SUMMARY.Three experiments are described in which the rates of loss from suspension ofLycopodiumspores introduced at the water surface in a large butyl rubber enclosure (‘Lund Tube’) were monitored. Rates of arrival at the mud surface and in sediment traps placed in the water column were also measured. Results obtained by each method are compared with the predetermined intrinsic sinking behaviour of the individual particles. The experiments, which were carried out under differing conditions of water‐column stability (isothermal mixing, thermal microstratification, and autumnal destratification), demonstrate the importance of turbulent mixing in determining the effective settling rate of seston. A simple predictive model is developed to relate sinkinglosses to the vertical extent of broadly defined water layers. Recoveries in sediment traps can give an accurate synopsis of net sedimentation when they are operated in non‐turbulent water layers, but trap performance in turbulent flow is significantly modified by hydrodynamic perturbation generated by the traps the