Description of the Hole 735B section required logging the features of 866 m of rock from a 1-km section of the lower ocean crust and integrating those observations with those made previously from the overlying 500-m section drilled during Leg 118. The recovery was more than double that of Leg 118, which previously held the record, and was some seven times the amount logged during Leg 147 in similar mater-ials at Hess Deep. To accomplish this, the scientific party divided into specialist igneous, metamorphic, and structural teams, rather than into alternating watches. The teams, in turn, divided core description into specific tasks, with one individual responsible for logging a specific observation. This method proved remarkably efficient, generating more data, observations, and core descriptions than on any previous hard-rock drilling leg.
The greatest value of this approach, however, was a dramatic improvement between Legs 118 and 176 in consistency, precision, and accuracy of observation. For example, different individuals estimated the modal abundance of plagioclase, olivine, and pyroxene in each of the 457 discrete igneous intervals described, whereas modes were independently determined by point-counting 220 representative thin sections. Averaged for the hole, the macroscopic modal analysis gave 59.5% plagioclase, 30.1% augite, 8.8% olivine, 0.33% orthopyroxene, and 0.76% oxide; point counting gave 58.9% plagioclase, 30.6% augite, 8.2% olivine, 0.62% orthopyroxene, and 0.79% oxide. The difference in orthopyroxene abundances is because half of it occurs as rims around olivine and cannot be distinguished macroscopically. These consistent observations, in turn, allowed direct comparisons between independently observed features of the core. Oxide abundances, for example, logged on a centimeter scale down the entire core, proved to have a remarkable positive correlation with the degree of crystal-plastic and magmatic foliation. The combined observations greatly facilitated our interpretation of the evolution of the lower crust at Hole 735B.
The Hole 735B core from between 504 and 1508 mbsf was divided by the igneous team into 457 discrete igneous intervals, numbered from 496 to 952 following the succession from the upper 500 m of Hole 735B (Dick et al., 1991a). These were distinguished on the basis of igneous contacts, variations in grain size, and the relative abundances of primary mineral phases (Fig. F8 ). Individual contacts were then described and logged. The major lithologies in Hole 735B were gabbro and olivine gabbro, comprising 14.9 and 69.9 vol% of the core, respectively. The distinction between the two is arbitrary, set at 5% olivine following the classification of the International Union of Geological Sciences, and there is complete gradation between them. The separation, however, allowed distinction of areas of lesser olivine content in the core. Generally, these are equigranular rocks (Fig. F9A) that occasionally contain a weak magmatic foliation often overprinted by crystal-plastic deformation (Fig. F9D). Frequently, though, it is varitextured, with irregular coarse, medium, fine, and even pegmatitic patches (Fig. F9E, F9F).
The average grain size of samples varies from fine grained (<1 mm) to pegmatitic (>30 mm), with average grain sizes generally in the range of coarse (5-15 mm) to very coarse (15-30 mm). In general, the relative grain sizes for the major minerals are in the order augite > plagioclase > olivine, but are usually similar. Pegmatitic intervals are sporadic through the core (Fig. F9E). Peaks in average grain size occur at 510, 635, 825, 940, 1100, 1215, 1300, 1425, and 1480 mbsf, and the grain size data for augite, plagioclase, and olivine all follow similar trends.
Weak modal and grain size layering is present in 22 intervals (12 vol% of the core). The types of layering observed include (1) grain-size layering characterized by either sharp breaks in grain size or gradational variations in grain size (Fig. F10A), (2) modal layering marked by distinct changes in the abundance of plagioclase, olivine, clinopyroxene, and Fe-Ti oxide (Fig. F10B, F10C), (3) magmatic foliation (igneous lamination) defined by the preferred orientation of plagioclase and in some cases olivine and clinopyroxene, and (4) layering defined by textural changes such as layers with crescumulus texture. In several intervals, grain-size and modal layering are present in rhythmic sequences.
The olivine gabbro is locally crosscut by fine- to medium-grained microgabbros having contacts that range from sharp, with a slight but definable finer-grained margin, to irregular grading and swirling up through and into the adjoining olivine gabbro (Fig. F9B). These range in composition from primitive troctolites in lithologic Units IV and XII at the top and bottom of the Leg 176 section, to microgabbro and gabbronorite, although the majority are olivine microgabbros, which are nevertheless similar to the olivine gabbros they frequently cut. The origin of these bodies is speculative, as they could represent channels along which relatively hot primitive melt was fed up into the succession of gabbro intrusions, or they could be protodikes through which the typically primitive magmas of the Southwest Indian Ridge erupted to the seafloor.
