The freshwater microcosm experiments have demonstrated that the stocking of Nile tilapia ( Oreochromis niloticus) or silver carp ( Hypophthalmichthys molitrix) and bighead carp ( Aristichthys nobilis) not only altered the community structure and metabolism intensity of the microcosms but also induced the marked changes in the nutrient levels of the microcosms, and the effects of omnivorous Nile tilapia on the microcosms were far greater than those of planktivorous silver carp and bighead carp. It is generally considered that the major mechanisms by which fish regulate the nutrient recycling in aquatic ecosystems include egestion (voided undigested material), excretion (dissolved metabolic wast material), reducing the mean body size of zooplankton via size-selective predation and thereby enhancing nutrient excretion in zooplankton community, the release of nutrients in fish carcasses by microbial decomposition, and accelerating the release of nutrients in sediments by mechanical mixing processes, in which the importance of nutrient excretion by fishes to the regeneration of nutrients in ecosystems has received the greatest attention. In order to evaluate the role of experimental fish in phosphorus recycling in the microcosms, author estimated orthophosphate (PO4-P) excretion by the experimental fish in the microcosms by measuring the specific PO4-P excretion rates of the three species of fish and fish biomass in the microcosms following the estimation of PO4-P turnover time and PO4-P excretion by zooplankton in the microcosms in the later period of the experiments. The rates of PO4-P excretion by fishes were directly measured in size-classes of the each species of fish. In the test fish taken from the microcosms or culture aquaria were promptly transfer-red to clean containers (1-5 fish per container dependent on fish size) holding tap water prefiltered with active carbon, aerated and stored for 24h at 23-25 degrees C. The tested fish stayed in the containers for 1h, and in the meanwhile feces voided by fishes were immediately drawn out with absorption tubes. The specific PO4-P excretion rate of fish was determined as the increment of PO4-P concentrations in the container in 1h, correcting for the volume of water in the container and the biomass of tested fish, here assuming that a brief stop of feeding by fishes could not bring about the obvious effect on their normal excretion of nutrients. The results of direct measurement of PO4-P excretion by fishes were proved to be reliable by the results of estimation of phosphorus excretion by Nile tilapia using a bioenergetics approach for testing the above assumption. According to the results of direct measurement the specific PO4-P excretion rate of Nile tilapia was much higher than that of silver carp and bighead carp as they were in the same size, which seemed to be independent of phosphor-us content in food but dependent on phosphorus content in fish tissues. And the rate Of PO4-P excretion by the experimental fish in the microcosms stocked with Nile tilapia, on the average, was more than three times that in the microcosms with silver carp and bighead carp, and phosphorus excretion by fishes would supply 10.0%-15.8% and 6.1%-7.5% of the PO4-P requirements of phytoplankton in the microcosms, respectively.It was inferred from the relation between PO4-P supply and demand for PO4-P in the microcosms with fish and the distribution of phosphorusin the microcosms at the end of the experiments that the release of phosphorus in sediments was the dominant pathway for phosphorus regeneration in the microcosms and egestion by the experimental fish was the major mechanism to regulate the recycling of phosphorus in the microcosms.