The effects from a simulated accidental pollution event in a pond with polymeric MDI (4,4'-diphenylmethane diisocyanate and homologs) on different trophic levels of the aquatic ecosystem were investigated in small artificial ponds. Three 4.5-m3 volume ponds, interconnected with closable locks, were provided with natural lake sediment and ground water. Caged fish (rainbow trout, Oncorhynchus mykiss) were added to each pond, and the interconnecting locks were kept open to establish nearly identical physicochemical and biological conditions. At this stage, the ponds were isolated from one another and MDI was added at a dosage of 1 g/liter on top of the sediment of treated part of the first pond, 10 g/liter to the second pond, and 0 g/liter to the third pond (untreated control). Neither the applied monomer MDI nor its potential reaction product MDA (4,4'-diphenylmethanediamine) was detected in water or accumulated by fish. The MDI polymerized to inert polyurea on the sediment of the test ponds. This polymerization formed carbon dioxide, released as bubbles which floated to the water surface. Some carbon dioxide was solubilized in water and reduced the water pH of about 9 by 2.0 units as an average in the high-dosed pond and 0.7 in the low-dosed pond. This reduction caused some other minor changes in the physicochemical characteristics of the pond water. Neither application rate caused any direct effect on the pelagic community (phytoplankton, zooplankton, fish, macrophytes) of the test ponds. Some minor indirect effects caused by the production of carbon dioxide were observed in phyto- and zooplankton community structures. Also, an increase of macrophyte growth was noted. Organisms living in the untreated part of the sediment (macrobenthos) were affected as a result of physical obstructions in this habitat. These populations, however, regained densities equivalent to the control after some weeks, except for Bivalvia which have too long of a generation time for the test period of this study.