Stocking of bigheaded carp (mainly Hypophthalmichthys nobilis and H. molitrix) is commonly used in (sub)tropical Chinese reservoirs to control phytoplankton, but with ambiguous results. Whether these carp act as a phosphorus (P) source or sink for phytoplankton is debated. We compared the trophic structures in twenty-three reservoirs with different nutrient concentrations in the flood season (after bigheaded carp introduction) with the dry season (after bigheaded carp harvesting). Fish biomass was positively related to TP, and the slope of the relationship showed no difference between seasons. Bigheaded carp harvesting exceeded the amount introduced, which may explain an observed lower intercept of the relationship and fish biomass to the TP ratio in the dry season. Fish predation pressure on zooplankton (fish: zooplankton biomass ratio as a proxy) was highest in the flood season and increased with TP in both seasons. Accordingly, zooplankton grazing effect on phytoplankton (zooplankton: phytoplankton biomass ratio as a proxy) decreased with fish biomass. Furthermore, both the zooplankton biomass and the zooplankton: phytoplankton biomass ratio were among the lowest reported in the literature for the nutrient range studied. Fish grazing effect on phytoplankton (fish: phytoplankton biomass ratio as a proxy) was also highest in the flood season and decreased with TP in both seasons. Nanoplankton was the dominant phytoplankton group in oligotrophic to mesotrophic reservoirs, while filamentous cyanobacteria dominated in eutrophic reservoirs. Chlorophyll a increased with TP and fish biomass, whereas the yield of chlorophyll a per TP (Chla: TP ratio) increased with fish biomass. Accordingly, both chlorophyll a and the Chla: TP ratio were highest in the flood season. We conclude that bigheaded carp act as P sink at the ecosystem level but as P source for phytoplankton, and enhance the yield of chlorophyll a per TP and thus eutrophication. (C) 2020 Elsevier Ltd. All rights reserved.