Experiments were performed using several modifications of a micro-ecosystem made up of three subsystems connected by a unidirectional flow of recirculating water. The three subsystems represent roughly the trophic levels of autot>rophs (algae), herbivores (Daphnia magna), and decomposers (sandfilter with bacteria). The particle concentration was determined weekly, using a Coulter Counter, and the Daphnia population was counted. A t regular intervals, the species composit.ion of the algae was determined. The changes in t.he algal concentration and Dapimici popuhtion are described, and the different modifications are discussed.I n several micro-ecosystems, the herbicide dichlobenil was introduced once or several tiniest and in one micro-ecosystem, the herbicide diuron was added. The response of the micro-ecosystems is discussed. For the periods without a toxic substance present, a normal operating range is suggested in relation to which the response of the systems to the introduced herbicide is discussed.Besides the separate analysis of the data, the algal concentrations and the Daphnia numbers were correlated to each other. This was done by regression analysis and a,ccording t o a dynamic system approach. I n both cases, the effect of herbicides was better discerned than in the separate analysis of the data. I n the dynamic systems approach, a normalized ecosystem strain was defined a s a parameter to measure effects of stress on the ecosystem level.