Oxide-rich gabbros, including oxide olivine gabbro, make up 7 vol% of the recovered rocks, and gabbronorites and oxide gabbronorites make up approximately 8 vol%. These intervals range considerably in size, but they decrease noticeably downward in the section and never are as abundant as the upper 500 m of the hole. There is nothing like the nearly 100-m-thick polygenetic units of disseminated oxide olivine gabbro and oxide olivine gabbro containing numerous sheared intervals in Units III and IV between 170.22 and 274.06 mbsf. The latter requires that a very large flux of melt passed through them to account for the massive precipitation of intercumulus Fe-Ti oxides (Natland et al., 1991; Dick et al., 1991a), whereas the former could easily have crystallized from locally derived iron-rich melts sweated out of the crystallizing olivine gabbros. The oxide-rich gabbros in the lower two-thirds of Hole 735B are found as innumerable sheared and undeformed irregular patches and veins in olivine gabbro (Fig. F9C) and as consistently deformed larger intervals each as thick as several meters or more. A consistent and impressive feature of the oxide gabbros is their overall strong association with areas of magmatic and crystal-plastic foliation, as detailed in the structure section, and with the percent oxide found in the core. This is noteworthy because most of these rocks are cumulates and do not represent a liquid composition. Moreover, the liquids with which they were in equilibrium were far more evolved than any pillow basalt that has been dredged along the Southwest Indian Ridge (Natland et al., 1991; Dick et al., 1991a).
Dick et al. (1991a) proposed that this is because the oxide gabbros represent intrusion of late iron-rich melts that have migrated out of the olivine gabbros and along shear zones penetrating or originating in partially molten lower crust (synkinematic igneous differentiation). When such melts migrate upsection and down a temperature gradient, they should decrease in mass as they migrate and crystallize. In the case of a late Fe-Ti-rich melt near the end of crystallization migrating through gabbros sufficiently rigid to support a shear stress (80%-90% crystallinity), the liquid would precipitate abundant ilmenite and magnetite as it cooled. Thus, the greater the fluid flux through any volume, the greater should be the enrichment in precipitated oxides. Accordingly, the association between oxides and deformation throughout Hole 735B can be interpreted as evidence that deformation and the formation of shear zones influenced the flow and transport of late intercumulus melt throughout the section.
A wide variety of felsic rocks constituting 0.5% of the core were described. The majority are leucodiorite; however, diorite, trondhjemite, tonalite, and very little granite also are present. They are largely net veins, and rarely are sufficiently massive (5 cm) to be described as an igneous interval. Although many of these veins are clearly of igneous origin, having primary igneous textures and sharp intrusive contacts, many have experienced subsequent high- and low-temperature alteration and developed diffusive or reactive contacts with the host gabbros. Still others may be hydrothermal or metamorphic in origin.
Seven additional major lithologic units (VI through XII) were identified below the Leg 118 section. These are based on modal mineralogy and the relative abundance of rock types and include Unit VI--compound olivine gabbro, which continues from 382 mbsf in the Leg 118 core to 536 mbsf; Unit VII--gabbronorite and oxide gabbronorite from 536 to 599 mbsf; Unit VIII--olivine gabbro from 599 to 670 mbsf; Unit IX--gabbronorite and gabbro from 670 to 714 mbsf; Unit X--olivine gabbro and gabbro from 714 to 960 mbsf; Unit XI--olivine gabbro from 960 to 1314 mbsf; and Unit XII--olivine gabbro and troctolitic gabbro from 1314 mbsf to the bottom of the drilled hole at 1508 mbsf. These units are shown in Figure F8. The stratigraphy, at first glance, would seem to resemble that of a large layered intrusion, with the proportion of rocks crystallized from differentiated and evolved liquids increasing upward. However, this is misleading. There is little layering, and none that resembles that characteristic of a layered intrusion. Rather, the section consists of a series of individual olivine gabbro intrusions, best defined geochemically, which are crosscut repeatedly at higher levels in the crust by oxide-rich gabbros. Thus the lower ocean crust here is differentiated kinematically by intrusion of late melts into the top of the section. In fact, as discussed in the geochemistry section, despite irregularly increasing olivine content, the lower olivine gabbros are less primitive, more titanian and richer in soda, than those crosscut by the evolved intrusives higher in the section.
The average phase proportions in Hole 735B troctolites and olivine gabbros closely resemble cotectic proportions observed in low-pressure experiments on mid-ocean-ridge basalts, suggesting that the main body of olivine gabbro crystallized at relatively shallow depths (<6 km) and solidified after efficient expulsion of residual melts. The more evolved Fe-Ti oxide-bearing gabbros do not have good experimental cotectic analogs. The strong correlation between deformation and oxide-rich gabbros in the section suggests (1) lenses of late-stage magma may have acted as zones of weakness along which deformation was concentrated; (2) late-stage magmas may have been concentrated in zones that had been previously sheared, because these zones have greater high-temperature permeability; or (3) active shear zones acted as conduits for melt transport through the section. The high concentrations of oxides present in some samples require that large volumes of melt were transported through these zones.
