The effect of interference competition can be assessed by comparing the capture rate of a predator foraging alone with that of the predator within a group. Since such an effect could be prey density dependent, a constant density of prey must be maintained while assessing this effect, irrespective of the elimination of prey by predation. However, when studying a predator-harvester, such as a planktivorous fish, which collects zooplankton at a rate of up to 1 prey s -1 , instantaneous replacement of each consumed prey item is not feasible. This problem was solved in short-lasting mesocosm experiments by minute-byminute supplementation to replace eliminated Daphnia and maintain a constant average prey density. Such experiments were performed with different numbers of foraging roach (Rutilus rutilus) at three prey densities and in two ranges of ambient temperature. The number of Daphnia required at the start of each experiment to establish the initial prey density and the number that it was necessary to add per minute were determined in experiments conducted without prey supplementation and in preliminary experiments with prey supplementation. The results of this study revealed that fish foraging in a group eat less, due to both exploitation and non-aggressive competition for space. Moreover, the effect of interference competition was stronger at higher temperatures, irrespective of the prey density, indicating that natural populations of roach foraging in shoals may suffer more from competitive interactions in warmer waters.