An important consideration in the environmental release of a genetically engineered microorganism is the capability for reduction or elimination of microorganism populations once their function is completed or if adverse environmental effects are observed. In this study the decontamination treatments of burning and biocide application, alone and in combination with tilling, were evaluated for their ability to reduce populations of bacteria released on the phylloplane. Field plots of bush beans (Phaseolus vulgaris), sprayed with the bacterium Erwinia herbicola, received the following treatments: control; control + till; burn; burn + till; Kocide (cupric hydroxide); Kocide + till; Agri-Strep (streptomycin sulfate); and Agri-Strep + till. Leaves and soil from the plots were sampled −1, 1, 5, 8, 12, 15, 19, and 27 days after application of the decontamination treatments. Burning produced a significant reduction in the number of E. herbicola, whereas tilling, alone or in combination with the biocide treatments, stimulated a significant increase in E. herbicola populations, which persisted for several weeks. The individual treatments of the biocides, Kocide and Agri-Strep, produced a rate of decline in E. herbicola populations that did not significantly differ from that of the control treatment. Key words: decontamination, risk control, field release, genetically engineered microorganisms.