Several cylindrical and noncylindrical sediment trap designs were tested in a recirculating steady-flow flume. The laboratory study was conducted to achieve dynamic- and geometric-similarity to conditions in a specific field environment where traps eventually would be deployed. Relative (to a "standard" trap design) particle collection efficiencies of the traps were quantified in ∼10 cm/sec turbulent flows that were continuously seeded with particles having fall velocities of about 10−2 to 10−1 cm/sec (the upper range of silt-sized quartz sediments). The nature of flow through the trap mouths was qualitatively described using dye.The following trap biases were demonstrated in the study. For unbaffled cylinders, efficiency decreased over a range of increasing trap Reynolds number (Rt) when aspect ratio (H/D) was held constant, and efficiency increased over a range of H/D when Rt was held constant. Baffling cylinders with various H/D but constant Rt, gave mixed results. Any disturbance to flow near the trap mouth or through the trap tended to increase between-replicate variability. For unbaffled, noncylindrical traps, small-mouth, wide-body traps overcollected particles and funnel-type traps tended to undercollect particles, relative to cylinders of the same height and mouth diameter.The biases demonstrated here are for specific parameter combinations and cannot be generalized outside the range of values tested. The results do indicate that significant biased collections are possible by a variety of trap designs and may be flow-regime dependent. Trap-users thus are urged to interpret vertical flux results with caution. Further quantitative studies of trap biases for the ranges of conditions common in field trapping environments and process-oriented studies of physical trapping mechanisms are needed to determine the utility of sediment traps for flux estimates in ocean flows.