The metamorphic petrology team logged 2792 veins as well as groundmass alteration assemblages on a piece-by-piece basis downcore. This was supplemented by an examination of some 243 thin sections of the Leg 176 cores. The individual observations were made by two pairs of investigators, with one pair identifying, measuring, and logging veins and the other characterizing and logging alteration in the groundmass of the gabbros. The same types of observations, and others, were then made on the thin-section suite by a fifth investigator, who separately confirmed and expanded the observations of the macro-description teams. In hand sample, plagioclase alteration was based on the extent of a milky white appearance. This did not distinguish between hydrothermally altered plagioclase and plagioclase recrystallized by the high-temperature granulite facies deformation, which was pervasive in many sections of the core. Thus the total background alteration logged downhole is significantly greater than might be inferred in thin sections, in which secondary plagioclase that formed during hydrothermal alteration is more readily distinguished from that produced by dynamic recrystallization. However, such a distinction is often arbitrary and was not made here. Otherwise, the thin-section observations are remarkably consistent with the observations of alteration in hand specimen. The thin-section observations are presented in Figure F11. Total alteration and vein abundances downhole are plotted in Figure F12. This figure uses additional data on vein abundances for the upper 500 m (Dick et al., 1991a).
The gabbros recovered during Leg 176 range from fresh to 40% altered, although there are many small intervals where alteration is far more extensive. Typically, however, there is less than a few percent total hydrothermal alteration through long sections. The most intensely altered portion of the core is between 500 and 600 mbsf, where both amphibole and secondary recrystallized plagioclase are most abundant and the primary minerals are on average 10% to 40% replaced. This contrasts sharply with the amphibolite mylonites at the top of Hole 735B, where total alteration was more intense, coming close to 100% (e.g., Robinson et al., 1991; Stakes et al., 1991). Calcite veins, associated with low-temperature oxidation of the rocks, are also abundant between 500 and 600 m, evidently reflecting ongoing alteration at low temperatures as a result of the presence of open fractures. A second zone of intense recrystallization is present between 800 and 1030 mbsf, where many of the rocks exhibit high-temperature plastic deformation, and veins are rare. Two less altered zones are located in an interval of abundant smectite veins at 1300-1500 mbsf. Below 1030 m, the intensity of alteration is generally much less than 10%.
The Leg 176 gabbros preserve a complex record of high-temperature metamorphism, brittle failure, and hydrothermal alteration that began at near-solidus temperatures and continued down to very low-temperature conditions. The highest temperature metamorphic effects are transitional with magmatic processes; they most likely overlap both temporally and spatially, and in places the effects of these two processes are nearly indistinguishable. This is particularly true of the felsic vein assemblages, which range from clearly magmatic to apparently exclusively hydrothermal. Plagioclase and diopside veins, and combinations thereof, on splays from apparently igneous felsic veins were not unusual (Fig. F13). In general, felsic veins also served as conduits for late fluids, possibly of magmatic origin, commonly leaving a heavy overprint on the igneous assemblages, and which extended to partial replacement of some veins with clays.
Granulite facies metamorphic conditions (>800º-1000ºC) are clearly marked by localized, narrow zones of crystal-plastic deformation that cut igneous fabrics. These intervals characteristically have anastomosing layers of olivine and pyroxene neoblasts that are bounded by plagioclase-rich bands. In some places, the high-temperature shear zones are associated with impregnations of oxide gabbros; in many cases, these zones have abundant recrystallized brown hornblende, indicating that deformation continued down to amphibolite facies metamorphic conditions. Other high-temperature effects probably resulting from late-stage magmatic activity include the formation of plagioclase + amphibole veins and diopside-rich veins, which in some intervals are progressively transposed into localized zones of high-temperature shear. Many of these rocks, veins, and shear zones reflect the effects of late magmatic hydrous fluids, but these zones also acted as pathways for later hydrothermal fluids at various temperatures.
Static high-temperature alteration is commonly associated with vein formation, and it is patchy throughout the section. Extensive intervals (>300 m) are marked by less than 10% total background alteration. This alteration is generally manifested by coronitic alteration halos around olivine grains, and the common replacement of clinopyroxene by variable amounts of brown amphibole. In more evolved rocks, magnesium-amphibole ± talc typically replaces orthopyroxene. The secondary minerals most likely formed under low water-to-rock ratios over a range of temperature, from >600º to 700ºC down to much lower temperatures.
Ingress of very warm to hot fluids (400º-550ºC) was facilitated by subvertical fractures, now lined with amphibole, that were probably related to cooling and cracking of the rocks in the subaxial environment (e.g., Fig. F14). However, the abundance of amphibole veins decreases markedly with depth, as does the alteration, and below 600 mbsf amphibole veins are rare. Amphibole veins and groundmass alteration are associated with zones of deformation at the top and bottom of the Leg 118 portion of the hole, with a sharp drop in abundance in undeformed intervals. The greatest alteration, often with complete replacement of mafic phases by amphibole and the highest vein abundances, is situated in the upper 100 m in a zone of intense deformation (Dick et al., 1991a; Stakes et al., 1991). A second, more irregular zone of deformation and alteration is present from 400 to 500 mbsf, where amphibole replacement of mafic phases is only partial. Although this zone of deformation and associated amphibole alteration continues into the upper 100 m of the Leg 176 section, it does not reach deeper levels of the core, where foliated gabbros largely experienced crystal-plastic deformation at high temperatures. Microcracks filled with talc, magnetite, amphibole, sodic plagioclase, chlorite, and epidote are spor-adically present throughout the core and represent smaller scale fracturing and fluid penetration under greenschist facies metamorphic conditions.
