Chlorine and major elements in >400 mid-ocean ridge basalt (MORB) glasses from 20 suites are used to examine how spreading rate, magma flux, tectonics, and hydrothermal activity influences assimilation and crystallization beneath MOR. Crystallization depths were determined for fractionated glasses using published models that describe liquids saturated with olivine+ clinopyroxene+ plagioclase. Calculated depths are minima for the onset of crystallization and maxima for the completion of crystallization for each liquid. Glasses from fast spreading ridges and from medium and slow spreading ridges with low Na-8.0 define low-pressure liquid lines of descent (LLDs). Higher crystallization pressures and greater variability are obtained from slow and medium spreading ridges with high Na-8.0. Crystallization pressures do not vary regularly along individual segments. The correlation between average crystallization pressure and Na-8.0 suggests that magma supply (and perhaps mantle temperature) plays an important role in determining magma ascent and crystallization depths. Cl/K in glasses is an indicator of assimilation of hydrothermally influenced material. Suites of MORE with high crystallization pressures have low Cl/K: from below detection limits (approximate to 0.01) in normal MORE (NMORB) to about 0.05-0.08 in enriched MORE (EMORB). We propose that this trend defines the mantle limit of Cl/K and that higher values are related to assimilation. Cl/K is highest (up to 1.1) and is negatively correlated with MgO along the superfast spreading southern East Pacific Rise (EPR) and the propagating, low-Na-8.0 Galapagos Spreading Center (GSC) at 85 degrees W. Cl/K is also above mantle values, but is not well correlated with MgO, in MORE from fast and medium spreading ridges and from slow spreading ridges that have low Na-8.0 and low crystallization pressures, e.g., Reykjanes Ridge. Cl/K is not correlated with crystallization pressure for individual samples within any suite. We propose that the spreading rate and the extent of melting act together to determine the total magma flux to a ridge, which influences crustal temperatures and determines how magmas ascend. At the highest magma fluxes, Cl/K is correlated with MgO, consistent with continuous assimilation of material that has a uniform Cl content: crystallization and assimilation are steady state processes that occur in crustal magma bodies that are larger than the scale of crustal heterogeneity in Cl. The lowest magma fluxes occur on slow spreading ridges that have formed by small extents of melting. In this cooler environment, magma crystallizes at the base of the strong lithosphere, below the level of alteration, and then ascends rapidly with little crystallization at shallow levels, so Cl contamination is avoided. On slow and medium spreading ridges with high extents of melting, magma flux is intermediate and forms small or transient crustal magma bodies: whether a magma batch becomes enriched in Cl depends upon the particular crust that it encounters.
