We tested the performance of two stationary interrogation systems designed for detecting the movement of fish with passive integrated transponder (PIT) tags. These systems allowed us to determine the direction of fish movement with high detection efficiency and high precision in a dynamic stream environment. We describe an indirect method for deriving an estimate for detection efficiency and the associated variance that does not rely on a known number of fish passing the system. By using six antennas arranged in a longitudinal series of three arrays, we attained detection efficiencies for downstream- and upstream-moving fish exceeding 96% during high-flow periods and approached 100% during low-flow periods for the two interrogation systems we tested. Because these systems did not rely on structural components, such as bridges or culverts, they were readily adaptable to remote, natural stream sites. Because of built-in redundancy, these systems were able to perform even with a loss of one or more antennas owing to dislodgement or electrical failure. However, the reduction in redundancy resulted in decreased efficiency and precision and the potential loss of ability to determine the direction of fish movement. What we learned about these systems should be applicable to a wide variety of other antenna configurations and to other types of PIT tags and transceivers.