Gabbroic to granophyric rocks sampled and analyzed in this study show complex chemical and lithologic stratigraphies with highly fractionated oxide gabbroic rocks intimately interlayered with more primitive gabbro and olivine gabbro. Both normal and inverse cryptic geochemical trends exist downhole through the core. The vertical chemical stratigraphy does not appear significantly affected (reequilibrated) by porous late melt migration through the cumulate pile except locally along narrow late melt flow zones that upset the mineral chemistry within a single thin section. Chemical variation in all silicate minerals correlates well downhole. Geochemistry appears to document significant melt trapped within the many samples in the cumulate pile, although textural evidence of trapped melt is not abundant in the form of highly zoned minerals. The trapped melt abundance overall is similar to upper isotropic gabbroic rocks in the ophiolite complexes, possibly suggesting the rocks crystallized in the upper plutonic crust or near the roof zone of a magma chamber complex. The most primitive rock silicate minerals are not near equilibrium with mantle, indicating significant fractionation of magmas took place prior to crystallization of the most primitive gabbroic rocks in the section. The bulk average major element chemistry of the core may be close to the regional basaltic average, but bulk rock incompatible element averages of the core are significantly depleted with respect to the basaltic average, helping to demonstrate that the plutonic rocks solidified as cumulates and not as congealed liquids or average of any basaltic liquid.
Oxide-rich and primitive oxide-free gabbroic rocks are intimately interlayered in the Hole 1105A section on a small scale, as demonstrated by FMS images (Miller et al., Chap. 3, this volume). Likewise, cryptic chemical variations show extreme fine-scale variations (scales of tens of meters or less). The mechanism of boundary layer fractionation may help to explain this complex juxtaposition, as simple perfect fractional crystallization fails to explain the rock chemistries observed or their complex juxtaposition, unless the model of Dick et al. (2002) is used, which assumes that highly allocthonous late-stage melts infiltrate primitive gabbro of any composition in the cumulate pile as melts rise from below, predominantly along shear zones. This is a plausible explanation, but we find that, in general, oxides are present where silicate phases are highly fractionated and we also question why late-stage and dense iron-rich melts would rise from below through thick cumulate sections. Silicate mineral evidence shows that their fractionation state is not high enough to have caused precipitation of oxides utilizing a PFX model. A boundary layer fractionation model could produce the complex juxtaposition of rocks if the interior of the chamber was more primitive and the exterior of the boundary layer was strongly fractionated. Local melt migration of dense ferrobasaltic melts could explain certain melt flow zones, reaction relationships, and cross-cutting oxide lenses noted in both Holes 1105A and 735B. Reaction between an evolved melt and existing silicate minerals could cause silicate compositions to reequilibrate partially to more evolved compositions and precipitate oxides, but as shown, in some cases the residual melt phase would have to be somehow eliminated from the sample to explain the low incompatible element abundances. The boundary layer model would not have these constraints, as the residual liquid could migrate to the main part of the chamber.
Crystal-plastic deformation textures indicate increasing strain downhole in oxide-rich gabbroic rocks, and this strain appears localized dominantly, although not exclusively, where abundant oxides are present. This may indicate that strain localization in the section is related to the weak rheology of the oxide minerals or a melt phase that precipitated abundant oxides as they penetrated through the cumulate section. The high-temperature subsolidus textural overprint makes it impossible to identify a hypersolidus development phase in shear zone formation through the section. The nature of the progression of recrystallization and microstructural development during high-temperature crystal-plastic deformation in the gabbroic section suggests that the strongest to weakest minerals are in the progression from clinopyroxene to olivine to plagioclase and opaque Fe-Ti oxides. Plagioclase is a more significantly recrystallized silicate phase, more than olivine or clinopyroxene in moderately strained rocks, and clinopyroxene is the dominant porphyroclastic phase in highly deformed rocks. Synkinematic brown amphibole rimming clinopyroxene porphyroclasts and replacing neoblasts suggests the shear zones undergo high-temperature hydration during deformation and that plagioclase rheology may be affected by these hydrating fluids. Significant clinopyroxene recrystallization is observed only in highly strained rocks. Several mylonite zones have been identified in the section, and some are associated with large excursions in mineral and whole-rock chemistry. One shear zone examined appears to lack evidence of having abundant trapped melt, although the adjacent undeformed olivine gabbro shows abundant trapped melt.
As is typical of studies of plutonic sections, significant questions raised and remaining concern the igneous and structural evolution of the rocks in both Holes 1105A and 735B. Whether extensive oxide-rich zones originated within boundary layers at the edges of magma chambers or, alternatively, deep within a cumulate pile where porous flow allowed evolved intercumulus melts to infiltrate and precipitate oxides is a question that remains. The evidence presented here indicates that either may be a viable explanation but that neither can be excluded. Another related question is a classic chicken-and-egg argument of whether the oxide-rich zones that have undergone deformation are indicators of hypersolidus flow within shear zones where evolved melts accumulated and precipitated oxides or are zones where cumulate crystallization processes led to oxide precipitation within a boundary layer and creation of weak rheologic zones in the solid. In the solid state, the weaker rheology of oxide-rich zones could have acted to localize subsolidus strain as the core complex evolved and was structurally modified, partially hydrated along localized shear zones, cooled, and unroofed.
The ability to examine the short section of gabbro in Hole 1105A in the context of studies of Hole 735B allows extensive comparisons. To date, the bulk of the evidence suggests that the oxide-rich zones in both holes can be correlated laterally, if not necessarily physically, in terms of the processes that produced them. A series of short holes drilled between Hole 1105A and Hole 735B may help to answer further questions about the framework and size of subaxial magma chambers and the magmatic and structural processes operative in plutonic complexes unroofed at slow spreading centers.