Cessation of hydrothermal fluid flow is marked by abundant late smectite, carbonate, zeolite ± prehnite veins and iron oxyhydroxide minerals associated with intense alteration at 500-600 mbsf. These minerals reflect low-temperature alteration by circulating seawater solutions, and they are most likely related to the presence of a fault at 560 mbsf. Smectite veins (Fig. F15), unlike the higher temperature vein assemblages, are often associated with alteration haloes where olivine and even pyroxene are extensively replaced by smectite. Below this interval, veins of smectite ± pyrite ± calcite, together with associated smectitic alteration of surrounding wallrock, reflect low-temperature hydrothermal reactions under more reducing conditions. These effects are observed throughout much of the core, but the abundant smectite veins at 600-800 m are most likely related to another fault at 690 mbsf. This lower temperature set of veins formed in tensional fractures and is most likely related to cooling of the block during uplift of the massif to form the transverse ridge.
A peculiar feature of the Hole 735B cores is that high-temperature alteration assemblages are most abundant at the top of the section, whereas low-temperature assemblages dominate near the base. Greenschist assemblages, representing intermediate conditions, are relatively minor. This inversion of the normal order of things, which has, for example, been found in the in situ section of sheeted dikes and pillow lavas at Hole 504B (Alt, Kinoshita, Stokking, et al., 1993), is reasonably attributed to the particular cooling history of the section. Apparently, at an early stage, conditions for the percolation of water to great depth did not exist beneath the rift valley, and alteration was largely limited to the upper portions of the gabbroic crust. Before a normal cooling profile could be established, however, the massif was unroofed and rapidly uplifted to the seafloor. The rapid cooling and relatively static conditions in the interior of the uplifted block inhibited extensive greenschist facies alteration, and there is only a single large epidote vein found in the entire lower two-thirds of the hole. In contrast, low-temperature alteration phases are abundant in the lower two-thirds of the core. This likely reflects alteration in the rift mountains after unroofing and uplift of the gabbro massif. With re-establishment of a normal conductive geotherm, circulation of seawater through a relatively restricted set of open fractures and continued cracking resulting from cooling produced local zones of oxidative alteration and smectite-lined veins.
To represent a systematic sampling of the major lithologies in the Leg 176 cores, 180 whole-rock samples were selected for analysis, representing one analysis every 4.8 m of recovered core. In principle, at least one sample representing the main lithology was taken from each core, even when an apparently homogeneous unit spanned several cores. Seams of oxide gabbro and larger felsic veins were occasionally sampled to study the complete range of petrologic differentiation. Given the distribution of lithologies, the large majority of the samples chosen for analysis were olivine gabbros and subordinate gabbro, disseminated Fe-Ti gabbro, and microgabbros. A limited number of samples were selected from the intervals strongly affected by high- and low-temperature alteration. Of the 458 igneous intervals identified in the core, about 140 are represented in the analysis suite. Analysis was done using X-ray fluorescence (XRF) for major element compositions and for the abundances of the trace elements V, Cr, Ni, Cu, Zn, Rb, Sr, Y, Zr, and Nb. Sample preparation techniques are outlined in the "Explanatory Notes" chapter. Samples taken for analysis generally weighed 20 to 30 g. Larger slabs were cut from the very coarse-grained intervals, and a thin section was prepared from a billet from the same or adjacent material. A representative selection of analyses for the different lithologies is included in Table T1.
Figures F16 , and F17, show the downhole variation of Mg# and TiO2. The least evolved rocks are troctolites from 500-520 mbsf. Throughout the entire gabbro section there are numerous thin intervals of Fe-Ti oxide gabbros and felsic veins that are significantly to strongly differentiated. The Mg# should be used with some caution as a "differentiation index" in the case of the oxide gabbros. Mg#s of cumulate rocks decrease as a result of the accumulation of iron-rich minerals, such that low numbers overestimate the extent of crystallization. Overall, the gabbros are split into two groups. The most voluminous are olivine gabbros and troctolites with minor oxides that have high magnesium numbers. These are crosscut by later Fe-Ti-rich oxide gabbros and felsic veins with high TiO2 contents, low Mg#, and relatively sodic compositions that exhibit extreme variability in their composition because of the accumulation of iron oxides. Such narrow, crosscutting late rock types are of little value in establishing a chemical stratigraphy, which is based entirely on the chemistries of the main gabbro types and principally on the Mg#s. Examining the depth profile (Fig. F16), approximately five major cycles can be identified of decreasing Mg# going from high Mg at depth to low Mg with increasing TiO2 in the olivine gabbros (roughly 0-225, 250-525, 525-900, 950-1350, and 1350-1508 mbsf). A prominent feature of the observed variation is that the lowermost three units in this cyclic repetition are more iron rich than the upper two; this is consistent with a somewhat more sodic composition and a slightly higher TiO2 content of the rocks.
