Ridge subduction follows the approach of an oceanic spreading centre towards a trench and subduction of the leading oceanic plate beneath the overriding plate. There are four possible kinematic scenarios: (1) welding of the trailing and overriding plates (e.g., Aluk–Antarctic Ridge beneath Antarctica); (2) slower subduction of the trailing plate (e.g., Nazca–Antarctic Ridge beneath Chile and Pacific–Izanagi Ridge beneath Japan); (3) transform motion between the trailing and overriding plates (e.g., San Andreas Transform); or (4) divergence between the overriding and trailing plates (e.g., Pacific – North America). In case 4, the divergence may be accommodated in two ways: the overriding plate may be stretched (e.g., Basin and Range Province extension, which has brought the continental margin into collinearity (and, therefore, transform motion) with the Pacific – North America relative motion); or divergence may occur at the continental margin and be manifest as a change in rate and direction of sea-floor spreading because the pair of spreading plates changes (e.g., from Pacific–Farallon to Pacific – North America), spawning a secondary spreading centre (i.e., Gorda – Juan de Fuca – Explorer ridge system) that migrates away from the overriding plate.Mantle upwelling associated with sea-floor spreading ridges is widely regarded as a passive consequence, rather than an active cause, of plate divergence. Geological and geophysical phenomena attendant to ridge–trench interaction suggest that regardless of the kinematic relations among the three plates, a thermal anomaly formerly associated with the ridge migrates beneath the overriding plate. The persistence of this thermal anomaly demonstrates that active mantle upwelling may continue for tens of millions of years after ridge subduction. Thus, regardless of whether the mantle upwelling was active or passive at its origin, it becomes active if the spreading continues for sufficient time and, thus, must contribute to the driving mechanism of plate tectonics.