Coincident multichannel seismic reflection and refraction data acquired during a wide-aperture two-ship experiment provide evidence for a complex crust-mantle (C-M) transition under Oahu, Hawaii. Several large-aperture common depth point lines and three expanding spread profiles suggest the existence of an anomalously thick (3–6 km) C-M transition zone underneath the volcanic ridge which extends for distances of 100 km to the north and south from the center of Oahu. The anomalous C-M transition may represent a plutonic complex which intruded into the upper mantle and the lower crust in a 200-km-wide area centered at Oahu. The existence of such a large volume of intrusions near the base of the crust implies that the surficial expression of volcanism constitutes only a small fraction of the amount of melt generated at depth under the Hawaiian Islands. This interpretation is in accord with previous petrological models which predict trapping and accumulation of upwelling magma at and below the Moho. We have constructed a model which suggests that the interaction between the upwelling magma and the lithospheric flexural stress field may modulate the characteristic eruption history of Hawaiian volcanoes. In particular, the model for the plane stress field which accompanies the flexure of the oceanic crust around island chains indicates that the stress field under individual volcanoes varies considerably with its position relative to the tip of the chain. As a Hawaiian-sized volcano develops, the magnitude of deviatoric compressive stresses under it is probably sufficient to block the conduits of the upwelling magma within the oceanic crust and to terminate eruptions. Further upwelling magma is predicted by the models to be ponded at the base of the crust. Resumption of posterosional volcanism seems to occur at a constant distance behind the center of active shield volcanism, as the horizontal compressive stresses along the axis of the chain are released. Observed orientations of dikes of this volcanic phase agree with the directions of the maximum calculated stresses. Our model implies that magma upwells over a 300-km-wide zone and that the oceanic plate may not be fractured under the islands.