A set of 42 whole-rock samples was selected from Hole 1105A and analyzed by X-ray fluorescence for major elements (Si, Ti, Al, Fe, Mn, Mg, Ca, Na, K, and P) and selected trace elements (V, Cr, Ni, Cu, Rb, Sr, Y, Zr, and Nb). The results are shown in Table T15. The samples were selected to represent the main lithologies in the core and generally weighed 20-30 g. Sample preparation and analytical procedures and techniques are outlined in "Geochemistry" (in "Igneous Petrology and Geochemistry") in the "Explanatory Notes" chapter. Thin sections were prepared from splits of all samples for which XRF analyses were obtained.
The sampling was designed to characterize the main lithologies as well as to illuminate their downhole variation. The results are shown in Figure F50 for selected major elements and in Figures F51 and F52 for selected trace elements, as well as MgO and Mg#. The analyzed samples are identified in the figures according to the main lithologies as oxide gabbro (solid squares), olivine oxide gabbro (solid circles), gabbro (Xs), and olivine gabbro (open circles), based on thin-section examination of each sample analyzed. The analyzed trace elements are above the lower limit of detection, except for most of the Rb and some of the Nb determinations. Loss-on-ignition (LOI) values were generally <2% and reflect the low extent of alteration in the gabbros.
The high Mg/(Mg+Fe) ratios and low incompatible element abundances (e.g., TiO2, P2O5, Y, and Zr) in the olivine gabbros and gabbros clearly identified these as cumulates when compared to typical Atlantis II Fracture Zone basalts (Natland et al., 1991). The low incompatible element contents, such as Nb, Zr, and Y, in most of the samples indicate that generally low amounts of intercumulus melt congealed within the cumulates, as also concluded by Natland et al. (1991) from a study of Hole 735B samples. The oxide gabbros and oxide olivine gabbros are compositionally highly variable in Fe2O3, TiO2, and certain trace elements in comparison to the gabbros and olivine gabbros. Nevertheless, the relatively high Fe2O3, TiO2, and P2O5 in some of the analyzed oxide gabbros indicate that they contain an overabundance of Fe-Ti oxide minerals and occasionally apatite in comparison to a melt composition, again indicating that these rocks are cumulate in origin rather than simply congealed melts.
The two main groups of gabbros (olivine gabbro and gabbros vs. oxide gabbros and oxide olivine gabbros) are for the most part clearly distinguished by their respective TiO2 and Fe2O3 content (Fig. F50). However, these two groups are not clearly defined in terms of Mg/(Mg+Fe) ratio, and indeed there is overlap. The olivine gabbros and gabbros have Mg/(Mg+Fe) values between 0.4 and 0.8, TiO2 generally <1.0 wt%, and Fe2O3 <10.0 wt%. In contrast, the oxide gabbros mostly have Mg/(Mg+Fe) values <0.5, TiO2 >1.0 wt%, and Fe2O3 >10.0 wt%. These differences in Fe and Ti are a simple reflection of the modal variability in the content of Fe-Ti oxide minerals and do not necessarily reflect different parental magmas, although it may be the case for some of the oxide gabbros, based on geochemical and petrologic work on Hole 735B samples (Bloomer et al., 1991; Natland et al., 1991; Ozawa et al., 1991). In cores from Hole 1105A, there appears to be more of a continuum in magnesium numbers between the oxide gabbros and oxide-free gabbros indicating that the core contains a wide spectrum of fractionation extent over small vertical distances in the core. The findings at Site 1105 are consistent with the results obtained during Leg 118 (Robinson, Von Herzen, et al., 1989) in the sense that both primitive and very evolved gabbroic rocks were recovered in close proximity to one another. One difference between the previous studies of Hole 735B and the present data from Hole 1105A is that some oxide gabbros contain very high P2O5 contents that may reach 4 wt% (e.g., Samples 179-1105A-4R-4, 57-62 cm, and 170-1105A-5R-1, 115-118 cm). Similar high P2O5 oxide gabbros were reported from the MARK region during Leg 153 at Site 922 (Shipboard Scientific Party, 1995). Such high concentrations indicate a maximum of ~10 wt% of modal apatite in the oxide gabbros and indicate a high extent of fractionation past apatite saturation.