As seen in Figure F8, the most extreme lithologic variability exists in the upper 1000 m of the hole, with the lowermost 500 m being largely olivine gabbro. Within the upper kilometer there is also an apparent change in the proportions of different lithologies above and below 500 mbsf. A preliminary mass balance calculation using the measured thickness of the lithologies, their average compositions from the shipboard XRF measurements, and the average rock densities, shows that the bulk composition of the 500-1000 m interval has 0.69 wt% TiO2, 2.99 wt% Na2O, and a Mg# of 70.5. By contrast, the same calculation done for the 0-500 m interval (Dick et al., 1991a) gives a bulk composition with 1.41 wt% TiO2, 2.67 wt% Na2O, and a Mg# of 67.2. Although the 0-500 m interval is richer in titanium and has a lower Mg# than the 500-1000 m section, it also has lower Na2O, which appears inconsistent with the hypothesis that it simply crystallized from more evolved liquids. In fact, the olivine gabbros in the uppermost 500 m are more primitive than those in the middle 500 m (0.34 wt% TiO2, 2.49 wt% Na2O, 80.3 Mg# vs. 0.43 wt% TiO2, 2.87% wt% Na2O, and 71.6 Mg#). This difference in the bulk chemistry between 0-500 m and 500-1000 m (which would be even more pronounced with a comparison to the largely olivine gabbro 1000-1508 m section) is entirely due to the increasing volume of intrusive Fe-Ti oxide-rich gabbros and gabbronorites in the upper part of the hole.
The main conclusions that can be drawn from the shipboard chemical analyses from both Legs 176 and 118 are as follows:
Measurements made on the Leg 176 core by the structural team include intensity and orientation of magmatic deformation, crystal-plastic deformation, cataclastic deformation, igneous and metamorphic veins, as well as description of crosscutting relationships. Additional observations on thin sections were made as a team with the aid of a video monitor. The results showed considerable variation in the structures observed throughout Hole 735B, both in style and position. Several unexpected features, notably the occurrence of significant reverse-sense shear zones and a general decrease in deformation downhole, were found.
Wherever possible, observations on the Leg 176 cores were combined with results from the upper 500 m of the hole drilled during Leg 118. This was done using published results, and by combining our observations with data logs for the intensity and orientation of deformation and metamorphic veins for the Leg 118 section used by Dick et al. (1991a). The team also relogged the lowermost 50 m of the core recovered during Leg 118 to provide a basis for comparison of the Leg 176 observations to those for the upper 500 m of the hole.
The structural geologists marked all the Leg 176 cores orthogonal to the foliation for splitting. The cores were placed into the archive and working halves of split core liners with the foliation dipping in the 90º direction in the core reference frame (i.e., dipping to the right on the cut face of the core in the working half). This provided a consistent framework for description and measurement of structures in the core. Thus the strong local concentration of poles to foliation shown in the stereo plots in Figure F18 with respect to strike are a direct artifact of the method in which they were split and oriented in the core liners. They are not geographically referenced. However, shipboard paleomagnetic declinations for these cores show a similar cluster around 250º in the core reference frame throughout the entire cored interval. This result demonstrates that, overall, the foliations have a consistent orientation and thus can be reoriented into the geographical reference frame by making the assumption that the declination of the remanent magnetic vector dips toward the south. The resulting reorientation suggests that both the crystal-plastic and the late magmatic foliations predominantly dip to the north and, hence, toward the ridge axis.
A majority of the rocks from Hole 735B (78%) have coarse- to medium-grained hypidiomorphic-granular, intergranular, and sub-ophitic textures with no preferred mineral alignment caused by late magmatic deformation. The remainder contain a variably developed magmatic foliation that is defined by the preferred orientation of elongate plagioclase laths and, locally, by pyroxene crystals with weak magmatic foliations predominating and strong fabrics occurring sporadically (Fig. F18). Magmatic foliations vary in strike and dip, with no systematic variation with depth. The majority have dips between 20º and 50º.
Crystal-plastic deformation in Hole 735B, as shown in Figure F18. It is highly localized, with the most intense deformation observed in the intervals 0-50 and 450-600 mbsf, and with intervals of little or no crystal-plastic deformation generally increasing in number and length downhole. Overall, 77% of the rocks recovered during Leg 176 have no crystal-plastic fabric, and only 7% have more than a weak foliation, whereas 71% of the Leg 118 gabbros have no crystal-plastic fabric and 14% have more than a weak foliation. Again, there is no systematic variation of the dip of the foliation with depth, although there is a strong concentration at ~30º.
A striking feature of the crystal-plastic deformation fabrics from Hole 735B is the number of shear zones with reverse sense of shear concentrated within and below the 20-m-wide shear zone between 945 and 964 mbsf and above the fault at 690 mbsf. From the top of the core to 680 mbsf, the majority of the shear zones display normal shear. An example of a reverse shear zone is shown in Figure F19. These are as yet unexplained; however, the high temperature of their formation, and the origin of the Atlantis Bank at the inside-corner high, dictate that shear zones formed in the lower crust beneath the ancestral rift valley on the Southwest Indian Ridge.
