Microcosms designed to simulate the pelagic epilimnion of a lentic freshwater body were compared with the natural system from which they were derived. Emphasis was placed on determining the influence of water agitation on microcosm realism and replicability. In two experiments, excellent tracking of the natural system for 4 to 6 weeks and excellent replication among microcosms was observed; the dominant phytoplankton and zooplankton taxa in the microcosms were not statistically distinguishable from those measured in the natural systems. In a third experiment, the dominant taxa in the 50-L microscosms could be statistically distinguished from the variables in the natural system. In all three experiments, correlations between the taxonomic variables and the chemical variables, when present in the natural system, were observed in the microcosms.