Fish polyculture may increase the maximum standing crop of a pond by utilizing a wider range of available food and ecological niches. To identify a better polyculture system for the fish production practiced in southern South America, a 167-d experiment was conducted in 16 earthen ponds (250 m2). The control group contained a traditional combination of carps: 35% common carp (CC), 35% grass carp, 15% silver carp (SC), and 15% bighead carp (BC). The T25%, T50%, and T75% groups consisted of ponds in which CC was replaced by jundiá (JN), and SC and BC were replaced by Nile tilapia at substitution rates of 25%, 50%, and 75%, respectively. The yields of the T25%, T50%, and T75% groups were higher than those of the control group. The effluents generated in the control and T25% groups had lower total phosphorus levels than those generated in the other groups. Considering that all the substituted groups demonstrated better growth performance than the control group and that the T25% group was also better in terms of effluent quality and economic performance, we concluded that fish polyculture in South America would be optimized using a species ratio tested in T25% groups.