The possibility that aquatic and terrestrial insects, killed by operational spray doses of fenitrothion, may cause lethal or sublethal effects directly in salmonids, has been shown by these experiments to be unlikely. The highest level found in poisoned insects by Penney (personal communication) is 3,000 times less than the level at which behavioural effects were noted in Experiment 2. It remains possible that reductions in insect biomass, following spraying, may be of sufficient magnitude and persistence to significantly reduce ration and hence reduce salmonid production. This possibility can only be assessed by direct field observations.