shallow abyssal peridotite records the process of partial melting that leads to the formation of ocean crust, the mechanism of transfer of melt from the mantle to the crust, and the scale of variability underlying the parental diversity of ridge basalts. Transitions between peridotites and gabbros and between gabbros and basalts, whether as dikes or flows, are particularly crucial places to investigate because both the physical and chemical aspects of these processes change most significantly over those critical intervals. The transitions also are places to establish a firmer understanding of the seismic structure of the ocean crust.
Transverse ridges adjacent to transform faults in the Atlantic and Indian Oceans were known to be places that afforded the drill a short and in many cases immediate path into the lower ocean crust and upper mantle. Prior to Leg 118, the major morphological attributes of transform faults had been worked out at one or two places (e.g., Karson and Dick, 1983; Fox et al., 1985; Karsen et al., 1984). Transverse ridges adjacent to transform valleys also were also known as places where gabbros and peridotites are prevalent, and this was especially well established at more than a dozen fracture zones in the Indian Ocean (e.g., Engel and Fisher, 1969, 1975; Fisher et al., 1986). However, the reasons why these rocks are exposed were uncertain, and the extent to which they represent intact ocean crust was disputed (Francheteau et al., 1976). Nevertheless, gabbros and serpentinized peridotites had been recovered in the same dredge hauls at a number of places; thus, contacts between these two rock types on the flanks of a number of transverse ridges could conceivably be reached with the drill. However, none of these transitions had actually been observed by means of either a submersible or a near-bottom remote vehicle. Whether the transitions are fundamentally igneous or tectonic in origin was unknown.
The site survey in 1986 (Dick et al., 1991b) succeeded in swath mapping the entire transform domain of the Atlantis II Fracture Zone. Full gravity and magnetic profiles were also obtained, and dredging hauled up plutonic rock at numerous places. The placement of some of these hauls suggested that contacts between gabbro and serpentinized peridotite do indeed exist on the steep inward-facing walls of the transverse ridges adjoining the Atlantis II Fracture Zone. However, a small median ridge was discovered in the deep transform valley, following the trace of the transform fault itself (Fig. F1). Virtually nothing but serpentinite was obtained in several dredge hauls of this feature. Whatever the complexities of the adjacent high transverse ridges, this small ridge was evidently comprised of nothing but abyssal peridotite. Not surprisingly, the opportunity to drill into (altered) mantle rock on this small ridge became the first priority of drilling for Leg 118. Although the high transverse ridges were of interest, those who reviewed the survey data and dictated the priorities for drilling perceived them more skeptically.
The hard realities of drilling, however, sometimes dictate their own priorities. The brief box survey with the VIT camera prior to the drilling of Hole 735B during Leg 118 was actually a small but proper milestone in the scientific investigation of fracture zones. Besides documenting a superb place to drill, the survey encompassed a substantial block of massive rock, obviously large, with strong and coherent foliation throughout. Contrast this to the closely spaced normal faults observed during dive traverses of gabbro and peridotite exposed on walls of rift valleys in the North Atlantic (Mével et al., 1991; Cannat et al., 1995; Karson and Lawrence, 1997). There also, the details of rock structure are often obscured by lack of structural continuity of outcrops, the surficial cataclasis of fault surfaces, and the prevalence and sometimes confusing distribution of talus debris. For all the detailed near-bottom work that has been done on gabbros and peridotites in the North Atlantic, no place there is known even today that provides as large a potential array of drilling targets or as easy a place to core plutonic rocks as the 25-km2 flat summit of Atlantis Bank where Hole 735B was drilled.
Although foliated gabbros obtained by dredge and submersible had been described (e.g., Stroup and Fox, 1981; Malcolm, 1981; Fox and Stroup, 1982; Honnorez et al., 1984), the VIT survey of Leg 118 was the first documentation of the distribution of such rocks in outcrop. Drilling etched this story in high relief. Evidence for high-temperature and sometimes quite extensive crystal-plastic deformation was found throughout the core but most especially at the top (Robinson, Von Herzen, et al., 1989). All of the crystal-plastic deformation overlapped certainly the later stages of magmatic differentiation recorded by the rocks, perhaps even most of it, and extended into the high-temperature subsolidus (Dick et al., 1991a; Bloomer et al., 1991; Ozawa et al., 1991; Natland et al., 1991). By some means perhaps related to the deformation, gabbroic rocks of both fairly primitive and extensively differentiated compositions and mineralogy were juxtaposed in complex fashion, in ways not yet then seen anywhere else, up and down the core. Evocative but simplistic conceptions of dilating pop-bottle axial magma chambers or of infinite onions or leeks (e.g., Cann, 1974; Nisbet and Fowler, 1978) were inapplicable here (Dick et al., 1991a; Natland and Dick, 2001). Was this something strange, or something typical, of a slowly spreading ridge?
