Crystalline Rocks

Studies of cumulate rocks from the crystalline layer of the oceanic crust (layer 3), and abyssal peridotites from the upper mantle beneath ocean ridges have significantly advanced our understanding of the processes of crustal creation at ocean ridges. Bloomer et al. [1991] interpreted the textural and mineralogical variations in gabbroic rocks sampled from Hole 735B as the product of near-closed system crystallization of a small magma body that resulted when the crystallinity of the system reached a point where the pore melts were trapped, as the system moved off axis. Natland et al. [1991] combined bulk gabbro compositions, glass compositions and the compositions of oxides, sulfides and associated silicates to constrain the liquid line of descent responsible for the gabbros in Hole 735B. The anomalously high abundances of globular sulfides in the oxide concentrates, and the high MnO contents of the ilmenites were attributed to the creation of Fe, Mn, and S-rich magmas by immiscible separation of siliceous and very iron-rich liquids. Dick et al. [1991] present a detailed study of a 500 m section of layer 3 sampled by Leg 118 of the Ocean Drilling Program in the Atlantis II Fracture Zone. This study addressed the apparent contradiction between the extreme iron-enrichment of many of the gabbros sampled by the drilling project, as well as through dredging at other localities, and the lack of both the large, long-lived magma chambers thought required to explain such enrichment, as well as the lack of erupted ferrobasalts along slow spreading ridges in general, and at the Atlantis II Fracture Zone in particular. The contradiction was resolved by demonstrating that large magma chambers are not required to generate extreme iron enrichment in plutonic rocks, as well as by arguing that Fe-rich erupted basalts are not found because the differentiated melts are only mobilized along fault zones and largely freeze before escaping to the surface.

Elthon et al. [1992] and Ross and Elthon [1993] reported strongly depleted compositional trends in minerals from cumulates from a variety of oceanic basins which indicate that the rocks crystallized from basaltic liquids that were strongly depleted in incompatible elements relative to any sampled MORB. The existence of such strongly depleted magmas are consistent with fractional melting models which predict that magmas of a much wider range of compositions than erupted MORB should exist beneath a spreading center. The depleted cumulates provide evidence that the strongly depleted magmas produced by melting the shallow, most depleted mantle, are emplaced into the lower crust prior to aggregation with the more enriched magmas generated by near-fractional melting within the upper mantle.

Johnson and Dick [1992] successfully modelled the abyssal peridotite and basalt suites from the Atlantis II fracture zone as complements produced during `open system, near-fractional melting' of the upper mantle (although they found that in fact very little `open system,' or exotic component was required in their near-fractional melting model). In contrast, Elthon [1992] proposed the controversial hypothesis that the chemical trends exhibited by the global abyssal peridotite data set are controlled principally by refertilization processes rather than by partial melting. His arguments were based on linear compositional trends (i.e., NaO vs MgO) in the global abyssal peridotite data set that are incompatible with melting processes. Bonatti et al. [1992] investigated small-scale variations in the compositions of mantle-derived peridotites from 0-15N on the MAR to infer upper mantle heterogeneity below the MAR in this region.