A marine seismic system for recording near-vertical incidence to wide-angle reflected waves and refracted waves has been used to obtain detailed crustal structure off Canada's west coast. Profiles about 20 km in length were recorded in three regions: the Hudson '70 survey area near 51 °N, 133 °W; west of Queen Charlotte Sound; and in northern Cascadia basin, west of central Vancouver Island. In the first area, the interpretation was completely consistent with the Hudson '70 study, but more detail was provided for the upper crust. About 0.6 km of sediments with velocity 2.4 km/s overly layers 2A and 2B with velocities of 4.0 and 5.5 km/s and thicknesses of 1.1 and 1.5 km respectively. The oceanic layer has a velocity of 6.8 km/s. Off Queen Charlotte Sound, the sediments vary in thickness from 3–3.5 km and are divided into an upper sequence with low velocities (2.1 and 2.8 km/s) and a lower sequence with higher velocities (about 4.2 km/s). Basaltic basement beneath the sediments has a velocity of 5.85 km/s. The seismic data indicate that sediment deposition has been complex, possibly interspersed with thin basalt sills derived from a nearby spreading centre. On the basis of these and other data, Winona basin is proposed to extend northwestward as far as an imaginary line drawn landward from the trough between the Dellwood Knolls. In order to test this proposal and delineate in detail the total sedimentary section, high resolution reflection studies with greater than 2 s of subbottom penetration are required. In Cascadia basin, reflection and refraction interpretations gave consistent results. The entire sedimentary sequence has low velocity values (≤2.6 km/s) and is about 1.8 km thick. A thin layer (0.4–0.7 km) of basaltic basement with velocity ~5.1 km/s lies below the sediments, and in turn is underlain by a 2 km layer with velocity ~6.1 km/s. A near-vertical incidence profile recorded in this study and a stacked record section provided by an oil company show reflections to subbottom depths of ~4.5 km, corresponding to the top of layer 3. The latter is laterally variable and poorly defined. Reflections from within layer 2 are recorded and some may be related to flows of basalt during crustal formation.