Hole 735B was originally drilled during Leg 118 to a total reported depth of 500.7 m, recovering 435 m of olivine gabbro and oxide olivine gabbro and related rocks (Robinson, Von Herzen, et al., 1989). Subsequently, during reoccupation of the hole during Leg 176, the depth to the bottom of the hole was measured as 504.8 m before the resumption of drilling. Depth to seafloor was the same as during Leg 118, and there is no explanation for this discrepancy. The unanticipated volume and rate of recovery did not permit adequate time to fully describe the Hole 735B cores during Leg 118. A subset of the scientific party, including Dick, Meyer, Bloomer, Stakes, and Kirby with the help of Chris Mawer, redescribed the core at the Texas A&M repository (Dick et al., 1991a). The final igneous lithostratigraphy conforms in general to the original six units described in the Initial Results volume (Robinson, Von Herzen, et al., 1989). The precise boundaries of the principal units, however, were moved by as much as 2 m. The units were subdivided into a total of 12 different subunits, consisting of ~495 distinct igneous lithologic intervals. At the same time, a full inventory of all metamorphic veins, a new log of deformation intensity, and detailed structural measurements were made.
The Leg 118 cores largely consist of olivine gabbro with apparently minor cryptic chemical variations. There is little evidence of the process of magmatic sedimentation, which is important in layered intrusions. This body is crosscut by numerous small microgabbro intrusions ranging from troctolite to oxide gabbronorite as well as a suite of late felsic veins, principally trondhjemite. The microgabbros often have highly irregular, but sharp contacts with the olivine gabbro, suggesting assimilation-fractional crystallization processes accompanying upward intrusion of melt bodies through the section. Primitive troctolites and troctolitic microgabbros are relatively abundant at the base of the section (up to Fo87 olivine) and were (erroneously) believed to be a harbinger of a major body of troctolite intruding the olivine gabbro from below. The olivine gabbro and microgabbros were in turn cross-intruded by numerous bodies of oxide-rich or oxide-bearing (0.1%-2%) gabbro, olivine gabbro, and gabbronorite. These are frequently associated with rock deformed during both hypersolidus and subsolidus conditions, during which late magmatic oxides filled cracks and formed pressure shadows around pyroxene augen. Geochemically, this produced an unusual extremely bimodal chemical variation with depth, without any first-order vertical variation in chemistry. This was unlike anything yet reported from a layered intrusion or ophiolite (see "Igneous Petrology").
These results were interpreted to show that the section formed by continuous intrusion and reintrusion of numerous small, rapidly crystallized bodies of magma (Bloomer et al., 1991; Dick et al., 1991c; Natland et al., 1991; Ozawa et al., 1991). Each batch of melt was intruded into a lower crust consisting of crystalline rock and semisolidified crystal mush. This led to undercooling and rapid initial crystallization of new magmas to form a highly viscous or rigid crystal mush, largely preventing the formation of magmatic sediments. Initial crystallization was followed by a longer, and petrologically more important, period of intercumulus melt evolution in a highly viscous crystal mush or rigid melt-crystal aggregate.
An unanticipated major feature of Hole 735B was the evidence for deformation and ductile faulting of still partially molten gabbro (Cannat, 1991; Dick et al., 1991a; Bloomer et al., 1991; Dick et al., 1992; Natland et al., 1991). This deformation was apparently particularly important over a narrow window late in the crystallization sequence (probably at 70%-90% crystallization) when the gabbros became sufficiently rigid to support a shear stress. This produced numerous small and large shear zones with enhanced permeability into which late intercumulus melt moved by compaction out of the relatively undeformed olivine gabbro (Dick et al., 1991a; Natland et al., 1991; Bloomer et al., 1991). Migration of this late iron-rich intercumulus melt into and along the shear zones locally hybridized the gabbro by melt-rock reaction and by precipitation of ilmenite, titanomagnetite, and other late magmatic phases. The net effect of these synchronous magmatic and tectonic processes is a complex igneous stratigraphy of relatively undeformed oxide-free olivine gabbros and microgabbros crisscrossed by bands of ferrogabbro. This process, also termed differentiation by deformation (Bowen, 1920), has recently been inferred for portions of the Lizard ophiolite (Hopkinson and Roberts, 1995). Evidence of such hyper-solidus deformation is absent, however, in high-level East Pacific Rise gabbros drilled at Hole 894G at Hess Deep (Natland and Dick, 1996; MacLeod et al., 1996). Also, basalts corresponding to melts in equilibrium with the ferrogabbros of Hole 735B are absent along this region of the Southwest Indian Ridge, although some do erupt along the East Pacific Ridge. Therefore the late magmatic liquids that permeated the shear zones were evidently uneruptable throughout most of their crystallization.
The primary igneous assemblage recrystallized under granulite- to amphibolite-facies conditions, and this was accompanied by the formation of amphibole-rich shear zones (Cannat et al., 1991; Cannat, 1991; Dick et al., 1991a, 1992; Mével and Cannat, 1991; Stakes et al., 1991; Vanko and Stakes, 1991). Locally there is a strong association between amphibole veins and zones of intense deformation. At this stage, as during late magmatic conditions, formation of ductile shear zones appears to have localized late fluid flow, with the most intense alteration occurring in or near zones of deformation. High-temperature alteration is far more extensive than found in layered intrusions (Dick et al., 1991a; Stakes et al., 1991), which are typically intruded and cooled in a near static environment. A rapid change in alteration conditions was seen in the Leg 118 gabbros, however, in the middle amphibolite facies with the virtual cessation of brittle-ductile deformation (Dick et al., 1991a; Stakes et al., 1991; Stakes, 1991; Vanko and Stakes, 1991; Magde et al., 1995). Mineral vein assemblages change from amphibole rich to diopside rich, reflecting different, more-reacted, fluid chemistries. Continued alteration and cooling to low temperature occurred under nearly static conditions, similar to those found at large layered intrusions. These changes were probably the result of an inward jump of the master fault at the rift valley wall. This transferred the section out of the zone of extension and lithospheric necking beneath the rift valley into a zone of simple block uplift at the inside-corner high of the Atlantis II Fracture Zone. Hydrothermal circulation, little enhanced by stresses related to extension, was greatly reduced, driven largely by thermal-dilation cracking as the section cooled to the current low temperatures in the hole.
The uniform stable magnetic inclination measured in Hole 735B gabbros drilled during Leg 118 demonstrates that there was little significant late tectonic disruption of the section, although the relatively steep inclination suggests block rotation of up to 20º (Pariso et al., 1991). The section thus likely preserves a typical record of accretion, hydrothermal circulation, and brittle-ductile deformation beneath an active rift. However, the later history of the rocks reflects static block uplift to shallow depths and rapid cooling at the ridge-transform intersection. Consequently, they do not record the full record of thermal equilibration and alteration of mature ocean crust cooled at depth beneath an intact section of ocean floor during seafloor spreading.
The Leg 118 results from Hole 735B show the importance of ephemeral magmatism, deformation, and alteration on crustal accretion at ultra-slow-spreading ridges. This contrasts with the results from the East Pacific Rise crustal section exposed at Hess Deep. There, no evidence of crystal-plastic deformation was found either in the high level gabbros from Hole 894G or in dredged gabbros (Gillis, Mével, Allan, et al., 1993)--perhaps because of the presence of a near-steady-state melt lens at the top of the section during accretion.