Abstract: Abstract: There is a global boom in dams. Although dams bring a huge economic benefit, they obstruct fish migrations and gene exchange. Fish passages are being designed and constructed to mitigate the influence of the dams on fish migration and gene exchange. Fish behavior researches are useful to the design of fish passages. A large amount of research about critical swimming speed and burst speed, which were related to velocity design in the high velocity region of fish passages such as the entrance, was reported worldwide. However the reports about velocity design in the low velocity region of fish passages were limited. To study fish behavior and provide information for the velocity design in the low velocity region, induced flow speed of 5 fishes in the Lancang River was investigated with the fish behavior measurement equipment and a stepped velocity test was carried out in spring when fish migrated. Before testing, fish body length was measured and the test fish was allowed to adapt to experimental conditions at 0.5 bl/s (bl meant body length of fish) for 2 h. At the initial flow velocity (approximately 0.06 m/s), the fish was nearly motionless along the flow direction. The water velocity was increased by approximately 0.01 m/s at 5 s interval and, when the fish was no longer able to hold station and began to swim actively, the flow velocity was reported as induced flow speed. Results included: 1) The induced flow speed of the fish ranged from 0.40 to 1.18 bl/s (or from 0.07 to 0.23 m/s) at 11.1-22.6 ℃, which was similar to other references (0.4-1.5 bl/s or 0.05-0.15 m/s). 2) Lateral line receptors detected small changes in flow, and swimming was induced in fish when the stimulus from lateral line receptors was sufficient. The relatively induced flow speed (bl/s) was negatively correlated with fish body length and the absolutely induced flow speed (m/s) was positively correlated with fish body length, whether each species was considered individually or all species were considered together. The above relationships were consistent to both the relationship between critical swimming speed and body length, and the relationship between burst speed and body length. The relationship could be used to forecast swimming performance of fishes that were captured with difficulty or had small quantity. 3) Overmuch decreasing or increasing the designed velocity of fish passages would increase the difficulties for the engineering design and increase the cost for construction. The relationship between percentage of fish swimming trend and flow speed was obtained in this study, which could provide the reference to balance the fish demand and flow design. Based on the data of fish induced flow speed, the designed minimum velocity of the fish passages should be more than 0.2 m/s in order that 95% of fish could pass the passages successfully, if both the difficulties for the engineering design and the cost of construction were neglected, and if the effectiveness of fish passages was considered uniquely. The release site for the fish lifting machine and fish transport boats was set in the upstream, far away the dam site. If the velocity at the release site was too low, the fish lost orientation even went back to the downstream, resulting in the fail of fish passages. Therefore, the site where the velocity is higher than fish induced flow speed should be selected for releasing fish. The principle can be also applied to the selection of exit site in fish passages.