During 5 days of drilling, Hole 1105A penetrated 158 m into the Atlantis Bank gabbroic massif. The measured cored interval was 143 m, starting 15 m below the seafloor. Core recovery was 118.43 m of gabbroic rocks (total recovery = 82.8%), in line with the high recovery rates in Hole 735B. Together with logging results (Miller et al., Chap. 3, this volume; Zarian et al., 2002), this high core recovery provides nearly complete coverage of the rock types drilled and a comprehensive view of pseudostratigraphy and structure in the gabbroic section cored (see the Leg 179 Initial Reports volume; Pettigrew, Casey, Miller, et al., 1999).
The recovered cores record a wide variety of rock types ranging from gabbro, oxide gabbro, olivine gabbro, and oxide olivine gabbro to fewer troctolitic gabbro, gabbronorite, and felsic rocks such as trondhjemite. A total of 141 lithologic intervals (Fig. F5) were defined within the core on the basis of distinct changes in mode, modal proportions, grain size, and/or texture. Four rock types were delineated by the shipboard scientists in constructing Figure F5: gabbro, olivine gabbro, oxide gabbro, and oxide olivine gabbro. Well-defined igneous layer contacts or structural boundaries of these intervals are preserved in many sections of the core. The highly layered nature of the gabbroic rocks documented within the core is supported by visual and thin section examination of the core, high-quality continuous FMS logs of the borehole (e.g., Fig. F6), other logs, and whole-core magnetic susceptibility measurements (Pettigrew, Casey, Miller, et al., 1999).
Gabbroic rocks constitute the majority of the recovered igneous lithologies (>99% by volume) with only a very minor component of felsic veins (<1%) (Pettigrew, Casey, Miller, et al., 1999). The gabbroic rocks are characterized by mostly medium- to coarse-grained granular textures and display modal, textural, and grain size variations. Regularly to irregularly developed igneous layering and rare igneous lamination are present within intervals, and many of the contacts between defined intervals are marked by mesoscopic layer boundaries (Pettigrew, Casey, Miller, et al., 1999). Grain size variations are extreme within the core: 2% pegmatitic gabbro, 67% coarse-grained gabbro, 28% medium-grained gabbro, and 3% fine-grained gabbro (Pettigrew, Casey, Miller, et al., 1999). Modally, the gabbroic rocks range from gabbro (36%) to olivine gabbro (43%), oxide gabbro (17%), oxide olivine gabbro (4%), and minor gabbronorite (Pettigrew, Casey, Miller, et al., 1999; Thy, Chap. 2, this volume). The principal gabbroic rock types are defined based on the presence or absence of olivine, orthopyroxene, and/or Fe-Ti oxide minerals. The contacts of the intervals were visually characterized to be mainly caused by modal, modal proportion, grain size, and textural changes or combinations of these. As detailed modal analysis was conducted on the samples included in this synthesis, we are able to further refine rock types as they are presented in downhole plots.
Late magmatic veins of leuco-gabbro to granophyre are present throughout most of the core. They represent a large compositional range extending to diorite and trondhjemite, constituting the more silicic end-members. These veins are often thin, several centimeters or less in width, so they are not defined as separate intervals. Despite their volumetrically minor importance, they are of considerable interest because they may represent the final products of magmatic differentiation and strongly influence mass balance calculations (see the Leg 179 Initial Reports volume; Pettigrew, Casey, Miller, et al., 1999). Most of the recovered core is characterized by random orientations of coarse-grained, elongated, or tabular minerals such as plagioclase and clinopyroxene with no sign of compaction or preferred dimensional orientation. However, weak magmatic lamination can be observed locally, and evidence of ductile shear zones and crystal-plastic deformation is observed in localized intervals throughout the cored section. The distribution of magmatic lamination, layering, deformation, and magmatic veining is shown in Figure F7 (Pettigrew, Casey, Miller, et al., 1999).
Minerals identified as secondary or metamorphic under the microscope include talc, serpentine, smectite, magnetite, calcite, epidote, chlorite, brown hornblende, green hornblende, actinolite, and plagioclase. In addition, orange-brown to pale greenish brown phlogopitic mica is present in trace amounts in some olivine gabbro samples. It is not certain, however, whether this mica is primary or secondary (Pettigrew, Casey, Miller, et al., 1999). The secondary and metamorphic mineralogy of the core is limited to fewer than the 12 mentioned mineral species. This apparent simplicity in mineral assemblages microscopically examined is partly because most gabbroic samples from which thin sections were made were not taken from the late magmatic veins or their vicinities, and only optical determination of the minerals was conducted. Samples analyzed during this study were generally little altered.
