A series of four large earthquakes with epicenters very close to one another occurred in 1990--1992 in the Weber reion of the southeast coast of the North Island, New Zealand. The region is one of oblique plate convergence and subduction, the plate interface being at about 20 km depth. The first event, February 19, 1990, ML 6.1, Mw 6.2, occurred within the upper part of the subducting Pacific plate on a steeply northwest dipping normal fault, as defined by relocated aftershocks. The mechanism of this event reflects downdip tension within the subducting plate due to ''slab pull'' forces, as is common in New Zealand. The second event, May 13, 1990, ML 6.2, Mx 6.4, occurred on an imbricate fault in the overlying Australian plate, dipping shallowly northwest, with components of both thrusting and dextral strike-slip motion. The slip vector is similar to the direction of plate convergence, showing no arc-normal rotation as observed for subduction thrust events in other areas of oblique convergence. It is unlikely that the spatial and temporal closeness of these events is coincidence ( a similar pair occurred to the south in 1942), but the coupling mechanism is not clear, perhaps being related to aseismic slip downdip on the subduction interface. Two subsequent events, both of ML 5.5, occurred on August 15, 1990, and March 2, 1992, and extended the aftershock zone of the deeper first main shock to the northeast, matching the extent of the shallower second main shock. Mechanisms of aftershocks of the two larger events are quite variable but as a whole are similar to the relevant mainshock. The time history of the aftershocks of the first (deeper) event shows a period of quiescence beginning 35 days before the second (shallow) event; the mean magnitude of those events that did occur was high. This observation cannot be explained by station biases, weather, or changes in analysis methods and may be due to stress changes induced by precursory aseismic slip which would discourage aftershock occurrence.