Lights were attached to the headline of a bottom trawl during a series of tows at 1000 m depth in the Rockall Trough (north-east Atlantic). There was no significant difference in the overall numerical abundance of demersal fish between the trawl catches with light and the controls with no light. A principal components analysis based on the total fish catch between the experimental and the control stations showed no segregation. At the individual species level significant differences in catch rate were found for eight species. These are discussed in relation to the small amount of available information on the visual systems of deep-sea demersal fish and their diet.