The 141 lithologic intervals were grouped into summary Units I–IV in order to facilitate description and comparison with the core recovered during Leg 118 (Robinson, Von Herzen, et al., 1989; Dick et al., 1991). Units I and III are characterized by more abundant oxide-free olivine gabbro and gabbro, and Units II and IV are characterized by more abundant oxide-rich oxide gabbro, oxide olivine gabbro, and oxide gabbronorite. The summary units and intervals illustrated in Figure F5 are mainly based on modal oxide abundances and grain-size characteristics. Magnetic susceptibility is strongly correlated with oxide abundance in Holes 1105A and 735B (e.g., Natland, 2002). In particular, Unit II in Hole 1105A and Unit IV in Hole 735B are strongly similar in abundances of oxide-rich gabbroic rock types and high magnetic susceptibilities. They, in fact, may be correlative units within the Atlantis Bank gabbroic massif. Although Unit II in Hole 1105A is significantly thicker than Unit IV in Hole 735B, the magnetic susceptibility values and lithologies are strikingly similar in each hole. Both contain vertically persistent zones of gabbroic rocks with high average oxide content and significant thickness.
These oxide-rich units may be part of the same facies, with the thickness of the unit increasing toward the center of the Atlantis Bank platform (i.e., toward Hole 1105A). Alternatively, the differences in thickness from one hole to the other may result from structural complexities in one or both holes, caused by shear zones noted in each hole (e.g., three mylonites in Hole 1105A) (see "Hole 1105A Deformation Extent and Core Microstructure," and discussions of Hole 735B in Niu et al., 2002, and Dick, Natland, Miller, et al., 1999). Figure F8 depicts magnetic susceptibility values for Hole 1105A and the upper part of Hole 735B. The unit boundaries in Hole 1105A strongly correlate with sharp decreases or increases in average magnetic susceptibility, similar to the definition of Unit IV in Hole 735B (Natland, 2002). Further refinements and core-logging correlations between magnetic susceptibility, oxide-bearing lithologies, and FMS data for Hole 1105A are presented in Miller et al. (Chap. 3, this volume).
The nature of the contacts and the petrographic variability between and within individual intervals provide important constraints on the magmatic and tectonic history of Hole 1105A gabbros. Natland and Dick (2002) argued, based on the bulk rock data from the Leg 179 Initial Reports, against correlation between Unit II of Hole 1105A and Unit IV of Hole 735B because they suggested (1) that the range of TiO2, SiO2, Zr, and Y bulk rock values in Unit II of Hole 1105A did not correlate or reach the extreme levels (lows or highs) typical of Unit IV in Hole 735B and (2) that Unit II of Hole 1105A contained more primitive olivine gabbro than Unit IV in Hole 735B. Based on the more comprehensive data set for Leg 179 reported below, we are not able to confirm significant differences in the range of compositions between Hole 1105A (Unit II) and Hole 735B (Unit IV), with TiO2 contents in Hole 1105A ranging as high as 7.56 wt% and SiO2 contents as low as 33.37 wt% where oxides are abundant. The abundances of Zr and Y in Hole 1105A Unit II samples are comparable and in some intervals exceed the ranges documented in Unit IV samples from Hole 735B. We also note that the downhole variation across Unit IV of Hole 735B may not be as well documented as the oxide gabbro interval in Hole 1105A. Because the average sampling interval in Hole 1105A is significantly less than 1 m, there are nearly three times as many bulk and mineral analyses over the same interval of Hole 1105A when compared with Hole 735B. Therefore, the three cryptic units defined by Natland and Dick (2002) in Hole 735B in this interval may underestimate the number of more primitive intervals present. In particular, the three cryptic chemical trend lines defined for Unit IV in Hole 735B often cross trend lines that could be defined by adjacent samples, perhaps indicating underestimation of the scale of variation. Therefore, we cannot as readily dismiss the possibility that the units are related in some fashion and could, in general, be correlated laterally as part of the same magmatic facies or units.
