The use of a minimal lactose enrichment broth (MLB) in conjunction with immunomagnetic electrochemiluminescence detection (IM-ECL) was evaluated for the estimation of viable Escherichia coli O157 populations in surface water samples. In principle, E. coli O157 populations (C(initial E. coli O157)) can be derived from enrichment data according to the equation: C(initial E. coli O157) = C(initial coliforms) x C(final E. coli O157)/C(final coliforms)), assuming that the growth rates and lag times of water-borne E. coli O157 and collective coliforms are sufficiently comparable, or at least consistent. We have previously described a protocol for determining C(final E. coli O157) in MLB-enriched water samples. In the present study, 80% of coliforms (red/pink colonies on MacConkey Agar) grew in MLB, indicating that this provides reasonably accurate estimates of C(initial coliforms). Estimates of C(final coliforms) were determined from turbidity data. Initial E. coli O157 populations (C(initial E. coli O157)) were calculated for 33 Baltimore watershed samples giving a positive IM-ECL response. The majority of samples contained E. coli O157 concentrations of < 1 cell per 100 ml. These data indicate that E. coli O157 are present in surface water samples but at very low levels. Growth rates for MLB-enriched coliforms were highly variable (k= 0.47 +/- 0.13 h(-1), n= 72). There was no correlation between growth rates and any measured water parameter, suggesting that coliform populations in water samples are spatially and temporally unique. Although variability in growth rates was expected to yield some low values, the fact that most E. coli O157 concentrations were < 1 suggests that other factor(s) were also responsible. Studies with E. coli O157:H7 and wild-type E. coli suggest that increased lag times due to starvation were at least partially responsible for the observed data. Based on estimates of C(initial coliforms) and k(coliforms), MLB was evaluated for sensitivity and quantitativeness. Simulated populations of E. coli O157:H7 at stationary phase varied from ca. 10(3) to 10(8) cells ml(-1) enrichment culture. Although not suitable for quantitation, MLB enrichment in conjunction with IM-ECL can detect as few as one viable water-borne E. coli O157 cell per 100 ml surface water. Experiments are in progress to evaluate alternative media for sensitivity and quantitative detection of enterohemorrhagic E. coli.