Large water-use facilities are often equipped with vertically travelling debris barriers known as Ristroph screens. Although made in a variety of configurations, all such screens are equipped with some manner of fish-catching troughs or rails, and all operate on the principle of direct contact and active removal of impounded fish and debris. The imposed fish mortalities associated with these machines are commonly attributed to the consequences of impingements (to fish being flattened against the screening by the force of the inflowing water), but the laboratory and field experiments reported here imply that in those circumstances where the screens travel continuously and where water speeds are moderate, the major underwater injuries are attributable instead to buffeting of captured fish within the fish troughs proper. A reconfigured machine, including the redesigned fish-catching apparatus, was installed and tested at a nuclear generating station on the Hudson River estuary. In tests similar to those on the unimproved machine, injuries and deaths were reduced from 53 to 9% for striped bass Morone saxatilis , from 64 to 14% for white perch M. american , from 80 to 17% for Atlantic tomcod Microgadus tomcod , and from 47 to 7% for pumpkinseed Lepomis gibbosus .