A major feature of the Leg 176 cores is the strong association between magmatic and crystal-plastic fabrics (Fig. F18A). This is striking because only 22% of the core has a weak to moderate magmatic foliation, and less than 1% a strong magmatic foliation, whereas 77% of the gabbro has no crystal-plastic fabric at all. Obviously, no magmatic fabric should be preserved in regions with a strong crystal-plastic fabric (e.g., 710-730 mbsf); however, rocks with both fabrics are quite common and are always near each other and long intervals of the core contain neither. In addition, in areas where the crystal-plastic fabric is intense, a magmatic fabric is often observed in the adjacent gabbro. Because the two fabrics have similar orientations (Fig. F18B), it is evident that they developed under the same or similarly oriented stress fields. These observations indicate that crystal-plastic deformation localized within areas with a magmatic foliation, representing the transition from late magmatic to solid-state deformation. There are several interesting exceptions to this correlation, particularly in the lower portion of the hole, demonstrating that the relationship between the two may be complex in detail. The question naturally arises as to whether the "magmatic foliation" was truly produced by magmatic processes. This was addressed by a careful, independent assessment of 240 thin sections for the presence of crystal-plastic deformation and magmatic fabrics. Although weak magmatic fabrics are hard to see in thin section, undeformed samples with magmatic fabrics and the relative intensity of crystal-plastic fabric correlated remarkably well to the macroscopic observations of the core.
Retrograde shear zones at less than granulite facies conditions are evidenced by brittle-ductile deformation and the formation of amphi-bole in the plane of shear and by crosscutting veins and microcracks. Two principal retrograde shear zones are present in Hole 735B, both of which coincide with zones of earlier late magmatic deformation and crystal-plastic deformation. The most intense of these is in the upper 100 m of the hole, where pyroxene is often entirely replaced by amphibole and true amphibolites are found. The less intense zone is around 500 mbsf at the bottom of the Leg 118 section and the top of the Leg 176 section.
Lower temperature cataclastic fabrics and faults are found at several levels within the Leg 176 section and at the base of the Unit IV Massive Oxide Gabbro at 274 mbsf (Dick et al., 1991a). Potentially major faults are absent in the upper 450 m of the hole. Two relatively major faults with the potential for significant displacement (e.g., hundreds of meters to kilometers) exist at 560 and at 690-700 mbsf, in addition to several minor ones at different locations. Smaller discrete planar faults with associated gouge, breccia, cataclasite, and ultracataclasite cut all rock types. These were logged at 600 locations in the Leg 176 core but are concentrated in the upper half of the hole, and are virtually absent below about 1400 m. The most common of these features are small-offset microfaults filled with calcite, amphibole, and/or smectite (below 1050 mbsf). Mineral assemblages associated with the brittle tectonic structures include amorphous silica, prehnite, chlorite, epidote, actinolite, and secondary plagioclase. These reflect a range of conditions for cataclastic deformation. Downhole dip shows no consistent pattern, although there is a maximum of poles to foliation near a plunge of 90º in stereo plots indicating a preferred subhorizontal fault orientation in the lower 1000 m of Hole 735B. Of the faults, 8.8% exhibit oblique slip, 2.8% are pure dip slip, and 2.3% are strike slip.
The cores from Hole 735B contain many late brittle deformation features with evidence of cataclasis, which are associated with alteration assemblages extending from the lower greenschist facies to the current temperatures in the hole. These features are almost certainly associated with block uplift, unroofing, and subsequent cooling of the massif at the inside-corner high of the Southwest Indian Ridge. This in large part reflects the bimodal metamorphic history of the massif wherein low-temperature alteration is concentrated near the bottom of the hole, whereas high-temperature amphibolite-facies rocks are concentrated at the top.
The gabbroic rocks cored during Leg 176 display magmatic, crystal-plastic, and brittle deformation features together with associated overprinting relations consistent with synkinematic cooling and extension in a mid-ocean-ridge environment. The following observations provide a basis for interpreting the conditions of deformation during evolution of this block of lower oceanic crust.
These observations indicate that the processes that control crustal accretion at slow-spreading ridges are strongly influenced by localized deformation at conditions ranging from magmatic to low-temperature cataclastic. The correlation between structural domains and igneous intervals suggests that this segment of the Southwest Indian Ridge is not supplied by a steady-state magma source. Rather, intrusion and deformation are episodic phenomena that may occur separately or synchron-ously but at different rates. In many cases, zones of localized deformation remain active over a wide range of conditions, for example, cataclastic overprint of oxide-rich crystal-plastic shear zones that were initially active under partially molten conditions.
Paleomagnetic intensities, inclinations, and declinations were measured separately on all the archive halves of the Leg 176 cores and on a large number of minicores drilled from the working halves. The average natural remanent magnetization (NRM) intensity of the Leg 176 minicores is 2.5 A/m, the same as the estimated value of 2.5 A/m for the upper 500 m of Hole 735B once the drilling-induced component is removed (Pariso and Johnson, 1993). There appears to be no general decrease in magnetization downhole; however, magnetic susceptibilities vary from 8.12 × 10-4 to 0.123 SI, with mean values smaller than for gabbros from the upper 500 m of the hole and having less scatter and decreasing slightly with depth (Fig. F20). Magnetic "hardness" increases with decreasing grain size and compensates for the strong decrease in abundance of relatively coarse-grained oxide gabbro downhole. Although olivine gabbro is nearly oxide free in the upper 500 m (Natland et al., 1991), it has significantly more relatively fine-grained Fe-Ti oxides in the lower 1000 m of Hole 735B, where it is the most abundant lithology. Overall, the gabbros of Hole 735B have a very stable remanent magnetization, with a high and often very sharp blocking temperature suggesting relatively rapid acquisition of thermoremanence during cooling of the gabbros.
