Possible causes of the increased algal blooms in Lake Victoria in the 1980s have been disputed by several authors; some suggested a topdown effect by the introduced Nile perch, whereas others suggested a bottom-up effect due to eutrophication. In this article the potential impact is established of grazing by fish on phytoplankton densities, before the Nile perch upsurge and the concomitant algal blooms in the Mwanza Gulf. The biomass and trophic composition of fish in the sublittoral area of the Mwanza Gulf were calculated based on catch data from bottom trawls, and from gill nets covering the whole water column. Estimates of phytoplankton production in the same area were made from Secchi values and chlorophyll concentrations. The total phytoplankton intake by fish was estimated at 230 mg DW m-2 day-1. The daily gross production ranged from 6,200 to 7,100 mg DW m-2 day-1 and the net production from 1,900 to 2,200 mg DW m-2 day-1. Thus, losses of phytoplankton through grazing by fish were about 3–4% of daily gross and 10–12% of daily net phytoplankton production. As a consequence it is unlikely that the phytoplankton blooms in the second half of the 1980s were due to a top-down effect caused by a strong decline in phytoplankton grazing by fish.