Modal mineral abundances of primary phase olivine, clinopyroxene, plagioclase, opaque oxides, orthopyroxene, and inverted pigeonite are presented in Table T1 and facilitate comparisons with whole-rock and mineral chemistry presented below. Scans of each thin section are presented in the volume "Supplementary Material" for comparison, and Figure F9 shows six examples of igneous cumulate textures largely devoid of deformation features. Although most modes of moderate to fine grain-sized thin sections are considered reliable, there were several very coarse gabbroic rocks analyzed for which we consider the modes to be unreliable based on analysis of standard thin sections (e.g., Samples 179-1105A-9R-2, 17–20 cm, and 27R-4, 29–31 cm). These poorly defined modes are indicated in Table T1 by asterisks. Limited sample sizes and the necessity to analyze the bulk chemistry of each sample prohibited more detailed analyses of these coarse samples with oversized thin sections; however, as in shipboard studies, the data provide information on the nature and relative proportion of the phases crystallizing downhole and in most cases appear to directly correlate with whole-rock chemistry. In addition, the data are useful in establishing when oxides, pigeonite, and orthopyroxene begin to fractionate in the crystallization sequence. Except for samples that may have had mixed lithologies (silicic veins and compositional layer boundaries), rocks were classified based on modes of magmatic minerals as (1) oxide-free olivine gabbro and gabbro, (2) oxide-bearing (0.1%–5% opaque oxides) olivine gabbro and gabbro, and (3) oxide (>5% opaque oxides) olivine gabbro and gabbro. This classification conforms with the International Union of Geological Sciences (IUGS) classification in utilizing the term "oxide" as a modifier if the accessory mineral is >5%. In order to allow comparison with usage in Hole 735B (Dick, Natland, Miller, et al., 1999) and prior usage (Thy, Chap. 2, this volume) in Hole 1105A, in which rocks are labeled oxide if they are >1% opaque oxides (Dick et al., 2002), we simply indicate samples with any "magmatic" oxides in the range of 0.1–5 mod% as "oxide bearing." Exsolved oxide in clinopyroxene and oxide associated with alteration products of olivine and clinopyroxene were not included in the oxide mode. Gabbronorites were also described by Thy (Chap. 2, this volume), but although orthopyroxene or inverted pigeonite is present in thin sections, the samples described are not sensu stricto gabbronorites as noted by Thy. We agree with Thy (Chap. 2, this volume) that the presence of orthopyroxene or inverted pigeonite (or its inversion products) are important petrogenetic markers and worth noting, but they were sampled in a limited number of shipboard sections and they can be depicted in downhole modal plots by their presence in certain oxide-rich intervals without presenting a rock name that does not necessarily fit the mode. Orthopyroxene-bearing samples were more extensively described in Hole 1105A by Thy (Chap. 2, this volume) and in Hole 735B (Robinson, Von Herzen, et al., 1989; Dick, Natland, Miller, et al., 1999), but after review of modal and bulk data we refrain from using rock names that overlap because these phases are present within either oxide-bearing or oxide gabbros. They are also associated with dioritic rocks, resulting in multiple overlap in rock names and modifiers. Thy (Chap. 2, this volume) notes the overlap in terminology; he states gabbronorites are also oxide gabbros in his chosen terminology. We also have not used the term "apatite" gabbro as in Thy (Chap. 2, this volume) because we find that the more appropriate term for these gabbros is actually oxide gabbro because of the high abundance of oxide minerals associated with apatite. The association of high abundance of accessory apatite (<5%) with high modal abundance of magmatic oxide (>5%) renders the two names overlapping, and the modifier conflicts with IUGS nomenclature. Thus, we chose the IUGS nomenclature for clarity. This is also consistent with Leg 179 Shipboard Scientific Party (1999b). The terms "melanocratic" and "leucocratic" preface rock names if modal abundances of mafic or felsic phases warranted. However, these terms were seldom used, and mainly in shipboard visual core descriptions (VCDs). No troctolitic rocks were observed in the section. Granophyric veined gabbro, quartz diorite, and trondhjemites were also analyzed for bulk chemistry, but heterogeneities (intimate mixtures of gabbro and granophyre), alteration, or deformation in many of these samples prohibited accurate modal analysis and in some cases resulted in difficult separations and analysis. The irregular nature of the veins and the common inclusion of small xenoliths of gabbroic rocks complicated their analysis. In addition, thin sections were not returned on these mixed silicic samples and a few gabbroic sections because such heterogeneities caused plucking of slide fragments during the trim and grinding process (total of 11 sections lost). Thus, modal and mineral chemistry data are not available for these 11 samples because of limited sample availability, as most of the sample was used for bulk analysis. They are available for 137 of our samples. We were, however, able to collect whole-rock data on all 148 samples included in this study.
The relative abundances of various gabbroic rock types sampled in Hole 1105A are displayed in a histogram in Figure F10. Downhole variations in modal abundance are illustrated in Figure F11 and include both our modal analysis (137 sections) and shipboard modal analysis (42 sections). The data from this study are included in Table T1, which shows the considerable amount of fine structure within the hole where rock types and modal abundances vary significantly on a small scale (1 m or less). Plagioclase and clinopyroxene dominate modes in the gabbroic section. Modal layer boundaries are defined by the appearance or disappearance of modal oxide, olivine, orthopyroxene, or inverted pigeonite. Small-scale variation is also clear from magnetic susceptibility measurements and the large number of lithologic intervals established during visual core description. Although sampling in Hole 735B was much coarser, the intimate juxtaposition and interlayering of relatively primitive olivine gabbro and oxide gabbro in Hole 735B led Dick et al. (2002) to propose multiple small melt flow channels and oxide gabbro intrusions to explain such fine-scale variation. We will examine this hypothesis in more detail with the finer scale of sampling available to us from Hole 1105A. Figure F11 also displays good correlation with shipboard definition of major lithologic units in that there is a general lack of oxide minerals in Units I and III and abundant oxides in Units II and IV. The average gabbroic modal composition of the cored interval, based on available modes from Hole 1105A, consists of 56.98% plagioclase, 32.37% clinopyroxene, 6.79% olivine, and 3.84% Fe-Ti oxides.