The addition of penicillin G to combat microbial contamination in continuous fuel alcohol fermentations was performed using both continuous and pulsed addition regimes. In continuous fermentations where both Saccharomyces cerevisiae and Lactobacillus paracasei were present, the mode of addition of penicillin G determined final numbers of viable L. paracasei. When the same overall average concentration of penicillin G was added in both pulsed and continuous modes, the initial viable number of L. paracasei (8.0x10(9) cfu ml(-1)) decreased to a greater degree (1.02x10(5) cfu ml(-1) L. paracasei) when penicillin G was pulsed at 6 h frequencies at an overall average concentration of 2,475 U/l than when penicillin G was added continuously at 2,475 U/l (2.77x10(5) cfu ml(-1) L. paracasei). Pulsed additions over longer frequencies at 2,475 U/l were not as effective in reducing viable bacteria. Viable yeasts increased during both treatment conditions by more than 2-fold. The two addition regimes also eliminated the 40% decrease in ethanol concentration caused by the intentional bacterial infection. Although there was 3 times more bacterial death with 6 h pulsed additions compared to continuous additions of penicillin G at 2,475 U/l, there was, by that point, no practical difference in either final ethanol concentration or relative ethanol recovery.