Then there was the question of exposure and uplift of the transverse ridge itself. Besides high-temperature crystal-plastic deformation, the cores of the upper 500 m of Hole 735B record a less pervasive and cooler but still substantial history of static metamorphism centered on an entirely separate series of veins and fractures in the rocks (Stakes et al., 1991; Vanko and Stakes, 1991). These are largely orthogonal to preexisting foliation. Amphiboles of various compositions, particularly, but also feldspars, pyroxenes, and clay minerals, line the fractures and replace minerals in rocks adjacent to the veins. After crystal-plastic deformation, the block of rock, although obviously still quite hot, was subject to a different stress field and it responded by cracking, not stretching, as high-temperature fluids coursed through it. Temperatures were sufficiently high, however, that all of this occurred while the block of rock was still in close proximity to a strong source of heat, that is, while it was still buried beneath the axial rift.
The compositions of saline fluid inclusions indicated that ~2 km of material was removed from atop the gabbros (Vanko and Stakes, 1991). This was not all eroded away, however, at the summit of Atlantis Bank. From the site survey, the overlying basalts and dikes clearly separated from the gabbros, evidently along a low-angle detachment fault that penetrated deep beneath the rift valley and were carried intact by seafloor spreading to the north, leaving the gabbros to migrate to the south. These then were soon exposed on the southern rift valley wall and shortly thereafter lifted up to sea level on the transverse ridge (Dick et al., 1991b). The detachment stage of this process corresponded to the imposition of comparatively low-angle, high-temperature crystal-plastic deformation of the gabbros. Brittle fracturing and veining occurred after this, evidently during the first stages of uplift of the plutonic section from beneath the rift valley.
A surprise was that the stable remanent magnetization of all the gabbros has a single, consistent inclination of ~71° (Kikawa and Pariso, 1991). This is ~19° higher than expected for the latitude of the site, suggesting that the rocks were tilted southward as a single block by this amount. If extended to depth, the gabbros cored are quite sufficiently magnetized to provide a source for the magnetic anomaly observed at the site. The site survey demonstrated that magnetic lineations paralleling the ridge axis strike directly across the summit of the transverse ridge (Dick et al., 1991b). The gabbros of Hole 735B thus are a component of the magnetized source for magnetic stripes (Kikawa and Ozawa, 1992; Pariso and Johnson, 1993). This obviously cannot be because of the quenching of tiny magnetites, as seen in pillow lavas. The pillows are not there. Instead, Curie temperatures were reached well after the rocks crystallized and were deformed, and even after much of the static alteration to form amphibole had taken place. Secondary magnetite formed at this later stage has to be responsible for the magnetic signature of the gabbros.
Thus, whereas the objective of extending drilling in Hole 735B during Leg 176 was formally and most simply to continue obtaining a long gabbro section, the actual heart of the venture was to extend these trends linking magmatism, deformation, metamorphism, and rock magnetization, plus some others, as far as possible into deeper rock and ideally into the upper mantle. How close to the mantle did we actually expect to come? Talcose serpentinite had been dredged from the west-facing slope of Atlantis Bank ~3 km west of Hole 735B at depths potentially within reach of the drill by the end of one more leg of drilling (Dick et al., 1991b). In the meantime, a seismic refraction study (Minshull and White 1996; Muller et al., 1997) indicated that high-velocity, unaltered mantle peridotite is present at ~5 km beneath the summit of Atlantis Bank, suggesting that perhaps serpentinized peridotite should lie between gabbro and Moho beneath the site. Formally, then, the deep target for Leg 176 was to breach the gabbro-peridotite transition, testing whether or not the peridotite is partially serpentinized. Failing that, however, we were to recover as long a gabbro section as possible in order to understand the origin of the ocean crust at this location.