The vertical structure of the sources of lineated marine magnetic anomalies has remained poorly known ever since the recognition, more than 30 yr ago, that the ocean crust records reversals of the Earth's geomagnetic field. Several authors have suggested that some or most of the stable source might reside in the gabbroic crust, based on the magnetization of dredged gabbros (Fox and Opdyke, 1973; Kent et al., 1978). However, such surficial samples have been subjected to varying degrees of hydrothermal alteration and weathering during faulting and emplacement to the seafloor that would likely significantly affect their magnetic properties. During the site survey for Leg 118 (Dick et al., 1991c), well-defined magnetic anomalies were mapped over large regions of the rift mountains of the Southwest Indian Ridge adjacent to the transform fault. Extensive dredging of these regions, including Atlantis Bank, recovered largely gabbro and peridotite, suggesting that these lithologies must be responsible for the anomalies (Dick et al., 1991c), a possibility first raised by the laboratory work of Fox and Opdyke (1973) and Kent et al. (1978). The hypothesis that gabbro could be a major contributor to the magnetic anomaly over Site 735 was confirmed by downhole logging (Pariso et al., 1991) and by direct measurements on cores (Kikawa and Pariso, 1991; Kikawa and Ozawa, 1992; Pariso and Johnson, 1993). With the addition of the Leg 176 cores, however, it now appears that the 1.5-km Hole 735B gabbro section is the principal source of the lineated magnetic anomaly over the site, to the extent that this section is representative of the crust in three dimensions.
This conclusion presents the possibility that gabbroic crust may potentially contribute to, and even exceed, the contribution of pillow basalts and sheeted dikes to marine magnetic anomalies elsewhere. However, the rapid acquisition of thermal remanence of the Hole 735B gabbros is consistent with rapid cooling caused by unroofing and uplift to sea level at the ridge-transform intersection. This may not be the case for normal Southwest Indian Ridge crust away from transforms, which would cool under a 2-km carapace of pillow basalts and sheeted dikes and acquire its thermal remanence more slowly.
Thermal and alternating field demagnetization studies of discrete samples from Leg 176 reveal two magnetization components: a low-stability component apparently related to drilling and more stable components with a steeper inclination. The mean characteristic inclination for Leg 176 discrete samples is reversed, with Inc = 71.4º (+0.3º/-3.1º), uncorrected for any hole deviation from vertical (Fig. F20). This is statistically identical to that found in the upper 500 m (71.3º, +0.4º/-11.0º) when the latter are recalculated using the method of McFadden and Reid (1982). Because Antarctica has been relatively fixed in the plate reference frame over the last 11 m.y., the present latitude of Hole 735B is likely close to that of the its ancestral ridge axis at the time of remanence acquisition. Thus, the observed inclination is steeper than the expected 52º for the site and requires a tectonic rotation of the section of approximately 19º ± 5º (depending on the deviation of the hole from vertical).
Overall, the tight cluster of stable magnetic inclinations downhole is significant to the interpretation of the igneous petrology and structure of Hole 735B, because the inclinations indicate that there has been little tectonic disruption of the section since it cooled below the Curie point at ~580ºC. Moreover, there is an unexpected strong preferred orientation of declinations at ~260º in the reference frame of the core liner. This is because the structural geologists systematically oriented each section of core for splitting so that they were cut orthogonal to the foliation, with each half placed in the working and archive halves so that the foliation dipped consistently in one direction. Thus, the consistent declinations demonstrate that these foliations are not random and suggest that gross reorientation of structural features in the core may be possible. Assuming a south-pointing characteristic remanence declination, the mean declination of 260º would be restored to 180º by a counterclockwise rotation of ~80º. Structural planar features that preferentially dip toward 90º in the core reference frame would thus dip toward the axial rift to the north.
Physical properties were measured using the MST on all the whole cores (natural gamma ray, gamma-ray densiometry, and magnetic susceptibility). Also measured were index properties including density, porosity, compressional wave velocity, and thermal conductivity on archive-half core pieces and on minicores from the working half. Only the magnetic susceptibility measurements made on the MST proved to have geologic value.
Thermal conductivity measurements were made at 219 intervals through the borehole section. Over the section the thermal conductivity is 2.276 ± 0.214 W/(m·K). The thermal conductivity varied considerably in the region of felsic veins. These values lie within the range measured for the upper 500 m of Hole 735B and for gabbroic rocks in general (Clark, 1966).
