Knowledge of the collection efficiency of sediment traps, particularly under conditions of varying current speed, is presently more a matter of hope than confidence. We report here on a field experiment designed to determine, for a particular trap geometry, the effect of current speed and particle fall velocity on the collection efficiency of a moored trap relative to the presumably unbiased efficiency of an identical drifting trap. The experiment was performed in a deep estuarine tidal passage where a smoothly varying unidirectional flow and a spatially homogenous particle population mimicked laboratory flume conditions. A multiple-sample sediment trap integrated to a current meter partitioned the mass flux collected by the moored trap into one of four chambers according to the following speed intervals: <12, 12–<30, 30–<50, and ≥50cm/s. The magnitude and particle characteristics of the flux collected at <12 cm/s were indistinguishable from those simultaneously collected by drifting traps. At higher speeds, the relative efficiency of the moored trap ranged between 1% and 24% and the mean size and density of the trapped particles increased. These results support predictions based on laboratory studies that collection efficiency decreases with an increase in the trap Reynolds number or a decrease in particle fall velocity. The study demonstrates that consideration must be given to scaling both trap diameter and aspect ratio according to the expected flow conditions, and that knowledge of flow conditions at the trap mouth is necessary to properly interpret the flux data.