A major obstacle in defining the mechanism of chemical-induced toxicity has been the inability to distinguish between events that cause cell death and those that result from cell death. This problem results from measuring biochemical parameters in tissues or cell pellets containing both viable and nonviable cells. In the present study, we described a method for the rapid separation of viable hepatocytes from nonviable cells and medium prior to biochemical analysis. Separation of viable hepatocytes was accomplished in a microcentrifuge tube by layering a sample of isolated hepatocyte suspension over a dibutyl phthalate oil layer and centrifuging for several seconds. As a result, greater than 90% of the hepatocytes centrifuged through dibutyl phthalate were viable while greater than 90% of the cells recovered above the oil layer were nonviable. The separation of viable hepatocytes by the dibutyl phthalate method was not affected by the presence of the hepatotoxins, adriamycin (ADR) in combination with 1,3-bis(2-chloroethyl)-1-nitrosourea (BCNU) or ethyl methanesulfonate (EMS), though the ratio of viable to nonviable cells in the suspension was drastically reduced. The metabolic and morphological integrity of hepatocytes centrifuged through dibutyl phthalate was altered after cell suspensions were treated with the ADR-BCNU or EMS. These chemically treated viable hepatocytes showed degenerative ultrastructural changes and a greater than 80% reduction in intracellular K+ and glutathione concentrations. Because centrifugation through dibutyl phthalate does not significantly alter the concentration of intracellular constituents nor the ultrastructure of control hepatocytes, the signs of reversible injury observed in hepatocytes centrifuged through oil resulted from the chemical treatment. These data indicate that the dibutyl phthalate separation technique offers the advantage of monitoring only viable hepatocytes for changes in membrane integrity or metabolic performance during a toxic chemical insult.