The application of preexisting DNA hybridization techniques was investigated for potential in determining populations of specific gene sequences in environmental samples. Cross-hybridizations among two degradative plasmids, TOL and NAH, and two cloning vehicles, pLAFR1 and RSF1010, were determined. The detection limits for the TOL plasmid against a nonhomologous plasmid-bearing bacterial background was ascertained. The colony hybridization technique allowed detection of one colony containing TOL plasmid among 10(6) Escherichia coli colonies of nonhomologous DNA. Comparisons between population estimates derived from growth on selective substrates and from hybridizations were examined. Findings indicated that standard sole carbon source enumeration procedures for degradative populations lead to overestimations due to nonspecific growth of other bacteria on the microcontaminant carbon sources present in the media. Population estimates based on the selective growth of a microcosm population on two aromatic substrates (toluene and naphthalene) and estimates derived from DNA-DNA colony hybridizations, using the TOL or NAH plasmid as a probe, corresponded with estimates of substrate mineralization rates and past exposure to environmental contaminants. The applications of such techniques are hoped to eventually allow enumeration of any specific gene sequences in the environment, including both anabolic and catabolic genes. In addition, this procedure should prove useful in monitoring recombinant DNA clones released into environmental situations.