|Abstract or Summary
- Prior to eruption at mid-ocean ridges, melts must travel through >6 km of crust from their origin in the mantle. The final composition of the melts is dependent on both the melting conditions and magmatic processes within the crust. While mid-ocean ridge basalt (MORB) glasses are commonly used to infer melting conditions, including upper mantle composition, extent of melting and melt aggregation, evidence suggests that the final erupted MORB melt is significantly affected by magma mixing, fractionation, and melt-wallrock interactions, which significantly alter the original composition of the melt. A yet unresolved problem in interpreted erupted basaltic glass is the relative role of mantle processes and crustal processes in their contribution to MORB. The three manuscripts in this dissertation use plagioclase phenocrysts in plagioclase ultraphyric basalts (PUBs) to attempt to address the nature of the magmatic processes occurring within the oceanic crust prior to eruption. Plagioclase megacrysts and host glasses were analyzed for Sr-isotopic compositions in multiple samples from multiple spreading ridges. The resultant data reveals that the plagioclase megacrysts are often in isotopic disequilibrium with the melt they are erupted in. Some of the samples also display significant inter- and intra-crystalline disequilibria. As Cl concentrations in melt inclusions were low, the observed isotopic heterogeneity cannot be caused by interaction with seawater. Rather, the varied Sr-isotopes in plagioclase are the result of growth from distinct mantle components and some physical process is responsible for entraining the plagioclase into the final melt. To better interpret this isotopic disequilibria recorded in plagioclase, this physical process which is responsible for the eruption of these plagioclase-rich magmas must be understood. No systematic difference exists between the lavas that erupt abundant plagioclase phenocrysts and those lavas from the same ridge segments that are aphyric. Because these plagioclase crystals are typically high anorthite (>An₈₀), the phenocrysts are typically more dense than the melt they are erupted in. Thus a simple floatation model is not sufficient to explain the eruption of PUBs. PUBs only erupt in locations without currently imaged magma chambers, mainly fast spreading ridges, because prolonged residence in a magma chamber would cause the dense plagioclase to settle out. PUBs may erupt, however, when magmas carrying plagioclase crystals travel through conduits without magma chambers and do not stall in the crust for extended periods of time. Finally, the plagioclase megacrysts in this study were analyzed for trace element concentrations, in addition to melt inclusions, olivine and host glass analyses. Collection of this suite of data allows for a more complete representation of the pre-eruptive history of individual basalts. Using appropriate partition coefficients for anorthitic plagioclase, liquids that were in trace element equilibrium with the megacrysts can be calculated and then compared to the array of lavas that are erupted on the surface. Liquids in equilibrium with plagioclase are more primitive than erupted MORB glasses in the same region and are typically not parental to the host lava in which the megacrysts were erupted. These primitive liquids often appear to be parental to other erupted lavas by fractional crystallization alone. Plagioclase equilibrium liquids that cannot be modeled as parental to the host lava are typically display isotopic disequilibrium as well. For samples that display trace element heterogeneity, many of the liquids in equilibrium with plagioclase can be modeled as resulting from progressive melting of a single mantle source.