The kinetics of response curves from Biolog community profiles for heterotrophic bacteria from a river in Nova Scotia, Canada have been analyzed to generate lag, slope, and asymptote parameters. The river water samples were treated with one of three supplements of Escherichia coli (in situ levels, 10(3) CFU/mL, or 10(6) CFU/mL) and one of five concentrations of chlorine (0, 1, 3, 5, or 7 ppm) to satisfy a full factorial design. The chlorine treatments decreased the inoculum density by up to 2 log values and decreased the species evenness. The E. coli supplements increased the inoculum density and decreased the species richness. Examination of the asymptotes did not reveal any significant effects owing to E. coli, but differences owing to the chlorine were detected. Analyses of the slopes showed a similar insignificance of the effects of E. coli and a lack of treatment effect owing to chlorine. The lag analyses also showed no significant E. coli effects, but showed a significant effect owing to chlorine. The discrepancy produced with the slope analysis (i.e., no chlorine effect) may represent an anomaly of the Biolog community approach. The use of lag phase was impaired because of the problem of infinite lags from wells that had no response, but a principle component analysis with a reduced set of substrates did suggest some influence of E. coli on the community profile. An examination of the substrates metabolized by the river water compared with pure E. coli revealed that the Biolog profiles of the river communities were not a simple summation of the component parts. In light of the lack of uniformity between these analyses, where the outcome depended on which parameter was used, caution is advised in interpreting Biolog community profiles on the basis of only one parameter.