The concentrations of some trace elements are positively correlated with the TiO2, Fe2O3, and P2O5 contents (Fig. F51). For example, some of the oxide gabbros have relatively high concentrations of Nb, Y, Zr, and Zn (see Shipboard Scientific Party, 1995). Yttrium and to a lesser extent Nb show positive correlations with P2O5 concentrations, indicating that these elements in part may be controlled by the amount of apatite present. The elements Zr and Zn show good to poor positive correlation with TiO2 and Fe2O3 and may reside in Fe-Ti oxide minerals, zircon, and sulfides. These correlations imply highly evolved residual melts and indicate that crystallization of Fe-Ti oxides, apatite, sulfides, zircon, and titanite were involved in the late stages of the differentiation and solidification processes.
The olivine gabbros of parts of Units I, II, and Unit III have relatively high concentrations of the compatible elements Ni and Cr that correlate with high MgO content but not with the Mg/(Mg+Fe) ratio (Figs. F50, F52). This may reflect relatively higher concentrations of modal olivine and pyroxene (or low plagioclase) in these intervals or significant changes in mineral compositions. It is also significant that for some incompatible trace elements, olivine gabbros of the oxide-rich Units II and IV tend to be relatively similar to the associated oxide gabbros (e.g., Zr and Ni).
It has been demonstrated that the variation in Mg/(Mg+Fe) for the olivine gabbros is a reflection of the compositional variation in the constituent minerals (i.e., olivine and clinopyroxene; Hebert et al., 1991, Bloomer et al., 1991; Dick et al., 1991a, 1992; Ozawa et al., 1991; Natland et al., 1991; Casey, 1997); therefore, the Mg/(Mg+Fe) ratio of whole-rock compositions can be used as a proxy for the cryptic variation in the mafic phases when Fe-Ti oxides are not present. Because the Mg/(Mg+Fe) ratio is calculated assuming that all iron is FeO, the presence of even small amounts of Fe-Ti oxides will lower the Mg/(Mg+Fe) ratio considerably and this no longer will reflect the mafic silicate mineral compositions.
The analyses of Hole 1105A cumulates illustrated in Figures F50, F51, and F52 suggest a relatively variable Mg/(Mg+Fe) ratio for the olivine gabbros that appears to overlap with the Mg/(Mg+Fe) ratio of some oxide gabbros. In contrast, the oxide gabbros and olivine gabbros generally show a range of low ratios, reflecting their high Fe-Ti oxide modal abundances. The continuum in the range and overlap of compositions among olivine gabbros, gabbros, and oxide gabbros may suggest a common heritage for each group, but this tentative interpretation must await detailed cryptic mineral chemistry studies. The range of compositions sampled is relatively similar to previous results from Leg 118 (Robinson, Von Herzen, et al., 1989; Dick et al., 1991c) and Leg 153 (Cannat, Karson, Miller, et al., 1995; Casey, 1997). The stratigraphic position of the Site 1105 gabbros, relative to the Site 735 gabbros, is at present uncertain and requires further detailed postcruise correlation efforts, but there may be significant overlap based on the initial attempts to correlate sections. There appears to be a relatively good correspondence between the results from Hole 1105A and the uppermost parts of Hole 735B in terms of Mg/(Mg+Fe) ratios and the presence of both primitive olivine gabbros and evolved oxide gabbros in close proximity over the same vertical intervals. Dick et al. (1991c) interpreted these two contrasting rock types to result from the mixing of primitive gabbros with highly fractionated oxide-rich intercumulus melts. The continuum of Mg/(Mg+Fe) ratios between oxide-free and oxide-bearing gabbros in Hole 1105A, however, was also noted for Hole 735B (Hebert et al., 1991). The question of the similar origin of both suites of rock clearly deserves very detailed studies of cryptic variation through the cumulate pile.
