Prey catching in the archer fish: does the fish use a learned correction for refraction?

An answer to the question of how the archer fish hits an aerial insect, despite the refraction of light at the surface of the water has not yet been found. The aims of the present studies are to find out: (1) whether the fish applies a learned correction with the virtual image as a point of reference; (2) whether deprivation of practice in squirting affects performance. For the first aim the accuracy of squirts was measured in 30 subjects. Contrary to suggestions from the literature, elevation failures were prominent but the frequencies of over- and under-squirting did not differ, which does not support the idea that the fishes applied a learned correction for refraction by using feedback from the efficacy of squirts. For the second aim, five experimental subjects were deprived of practice, whereas six control subjects got daily practice, during 6 months. The only significant difference, found thereafter, was that during the first session experimental subjects aimed more often before squirting than control subjects did, but hitting was not affected. A number of subjects developed abnormal mandibles which inevitably led to squirting too high. Our findings do not support the hypothesis that the archer fish uses learned corrections for refraction.