Hydraulic characteristics inside two research circular tanks (1.5-m and 1.2-m diameter) with the same volume of water were studied to understand how they might affect experimental bias by influencing the behavior and development of juvenile fish. Water velocities inside each tank were documented extensively and flow behavior studied. Surface inflow to the 1.5-m tank created a highly turbulent and aerated surface, and produced unevenly distributed velocities within the tank. A low-flow velocity, or "dead" zone, persisted just upstream of the surface inflow. A single submerged nozzle in the 1.2-m tank created uniform flow and did not cause undue turbulence or introduce air. Flow behavior in the 1.5-m tank is believed to have negatively affected the feeding behavior and physiological development of a group of juvenile fall chinook salmon, Oncorhynchus tshawytscha. A new inflow nozzle design provided comparable flow behavior regardless of tank size and water depth. Maintaining similar hydraulic conditions inside tanks used for various biological purposes, including fish research, would minimize experimental bias caused by differences in flow behavior. Other sources of experimental bias are discussed and recommendations given for reporting and control of experimental conditions in fishery research tank experiments.