We first set down the way in which Hole 735B came to be, as there are some misunderstandings about this. The drilling was a long-term project with objectives shifting over the years, not least because of the success of the first leg, and these need to be set out as well. We were principal proponents for the drilling from the outset and participated in decisions onboard ship, somewhat peripherally during Leg 118 and more centrally during Leg 176.
Hole 735B has been described as the result of "serendipity" by those who prefer a more rigorous method of planning and conducting scientific ocean drilling than that which took place. Although Hole 735B was not the initial target of Leg 118, which set out to drill serpentinized peridotites in the transform valley, the results of the drilling are fully consistent with the original objectives of that leg and its successor. Serendipity (the faculty of making desirable but unsought-for discoveries by accident; American College Encyclopedic Dictionary), therefore, is not an accurate term to describe what actually happened. At the least, the drilling was not unsought. The decision to drill at Site 735 was also an informed one, dictated to be sure by circumstance, but there were solid reasons behind it.
In 1987, during ODP Leg 118, Hole 735B was drilled for >500 m into gabbroic rock atop a shallow platform on the eastern transverse ridge of the Atlantis II Fracture Zone, Southwest Indian Ridge (Fig. F1). The drilling capped a tumultuous leg that began with unsuccessful attempts to core at three other locations in the transform valley and the steep eastern wall of the transform, all of them thwarted by the impossibility of penetrating boulders and the loose surficial material of debris flows that had flooded the transform valley. The two halves of a hard rock base, constructed while sailing east to the fracture zone in the first week out of Cape Town and needing only an optimistic signal to be bolted together over the moonpool and lowered to the seafloor, sat rusting on deck. For more than a month, each of the highest-priority drilling targets was attempted, to no avail. Video monitors in various locations on the ship usually showed a great churning of rock and mud during the attempted spud-ins, but no coherent core was recovered. The holes were barely worth the expenditure of a core bit, let alone the hard rock base. In due course, a groundswell arose among members of the scientific party to pull out of the rubble-glutted ditch and try a target atop the transverse ridge, where, at least, no such material could have accumulated.
At hand were the preliminary charts of a swath-mapping site survey, funded and carried out only a year earlier. The survey was one of the first ever of a transform fault using multibeam technology, and the Atlantis II Fracture Zone is still is the only one of its kind to have been surveyed by such means in the Indian Ocean. The Indian Ocean, of course, was well known as the location of numerous major rents in the seafloor, displacing segments of the Central and Southwest Indian Ridges (e.g., Fisher and Sclater, 1983). Most of these have large high-standing transverse ridges adjacent to transform valleys, similar to those of the Atlantis II Fracture Zone, from which large quantities of gabbroic and ultramafic rock had been routinely dredged (e.g., Engel and Fisher, 1969, 1975; Fisher et al., 1986). The proposal for drilling a fracture zone in the Indian Ocean, conceived in 1984 during a workshop on the Indian Ocean, was a response to the projected circumnavigation of the globe by JOIDES Resolution, then formally named SEDCO/BP 471, as part of the initial scientific plan of the new Ocean Drilling Program and the knowledge that these enticing targets would unavoidably lie in her path.
Thus, amidst the difficulties of Leg 118 the shipboard scientists fixed their attention on the heights shown on the unrolled survey charts, rather than the depths, and the principal remarkable fact about the eastern transverse ridge of the Atlantis II Fracture Zone, and the key to the eventual success of Leg 118 emerged. The ridge has a flat top that lies in only ~700 m of water. The top is so flat and shallow that the prospect of wave planation immediately came to mind, although a submerged reef was not an utter impossibility either. Whatever the case, statistics were presented to show how frequently gabbros and ultramafic rocks have been obtained by dredging at all levels of transverse ridges regardless of elevation in the Indian Ocean (Fig. F2). In generic form, such a target was also among the prioritized objectives for the leg. The scientific party deliberated, then voted. The Co-Chief Scientists obtained dispensation from authorities on "the beach," and JOIDES Resolution moved to what became Site 735.
A first glimpse of the seafloor using the video camera attached to the vibration-isolated television (VIT) frame lowered around the drill pipe and suspended just off bottom showed neither reef nor pillow lava but banded rock. Optimism rose. Coring at a short pilot hole brought up foliated gabbro, explaining the banded appearance of the outcrop and raising the level of optimism further still. A box survey with the video camera showed such rock in abundance between thin drifts of sediment. The signal was given, and the hard rock base was at last bolted together and lowered as quickly as possible to the seafloor. The almost perfectly flat bottom made the landing simple and assured the prospects for coring. No other preliminary operations were conducted; no casing was set in place. Instead, coring began as soon as the drill string could be turned around with a core bit at its end, and the funnel on the hard rock base, now perched on the seafloor, was reentered. Drilling went slowly at first until the bottom-hole assembly was safely below the seafloor, and then at an accelerating rate, one core of gabbro after another was brought on board. Reentry of the hole became routine. The material was cored with exceptional ease and astonishing recovery, nearly 87% over the entire hole. The scientific party, which for a month had made do with small chips and scraps from failed holes, was now nearly overwhelmed. The rig crew rushed to achieve the milestone of a 500-m penetration. A respite was finally provided to those describing the rocks by the staging of a comprehensive downhole logging program in the final days before leaving the site. Over the course of 17 days, Leg 118 went from being an abject failure to one of the most astounding successes in the history of drilling of the ocean crust. The first long section of the lower ocean crust ever drilled was in hand, logged, and revealing its secrets. With far more rock on board than anyone anticipated, the ship departed for Mauritius.
