Experiments conducted on the Oregon continental shelf in June 1979 indicate that the boundary layer flow at the seafloor was hydrodynamically smooth. Fine-resolution velocity profiles are used to test the assumption that the flow behaved like a universally similar, neutrally buoyant flow over a smooth wall. The non-dimensional thickness of the viscous sublayer is far more variable than has been observed in laboratory studies over perfectly smooth walls. This variability may conceivably be related to upstream changes of surface roughness or to the presence of distributed roughness elements. Laboratory results indicate that the near-bed flow adjusts extremely slowly to upstream roughness changes of surface roughness or to the presence of distributed roughness changes and suggests that flow near the sublayer in the ocean may be influenced by roughness changes as far as tens of meters upstream.