Plate divergence at mid-ocean ridges results from a combination of three main processes: magma emplacement, faulting and hydrothermalism. Melt fluxes at fast and most intermediate spreading ridges are high enough that the axial lithosphere (the plate between the plates) remains thin, or at least thinner than the cumulative melt thickness (melt flux divided by expansion rate) per unit length of ridge. In this configuration, the magma can fully accommodate plate divergence. This is not the case for slow and ultraslow ridges, where melt fluxes are lower, resulting in colder axial geotherms and an axial lithosphere that is in the general case thicker than the cumulative melt thickness per km of ridge. In this configuration, faults must accommodate a significant part of plate divergence, while magma may be emplaced at a range of depths, over the full thickness of the axial plate. This creates the conditions for the common exhumation of mantle-derived peridotites and the formation of a composite (variably serpentinized peridotites and magmatic rocks) oceanic crust. Melt fluxes, and probably melt emplacement depths, are also highly variable at slow-ultraslow ridges. This allows for complex interactions between magma and faults, and between magma and hydrothermal circulation, resulting in spatially and temporaly variable spreading modes (the combination of the mid-ocean ridge tectonic, magmatic and hydrothermal processes that determine the composition and structure of the oceanic lithosphere). In this presentation, I revisit these concepts, discuss the extent to which they have been tested by experiments and modelling, and point to several remaining questions.