Open reentry holes drilled deep into the ocean crust have lives of their own, acting both as magnet to some and unwelcome burden to others. A proposal to return to Hole 735B was submitted within months of the completion of Leg 118. The Indian Ocean itself, however, was not so attractive to those with other interests. Long sojourns in the Atlantic and Pacific Oceans were in the future of JOIDES Resolution, and the deepening of Hole 735B, remotely situated in the Indian Ocean, did not fit with these plans. A well-attended workshop at Woods Hole Oceanographic Institute featuring archive portions of the Leg 118 core in the atrium outside the conference room launched a program of offset-section drilling of the lower ocean crust and upper mantle that eventually led to the scheduling of legs to drill gabbro and peridotite in the eastern Pacific and North Atlantic Oceans. The Southwest Indian Ridge, however, had to wait.
Again, site survey was an issue. The survey from 1986 provided a fairly broad regional context for Hole 735B, but the depth to the mantle transition, where other holes might be drilled if it became necessary to move, and the mechanism of exposure and uplift of the shallow platform—whether indeed it had been eroded at wave-base—were all unanswered questions. To be sure, simply reentering a hole does not of itself require a survey, and to some this meant that the drilling should proceed without it, allowing other surveys to be funded. Within the drilling planning structure, however, the lack of any more detailed survey information for a long time threatened the drilling. In the end, a multinational survey focused on the shallow platform was conducted 6 months after Leg 176. Dive programs using submersibles and remotely operated vehicles organized from Japan shortly followed. A British geophysical expedition prior to Leg 176 established something about the depth to the mantle (Muller et al., 1997). The environs of Hole 735B are now perhaps the best surveyed, best sampled, and, as we believe, best understood portion of a major fracture zone anywhere in the ocean basins.
Leg 176 was finally scheduled without a new survey, sandwiched between a paleoceanographic leg in the South Atlantic and another in the far southern oceans, both of them ending or beginning at Cape Town and fully 10 years after Leg 118. In the event, Hole 735B was occupied for the entirety of Leg 176. Following a short logging program to measure temperatures and obtain an image of the hole with the Formation MicroScanner (FMS), a tool unavailable during Leg 118, we began coring, picking it up with the same high recovery experienced in 1987. On first touching the bottom of the hole with the drill, we were surprised to find that it was 4 m deeper than when last measured during Leg 118. The cause of the discrepancy remains unknown, but perhaps has to do with small differences between average lengths of stands of pipe—always assumed to be a standard 9.5 m long—used during the two legs, accumulated over the >500 m length of the hole. Over the next several weeks, an additional 1004 m of rock—all of it gabbro—was cored, carrying penetration to 1508 meters below seafloor (mbsf). This was not entirely without incident, but because the drillers were both ingenious and resourceful there were no major difficulties. Recovery again remained astonishingly high and was almost identical to that of Leg 118.
A sense of the sheer amount of rock drilled and recovered over the two legs perhaps is conveyed by comparison with the backdrop to Cape Town, the port in South Africa that served both Legs 118 and 176 (Fig. F3). Table Mountain, like Atlantis Bank, is flat topped. It consists of Precambrian Cape Granite unconformably overlain by Cambrian sedimentary rock (Haughton, 1969). An aerial tram lifts tourists from the city to the summit. At the outset of Leg 176, our conjecture was that we would very likely core deeper than the tourists ascend. Perhaps we would even reach mantle peridotite. We did the one but not the other.
With a week to go before ceasing drilling and commencing logging, while some high-strength pipe joints near the top of the drill string were being shifted, two consecutive sharp heaves of the vessel in fairly heavy seas caused the end of the temporarily suspended pipe to jam against a ledge ~100 m above the bottom of the hole. The pipe string lifted off its supports in the rig and went into compression on each downswell. On the second wave, it snapped at the guide base. In the hole the severed part of the string broke through the ledge. Several tons of steel, including the entire heavy bottom-hole assembly, hurtled downward, coming to an abrupt halt at the bottom of the hole. Only the rock around the hole prevented the falling string from collapsing under its own momentum, but it undoubtedly was forced strongly, in a tight spiral, against the wall of the hole. During the next few days, the rig crew retrieved a portion of the lost pipe but the decision during Leg 118 not to string casing from the reentry cone into the upper part of the hole severely limited the size of fishing tools that could pass from the cone into the unconnected hole beneath (Fig. F4). In the end, more than 900 m of pipe was irretrievably wedged in the hole. The drilling was over, and our opportunities to continue—perhaps into peridotite—and to log all but a small portion of the newly cored section were also lost. Thus, on the brink of truly dramatic success came the untimely demise of Hole 735B.
The reports in this volume therefore mainly summarize work on the core. This in itself has been a considerable effort, comprising first the two shipboard studies and then shore-based work of dozens of scientists—participants in both legs—funded by various agencies of the partners in the Ocean Drilling Program. However, whereas most of the studies following Legs 118 and 176 deal only or mainly with those portions of the hole drilled during either leg, our synthesis (Natland and Dick, Synthesis Chapter, this volume) attempts to provide a synopsis of the entire core and place it properly in its geologic setting.