A novel method for selecting overproducing strains of rifamycin B was developed. This technique involves the use of lysozyme and the effect of barbital on the growth of A. mediterranei. Complete medium added with glycine and barbital was inoculated with mutagenized mycelium, incubated for 48 hours and treated with lysozyme. The lysozyme resistant mycelium was washed with dilute detergent. Complete medium with glycine but without barbital was inoculated with the washed mycelium. Protoplasts were obtained and regenerated and the colonies were picked and seeded on Bennet agar plates with and without barbital. Two selected mutants were sensitive to 0.5% barbital producing 200% more rifamycin than the parental strain. In addition, 30 barbital resistant mutants were isolated and their production level was lower than the one observed with the parental strain. These results suggest that the effect of barbital on secondary metabolism (rifamycin production) is related to its effect on primary metabolism.