The MST system measured in excess of 22,000 magnetic susceptibility points. These are shown in Figure F21 and exhibit two characteristic features. The first is an overall decrease in susceptibility downhole defined largely by a gradual decrease in the baseline value with depth. The second is the occurrence of more than 600 extreme spikes of high susceptibility that can be individually correlated to the location of intervals of oxide-rich gabbros and gabbronorites enclosed within oxide-poor olivine gabbro and crosscutting felsic veins in the Hole 735B cores. The MST spikes indicate that the typical interval is no more than 10 to 15 cm thick, and is in many cases significantly smaller. These intervals should be most abundant in the upper 500 m of the core, where susceptibility measurements on individual samples define the highest overall unit susceptibility, but where the MST measurements were not made. The proportion and frequency of these oxide-rich intervals decrease systematically downhole from 274 m at the base of the Unit 4 massive oxide olivine gabbro.
Mass and volumetric measurements were made on 218 minicores with a mean porosity of 0.649% ± 2.884% (the population being heavily skewed by the large number of minimal porosities and the small number of porosities significantly greater than 1%). The mean bulk density was 2.979 ± 0.10 g/cm3, and a mean grain density was 2.991 ± 0.107 g/cm3, close to the density of the typical olivine gabbro (2.96 g/cm3). Density varied with mineral content and was highest in oxide gabbros (3.21 g/cm3) due to the presence of substantial ilmenite and magnetite (4.7-5.2 g/cm3) and lowest in troctolitic gabbros and olivine gabbros, where the proportion of plagioclase (2.7 g/cm3) was greatest. Density variations downhole are consistent (Fig. F22 ) with lithologic variations, with the greatest scatter in the upper 500 m where oxide gabbros are most abundant (between 2.8 and 3.3 g/cm3). Densities show increasing scatter below 925 mbsf (between 2.8 and 2.9 g/cm3). A slight decrease in mean density at the bottom of the hole reflects an increasing proportion of plagioclase-rich olivine gabbro and troctolite.
The physical properties measurements provide considerable insight into the origin of the vertical incidence seismic profile (VSP) reflectors identified from the Leg 118 VSP (Swift et al., 1991). The compressional velocities of 217 minicores were measured; they averaged 6777 ± 292 m/s and are shown together with the data from Leg 118 in Figure F22. No significant variation in minicore compressional velocities occurs for the Leg 176 cores downhole, not even at the VSP reflectors at 560 m and between 760 and 825 m. Higher in the hole, however, there is a clear dip in velocities for the massive oxide gabbro Unit 4 that corresponds to a striking increase in the density of minicores and the vertical-incidence seismic profile reflector at 225-250 mbsf (Iturrino et al., 1991; Dick et al., 1991a). A sharp increase in seismic velocity of the Leg 118 cores near the top of the hole occurs in the vicinity of another reflector at 50 m. The latter corresponds to a zone of amphibolites with intense shear and crystal-plastic deformation that has produced a strong crystal-fabric orientation. Iturrino et al. (1991) demonstrated that these are elastically anisotropic, with strong directional variations in Vp related to foliation. The reflector is located at a break in the deformed interval where highly deformed amphibolites are juxtaposed against relatively undeformed gabbro.
The absence of a break in density or compressional P-wave velocity corresponding to the two lower VSP reflectors (560 and 760-825 m) demonstrates that these do not result from the intrinsic physical properties of the gabbros, whether the development of a strong preferred mineral fabric, or variations in density. Rather, they must correspond to another effect such as the presence of faults and fractures. This is confirmed by examination of the recovery record from Hole 735B (Fig. F22). Recovery in both massive fine- and coarse-grained, foliated and unfoliated gabbros was high, generally close to 100% in Hole 735B. Drilling rates, however, varied considerably in these lithologies, dropping dramatically for fine-grained intervals. Although the low recovery at the top of the hole likely reflects lack of drill-string stability, elsewhere, where the rocks are believed to be highly fractured, drilling rates increased dramatically, and recovery dropped to as low as 31%. The precise coincidence of the two lower VSP reflectors with intervals of dramatically reduced recovery, therefore, apparently confirms the hypothesis that these are highly fractured zones of rock corresponding to some form of faulting associated with the late uplift of the platform.
The downhole measurements performed at the end of the leg show that there are distinct boundaries between each of the top six litho-stratigraphic units of Hole 735B. Logging data also suggest that there are two faults centered at approximately 555 and 565 mbsf which are 2 and 4 m thick, respectively. These log measurements show reduced velocities, densities, resistivities, and elevated porosities at these intervals. These zones also correlate with the approximate depth of a seismic reflector identified during the Leg 118 VSP experiment. Data from the Formation MicroScanner (FMS) processed postcruise show a marked improvement over the initial raw data obtained at the end of the leg. Preliminary analysis shows that several hundred structural features strike generally west-northwest from 280º to 310º. The dip azimuth of these features ranges from 340º to 20º with several features also dipping from 180º to 220º; the magnitudes mostly range from 10º to 50º. However, the degree to which the high concentration and magnetization of the oxide minerals influence the FMS magnetometer, which is essential to orient the structures observed in the FMS log throughout the upper 600 m of Hole 735B, is not yet known. Preliminary reorientation of core pieces shows a good correlation between Leg 118 borehole televiewer downhole tool data, digital unrolled core images, and Leg 176 FMS logs. A steep fracture dipping to the west is clearly identified in Core 118-735B-77R (403.5-409.5 mbsf) and the oriented logs.