Hole 1105A provides an important, but restricted, section through part of the lower oceanic crust. These results, together with a much more extensive coring from Hole 735B (Robinson, Von Herzen, et al., 1989), represent the most comprehensive sampling of an oceanic gabbroic complex to date. The previous study of these cores has contributed significantly to how we view magma chambers, their solidification processes, and their interaction with spreading ridge tectonics (e.g., Bloomer et al., 1991; Dick et al., 1991c, 1992; Ozawa et al., 1991; Natland et al., 1991).
The core obtained at Site 1105 is composed of a sequence of gabbro, olivine gabbro, oxide gabbro, and oxide olivine gabbro intervals that are grouped into four lithologic units. These intervals and units do not show mutual intrusive relationships and are considered to have been an integral part of a growing cumulus pile at the floor, roof bottom, or walls of a magma chamber (or multiple chambers). The gabbros at Site 1105 are dominantly typical adcumulates and heteradcumulates with some evidence for trapped liquid crystallization in the form of zoning in plagioclase. This indicates that some rocks solidified with significant trapped melt. Some incompatible trace element abundances are elevated over that expected for adcumulates and would suggest that a portion of the core contains significant amounts of formerly trapped melt that solidified in situ. In addition, sections of the core, which are highly deformed by ductile deformation, commonly mask evidence of mineral zonation because of the extreme strain in remnant porphyroclasts that leads to undulose extinction and recrystallization. Igneous lamination is scarce, but modally defined layers and variation are present throughout the core, particularly defined by changes in olivine and Fe-Ti oxide modal abundances. Nevertheless, a major criteria in defining igneous layering is grain-size variations from fine-grained to pegmatitic varieties of gabbroic rocks with gradational to planar interfaces seen on a centimeter to meter scales.
Some of the gabbroic rocks drilled at Site 1105 were subjected to crystal-plastic deformation to various extents and were locally dynamically recrystallized at temperatures possibly as high as granulite facies grade. Such plastically deformed zones within the gabbroic section may likely have provided pathways through which seawater infiltrated into the lower oceanic crust. In addition, gabbroic rocks were subjected to alteration and metamorphism dominantly under greenschist facies conditions.
Despite their variability and cumulate nature, preliminary analysis of the Hole 1105A gabbros does not show a simple and systematic pattern of mineralogical and compositional variation with relative stratigraphic position. The core is composed of an intercalated interval or groups of intervals having distinct mineralogical and compositional ranges. One group is a relatively more primitive lithologic type composed of gabbro and olivine gabbro. The other group is an evolved lithologic type composed of oxide gabbro and oxide olivine gabbro. This spatial arrangement of the relatively primitive and evolved gabbros is very different from the systematic mineralogical and compositional variations commonly documented in gabbroic complexes in ophiolite and continental settings. Although the recovered rock types mimic the mineralogical variability and crystallization range expected in a fractionating basaltic magma chamber and there appears to be a compositional continuum, the close spatial arrangement of oxide gabbroic rocks and oxide-free gabbroic rocks does not lead to such a simple interpretation as classically portrayed in layered basic intrusions on land.
Despite their close association in the core and their contrasting mineralogical and compositional features, there is surprisingly limited evidence for interactions and mixing between the two compositional end-member components. The package of cumulates appears to have shared a common crystallization, postcumulus, and deformational history despite distinctly different parental liquid compositions (basalt and ferrobasalt). There is no conclusive evidence preserved in the Hole 1105A gabbros to indicate that the formation of the oxide gabbros postdates or is intrusive into the olivine gabbros. These findings are generally similar to the results obtained from Site 735 located only 1.2 km from Site 1105. Dick et al. (1991c) interpreted the intercalated oxide-rich and oxide-free gabbros to result from syntectonic differentiation caused by compaction of a partially molten olivine gabbro. They envisioned that intercumulus melt would migrate laterally and perhaps vertically into and along locally distributed ductile shear zones. This migrating melt would react with olivine gabbro, precipitate Fe-Ti oxide minerals, and finally solidify after cessation of deformation and thereby transform the olivine gabbro to ferrogabbro (Natland et al., 1991). Further studies of the 1105A core will be conducted to test the viability of such models.