To study the effects of grazing on a natural post-spring phytoplankton community a land-based mesocosm experiment was performed during late 1988 (7–18 June). 300-1 polyethylene cylinders, immersed in a flexible pool for temperature regulation, were filled with a plankton community filtered through a 90-μm nylon net. In duplicate cylinders, the ambient copepod concentration (mainly an Acartia sp.), five times the ambient concentration, ambient concentration + medusae of the scyphozoan Aurelia aurita (L.) and ambient concentration + mussels (Mytilus edulis L.) were added. Nutrients were added throughout the experimental period to keep the concentrations at 5 μM nitrate, 0.5 μM phosphate and 10 μM silica. Chl a, nutrients and densities of phytoplankton, ciliates and copepods were measured throughout the experiment. We observed that the diatoms remaining from the spring bloom could, with small additions of nutrients, outgrow their grazers. Only a small reduction in copepod grazing pressure through medusae predation on copepods was required for an even stronger growth and dominance by diatoms. This pattern of development could be broken by an increase in Acartia sp. biomass. An initial addition of five times the natural concentration of Acartia sp. led to a rapid growth of monads, < 3 μm, which constituted almost 100% of the biomass after 3 days. When Acartia sp. decreased in biomass, diatoms and cryptomonads became dominant. As the small naked ciliate Urotricha sp. had already been eliminated by the initially abundant copepods in the cylinders where the copepod concentration was five times the ambient concentration, there were few grazers left to make use of the ample food supply. Urotricha sp. could possibly control the < 3-μm monads if not or only moderately predated by Acartia sp., indicated by the lower increase in monad biomass in the cylinders where mussels or medusae were added. M. edulis was capable of grazing down all phytoplankton species to very low densities, except monads < 3 μm, in spite of nutrient additions, indicating a possible nonspecies-selective control of phytoplankton abudance by benthic filter-feeders in eutrophicated areas.