Aims: The effect of a lactic acid producing bacterial (LAB) inoculant on the elimination of Escherichia coli O157:H7 from barley forage was assessed. Methods and Results: Triplicate mini-silos were prepared for four treatments and six sampling times (1, 3, 7, 15, 30 and 42 d post-ensiling). The treatments were (i) 10(5) cfu g(-1) Pediococcus pentosaceus and Propionibacterium jenzenii (P2); (ii) 10(5) cfu g(-1) E. coli O157:H7 strain 3081 and 10(5) cfu g(-1) E. coli Biotype 1 strains 719IE10, 719IE14 and 614ME49 (EC); (iii) P2 + EC; and (iv) the control (sterile distilled water). Triplicate mini-silos were opened at each sampling time for pH, volatile fatty acid (VFA) and lactate determinations and E. coli , E. coli O157:H7 and LAB were enumerated. On d 3 and 7, numbers of E. coli O157:H7 in P2 + EC were significantly lower than in EC (P < 0.05). Escherichia coli O157:H7 was not detected in P2 + EC and EC at 7 and 15 d post-ensiling, respectively. On d 15 through 42, E. coli Biotype 1 was not detected in P2 + EC or EC. Populations of LAB were higher in P2 and P2 + EC than in the control and EC on d 3 and 7 (P < 0.05). After 3 d of ensiling, lactate levels were higher (P < 0.05) and pH was lower (P < 0.05) in P2 and P2 + EC as compared to the control and EC. Bacteriocins of P2 were not found to be inhibitory to E. coli O157:H7 using the agar-spot procedure. Escherichia coli O157:H7 inoculated into the control silage at a level of 10(3) cfu g(-1) and exposed to aerobic conditions at 22degreesC was not detected after 1 d and remained undetectable for the 28 d exposure period. Conclusions: Silage inoculant P2 increased lactate levels and decreased pH more rapidly during ensiling, which appeared to hasten the elimination of E. coli O157:H7 from the silage. Significance and Impact of the Study: Results emphasize the importance of adequate ensiling since E. coli O157:H7 may be maintained and spread among cattle through feed.