Microcosms containing intact soil-cores are a potential biotechnology risk assessment tool for assessing the ecological effects of genetically engineered microorganisms before they are released to the field; however, microcosms must first be calibrated to ensure that they adequately simulate key field parameters. Soil-core microcosms were compared with the field in terms of ecological response to the introduction of a large inoculum of a rifampicin-resistant rhizobacterium,Pseudomonas sp. RC1. RC1 was inoculated into intact soil-core microcosms incubated in the laboratory at ambient temperature (22°C) and in a growth chamber with temperature fluctuations that mimicked a verage field values, as well as into field lysimeters and plots. The effect of the introduced bacterium on ecosystem structure, including wheat rhizoplane populations of total and fluorescent pseudomonads, total heterotrophic bacteria, and the diversity of total heterotrophic bacteria, was determined. Fluorescent pseudomonads were present on the rhizoplane in significantly lower numbers in soil inoculated with RC1, in both microcosms and the field. Conditions for microbial growth appeared to be most favorable in the growth chamber microcosm, as evidenced by higher populations of heterotrophs and a greater species diversity on the rhizoplane at the three-leaf stage of wheat growth. Ecosystem functional parameters, as determined by soil dehydrogenase activity, plant biomass production, and(15)N-fertilizer uptake by wheat, were different in the four systems. The stimulation of soil dehydrogenase activity by the addition of alfalfa was greater in the microcosms than in the field. In general, growth chamber microcosms, which simulated average field temperatures, were better predictors of field behavior than microcosms incubated continuously at 22°C.