The direct viable count (DVC) method was modified by incorporating radiolabeled substrates in microautoradiographic analyses to assess bacterial survival in controlled laboratory microcosms. The DVC method, which permits enumeration of culturable and nonculturable cells, discriminates those cells that are responsive to added nutrients but in which division is inhibited by the addition of nalidixic acid. The resulting elongated cells represent all viable cells; this includes those that are culturable on routine media and those that are not. Escherichia coli and Salmonella enteritidis were employed in the microcosm studies, and radiolabeled substrates included [methyl-3H]thymidine or [U-14C]glutamic acid. Samples taken at selected intervals during the survival experiments were examined by epifluorescence microscopy to enumerate cells by the DVC and acridine orange direct count methods, as well as by culture methods. Good correlation was obtained for cell-associated metabolic activity, measured by microautoradiography and substrate responsiveness (by the DVC method) at various stages of survival. Of the cells responsive to nutrients by the DVC method, ca. 90% were metabolically active by the microautoradiographic method. No significant difference was observed between DVC enumerations with or without added radiolabeled substrate.