Concurrent, rice-fish systems in the Mekong Delta, Vietnam, are characterized by a rice field surrounded by a trench. Rice is direct seeded and fish are reared in polyculture. The most common species for polyculture are silver barb Barbodes gonionotus (Bleeker), common carp Cyprinus carpio L. and Nile tilapia Oreochromis niloticus (L.). In a 1996-97 rice-fish experiment, the effects of (1) absence or presence of such a polyculture, and (2) different rice-seeding rates (100 and 300 kg pre-germinated rice per ha) on the field floodwater ecology were studied. Fish obtained from local nurseries were stocked in six out of 12 rice-fish plots at the Co Do experimental rice-fish station (10 degrees 10' N and 105 degrees 20' E). Silver barb, common carp and Nile tilapia were stocked in polyculture at densities of 3150, 310 and 550 fish ha(-1), respectively. About 50% of the observed variation in the floodwater could be attributed to the fish stocked and to rice-seeding rate. The above-ground dry weight of the rice plants increased in the presence of fish and by an increase in rice-seeding rate. Greater shading by denser plants resulted in changes in variables related to aquatic primary production, including the amount of chlorophyll-a, dissolved oxygen concentration and pH. The presence of fish quadrupled the ammonium concentration, but decreased that of ortho-phosphate. Though fish increased the ammonium concentration, this did not result in a higher aquatic photosynthesis. Light, not ammonium, was the most limiting factor. However, increased ammonium concentrations in the presence of fish did have a positive effect on the density of Euglenophyta and the amount of chlorophyll-a, though only at the lower rice-seeding rate. At the higher rice-seeding rate, light seemed to be lacking. At the lower rice-seeding rate, fish also had a significant effect on the Cladocera-Rotifera ratio. Rotifera numbers increased as fish stimulated phytoplankton. We also demonstrated that fish most probably do not reduce ammonia volatilization losses. The lower rice-seeding rate seemed better for fish production than the higher rate: at the lower rate, dissolved oxygen concentrations were higher throughout the rice cycle. In addition, the amount of chlorophyll-a and the density of Euglenophyta were higher at the lower rice-seeding rate. Finally, we could show that even at high rice-seeding rates, fish had an impact on the aquatic environment. Fish access to the field is not hindered by dense rice stands.