A number of studies have revealed ammonia to be toxic to aquatic organisms; however, little is known about its effects under natural conditions. To elucidate the role of ammonia, we conducted 96-h acute toxicity tests as well as a whole-ecosystem chronic toxicity test for one year in ten 600-m(2) ponds. Three common cyprinids, silver carp Hypophthalmichthys molitrix Val. (H.m.), bighead carp Aristichthys nobilis Richardson (Am.), and gibel carp Carassius auratus gibelio Bloch (C.g.), were used as test organisms. The 96-h LC50 values of un-ionized ammonia (NH3) for H.m.,A.n., and Cg. were 035, 033, and 0.73 mg L-1, respectively. In the ponds, annual mean NH3 ranged between 0.01 and 0.54 mg L-1, with 4 ponds having a NH3 higher than the LC50 of A.n. (lowest LC50 in this study). No fish were found dead in the high-nitrogen ponds, but marked histological changes were found in livers and gills. Despite these changes, the specific growth rate of H.m. and An. increased significantly with NH3. Our pond results suggest that fish might be more tolerant to high ammonia concentrations in natural aquatic ecosystems than under laboratory conditions. Our finding from field experiments thus suggests that the existing regulatory limits for reactive nitrogen (NH3) established from lab toxicity tests might be somewhat too high at the ecosystem conditions. Field-scale chronic toxicity tests covering full life histories of fish and other aquatic organisms are therefore encouraged in order to optimize determination of the effects of ammonia in natural environments. (C) 2017 Elsevier B.V. All rights reserved.