Thesis (Ph. D.)--Joint Program in Oceanography (Massachusetts Institute of Technology, Dept. of Earth, Atmospheric, and Planetary Sciences; and the Woods Hole Oceanographic Institution), 2006.Includes bibliographical references.Between 90-25° E on the ultraslow-spreading Southwest Indian Ridge lie two sharply contrasting supersegments. One 630 km long supersegment erupts N-MORB that is progressively enriched in incompatible element concentrations from east to west. The second 400 km long supersegment contains three separate volcanic centers erupting E-MORB and connected by long amagmatic accretionary segments, where mantle is emplaced directly to the seafloor with only scattered N-MORB and E-MORB erupted. Rather than a major break in mantle composition at the discontinuity between the supersegments, this sharp contrast in geometry, physiography, and chemistry reflects "source" versus "process" dominated generation of basalt. Robust along-axis correlation of ridge characteristics (i.e. morphology, upwelling rate, lithospheric thickness), basalt chemistry, and crustal thickness (estimated from gravity) provides a unique opportunity to compare the influence of spreading geometry and rate on MORB generation. What had not been well established until now is the importance of melting processes rather than source at spreading rates < 20 mm/yr.(cont.) Along the orthogonally spreading supersegment (14 mm/yr) moderate degrees of partial melting effectively sample the bulk mantle source, while on the obliquely spreading supersegment (7-14 mm/yr) suppression of mantle melting to low degrees means that the bulk source is not uniformly sampled, and thus "process" rather than "source" dominates melt chemistry.by Jared Jeffrey Standish.Ph.D.