Direct and indirect effects of the fungicide carbendazim on ecosystem structure and functioning were studied < or =8 weeks after application (nominal concentrations: 0, 3.3, 33, 100, and 1000 microg/L) to outdoor microcosms in Thailand. Direct effects on macroinvertebrates are discussed in detail in a separate article. The present article presents the effects on other end points and discusses the hypothesized ecologic effect chain. Negative treatment effects on the zooplankton community were only recorded for the highest carbendazim treatment (NOEC(community) = 100 microg/L). The rotifer Keratella tropica, cladocerans (Moina micrura, Ceriodaphnia cornuta, and Diaphanosoma sp.), and cyclopoid copepods were decreased or even eliminated at this treatment level. The decrease in zooplankton and macroinvertebrate abundances was accompanied by an increase in numbers of several tolerant invertebrates, presumably caused by a release from competition and predation. The death of sensitive invertebrates probably also led to an overall decreased grazing pressure because increased levels of chlorophyll-a and bloom of the floating macrophyte Wolffia sp. were noted. The increase in primary producers is discussed to be the probable cause of changes in physicochemical water conditions, eventually resulting in an anoxic water layer during the last 3 weeks of the experiment. This is likely to have resulted in decreased invertebrate abundances noted in that period. Furthermore, the decreased decomposition of Musa (banana) leaves observed 8 weeks after application is considered to be the indirect effect of a decreased microbial activity resulting from these anoxic water conditions, rather than a direct toxic effect of carbendazim.