The isolation of Listeria monocytogenes from food is carried out using a double enrichment. It is believed that the double enrichment can allow the overgrowth of Listeria innocua in samples where both species are present. In this study, we have evaluated the impact of overgrowth between Listeria species and strains during each step of the enrichment process. The effect of factors minimizing interactions between strains or phage inhibitory effects has also been estimated. In an artificially contaminated food undergoing enrichment, overgrowth could result from competitive interactions between Listeria spp. resulting from the production of bacteriocins and bacteriophage at high initial contamination levels (>10(4) cfu/g), but not at lower levels (50-100 cfu/g) as generally found in contaminated foods. At high levels of inoculation, the competitive effect could be reduced by solidification of the selective broths, to limit the diffusion of the inhibitors. Overgrowth resulting from differences in growth rate occurred independent of the initial contamination level. However, in naturally contaminated foods undergoing enrichment, there were no absolute correlations between growth rates or inhibitory profiles in terms of strain evolution during enrichment. In fact, Listeria strains which were predominant in the original sample in most cases remained the dominant strains at the end of the enrichment, although the relative proportion of any given strain could change significantly over the enrichment process. Additional factors which have yet to be identified impact on the evolution of Listeria in the two-step enrichment process. Analysis of strain evolution in eight naturally contaminated foods has indicated that the second enrichment step in Fraser broth can be reduced from 48 to 24 h without impacting on the recovery of L. monocytogenes. Our limited survey of naturally contaminated foods also demonstrated that maximum recovery of L. monocytogenes and other Listeria strains was found following 24 h incubation in 1/2 Fraser Broth. This finding suggests that it may be possible to shorten the current two-step isolation method further without reducing method sensitivity.