In a laboratory flume, a comparative study on the near-bottom performance of the Acoustic Doppler Velocimeter (ADV) was conducted. Two different ADV systems were tested for different configurations and two flow velocities (9 cm s-1, 18 cm s-1). The results were compared with synchronous measurements with a Laser Doppler Anemometer (LDA). Near-bottom velocity measurements with the ADV have to be interpreted carefully as the ADV technique underestimates flow velocities in a zone close to the sediment. The height of this zone above the sediment varies with different ADV systems and configurations. The values for nominal sampling volume height (SVH) given by the software often underestimate the true, effective sampling volume heights. Smaller nominal SVH improve the ADV near-bottom performance, but the vertical extent of the zone in which the ADV underestimates flow by more than 20% may be larger than true SVH/2 by a factor of 2 (= true SVH). When the measurement volume approaches the bottom, ADV data quality parameters (signal-to-noise-ratio (SNR) and signal amplitude) exceeding the average 'open water' level, are clear indicators that the ADV has begun to underestimate the flow velocity. Unfortunately, this is not a safe indicator for the range of reliable measurements as the ADV may begin to underestimate velocities even with unchanged 'open water' data quality parameters. Thus, one can only recommend avoiding measurements below a distance from the bottom that was defined empirically comparing the ADV and the LDA velocity profiles. This distance is 2.5 times nominal sampling volume height for the tested ADV systems and experimental settings. © Springer 2006.