A large-scale biomanipulation trial was carried out on Lake Vesijärvi in Finland during 1989–1993. Following the mass removal of coarse fish the biomass of cyanobacteria collapsed from 1.4 g/m−3 to below 0.4 g/m−3, while total phosphorus concentration declined from 45 μg/L to 30 μg/L. No relevant changes in zooplankton communities were observed. The results suggest that the success of food web manipulation as a tool for lake restoration is not necessarily dependent on the grazing rate of zooplankton. The effects of reduced fish-mediated internal loading and recycling of nutrients are in many cases stronger than those of reduced planktivory. Alternative stable states of water quality may also exist in lakes not covered by macrophytes, owing to the changes in the behavior of fish stocks. Year-to-year variation in the littoral zone may cause large oscillations in lake ecosystems—for example, through the recruitment of fish. In addition, the nutrients translocated by fish from the littoral zone may affect the nutrient dynamics of the pelagial plankton community. In terms of phytoplankton species composition and the ratio of phosphorus to chlorophyll a, the water quality in Lake Vesijärvi has improved in a stepwise fashion within the last 10 years. This is probably due to the fact that the five-year mass removal of fish in Enonselkä fulfilled the requirement of sustained management of fish stocks in order to maintain nonequilibrial conditions between alternate stable states. The prediction of the water quality development is obscured, however, by spatial and temporal within-lake variation, which sets high requirements for sampling programs.