Bacillus cereus UW85 produces a novel aminopolyol antibiotic, zwittermicin A, that contributes to the ability of UW85 to suppress damping-off of alfalfa caused by Phytophthora medicaginis. UW85 produces a second antibiotic, provisionally designated antibiotic B, which also contributes to suppression of damping-off but has not been structurally defined yet and is less potent than zwittermicin A. The purpose of this study was to isolate genetically diverse strains of B. cereus that produce zwittermicin A and suppress disease. We found that most isolates of B. cereus that were sensitive to phage P7 or inhibited the growth of Erwinia herbicola produced zwittermicin A; therefore, phage typing and E. herbicola inhibition provided indirect, but rapid screening tests for identification of zwittermicin A-producing isolates. We used these tests to screen a collection of 4,307 B. cereus and Bacillus thuringiensis isolates obtained from bacterial stock collections and from diverse soils collected in Honduras, Panama, Australia, The Netherlands, and the United States. A subset of the isolates screened by the P7 sensitivity and E. herbicola inhibition tests were assayed directly for production of zwittermicin A, leading to the identification of 57 isolates that produced zwittermicin A; 41 of these isolates also produced antibiotic B. Eight isolates produced antibiotic B but not zwittermicin A. The assay for phage P7 sensitivity was particularly useful because of its simplicity and rapidity and because 22 of the 23 P7-sensitive isolates tested produced zwittermicin A. However, not all zwittermicin A-producing isolates were sensitive to P7, and the more labor-intensive E. herbicola inhibition assay identified a larger proportion of the zwittermicin A producers. Preliminary phenotypic characterization of the zwittermicin A-producing isolates obtained from soil and plant roots revealed a minimum of 31 distinct strains. We tested UW85 and 98 isolates from this