Native fish populations in the Murray-Darling Basin are estimated at 10 % of pre-European settlement levels. Whilst multiple key threatening processes have contributed to this decline, the impact of fish losses at water diversions has been largely underestimated and remains unaddressed. There is mounting evidence that significant numbers of fish (ranging from hundreds to millions at individual diversions) are being lost from rivers through water abstraction. Whilst the true extent of the impact across the entire Basin remains uncertain, given the extent of water diversion from most rivers, it is likely that fish entrainment is significant and will need to be addressed if a recovery in native fish populations is to be achieved.Fish screens can be used to protect fish populations whilst maintaining irrigator entitlements. Although several different screening approaches are currently applied elsewhere in the world, most of which would be suitable for application in the Murray-Darling Basin, it is essential that technologies are designed with the needs of local fish species in mind. In particular, screens will need to meet certain design criteria (e.g. maximum velocities at the screen face, or be made out of suitable material) that ensure fish are excluded from abstracted water and do not suffer injury or mortality. Currently no such criteria or guidelines exist for the design of screens suitable for Australian native fish, and this study has been the first to collect data relevant to the Murray-Darling Basin.A combination of field and laboratory-based experiments at simulated intake screens was used to test a variety of approach velocities (velocities in front of and perpendicular to the screen face) and screening materials. It was found that the installation of fish screens has great potential to significantly reduce fish entrainment at intakes and, in some cases, mortality at an experimental intake was reduced from over 90 % (unscreened) to less than 2 % (when screened) in the laboratory. Approach velocities (measured 8 cm from the screen) of up to 0.4 m/sec (1.5 m/sec slot velocity through the screen) were effective in reducing entrainment of juvenile golden perch and silver perch in laboratory trials, with very little injury or mortality resulting from incidental screen contacts or impingement. In comparison, field observations of an assemblage of fish at a screen in a river demonstrated that even modest increases in approach velocity (from 0.1 to 0.5 m/sec) produced a significant increase in the rate of screen contact for fish smaller than 150 mm, with the impact being more marked the smaller fish were.