Although there is general agreement that plate motions are maintained by thermal convection of the Earth's mantle, we still know rather little about the spatial and temporal details of the convective flow. Thus any observations that can be used to constrain the nature of this circulation are of particular interest. Furthermore, at significant palaeoceanographic gateways, vertical movements generated by changes in convective circulation are likely to have had an important impact on ancient thermohaline circulation. In the North Atlantic Ocean, the V-shaped ridges, which lie on either side of the mid-oceanic ridge south of Iceland, have long been regarded as a direct manifestation of transient convective behaviour within the Icelandic plume. Recent analysis has shown that the chemistry of rocks dredged from the mid-oceanic ridge combined with the bathymetric expression of the V-shaped ridges require hot (30-35 degrees Centigrade) pulses of plume material to spread out radially beneath the North Atlantic Ocean. These hot pulses are linked with temporal changes in vertical uplift which in turn moderate the overflow of Northern Component Water (i.e. the ancient precursor of North Atlantic Deep Water) across the Greenland-Scotland Ridge. Thus mantle convection has a direct and measurable effect on ancient thermohaline circulation in a key portion of the global conveyor. Despite these important advances, the detailed structure and chemistry of the V-shaped ridges are very poorly known. Models are reliant upon poor quality seismic reflection profiles which were acquired in the 1960s. We will acquire, process, interpret and model a set of seismic reflection profiles across the V-shaped ridges where they are best expressed southwest of Iceland. Ancillary underway data will include multi-beam bathymetric, gravity and magnetic profiling, and dredging of basalts. Our survey will achieve two important aims. First, a set of high resolution profiles which image the detailed structure of sedimentary cover together with the sediment-basement interface at each of the proposed IODP driill sites will constitute a site survey for the planned drilling leg. Secondly, these profiles will be used in their own right to determine the detailed chronology of transient plume behaviour over the last 15 million years. This chronology will be combined with ongoing studies of basalt geochemistry and palaeoceanography. Our results will have an important impact upon the quantitative understanding of Neogene overflow of Northern Component Water.