The purpose of this research was to examine the relationship between the duration of source exposure and the colonization of islands. Experiments used colonization of polyurethane foam (PF) artificial substrates by Protozoa as model systems. Species sources were obtained by colonizing PF substrates in natural systems for varying periods. Sources were placed in sterile laboratory microecosystems that contained barren PF islands. Short-term (7 d) experiments revealed no differences in colonization of islands exposed to sources ranging in age from 3 h to over 1000 d or isalnds exposed to sources either briefly or continuously. This suggested that propagule production was extremely rapid from all sources and confirmed the stochastic nature of early, non-interactive colonization. Further experiments using larger microecosystems with more islands and maintained for up to 28 d in the laboratory examined colonization resulting from differences in source maturity and complexity. Colonization was directly related to the number of species on the source and was reduced when the source was removed after a brief (24 h) period, although differences in colonization were small for experiments carried out over 28 d. All experimental systems reached species equilibrium suggesting the possibility of multiple stable states for these communities. Varying the period of source exposure demonstrated that the eventual equilibrium species number on islands was directly related to the natural logarithm of source exposure time. Colonization from short-term source exposures converged rapidly to that resulting from continuous source exposures. This suggested that a critical period of source exposure (much less than continuous) may also lead to maximal colonization. The ability to maintain complex communities in the laboratory for extended periods was used to develop a multispecies toxicity test based on protozoan colonization. Microecosystems identical to those used in theoretical experiments were amended with 0, 10, and 500 ug/l cadmium, and experiments were conducted over 21 d using mature species sources. Colonization of islands was reduced by over one-half (relative to controls) at 10 ug/l Cd and was nearly eliminated at 500 ug/l Cd. Sources were not affected at low concentrations but were severely altered at 500 ug/l Cd. These tests are attractive because they allow the use of indigenous communities in testing a complex biological response not obtainable by studying single species.