The pharmacokinetic concepts of bioavailability and biotransformation are introduced into the assessment of public health risk from experimental data concerning the emissions of potentially mutagenic and carcinogenic substances from motor vehicles. The inappropriateness of an automatic application in the risk assessment process of analytical or experimental results, obtained with extracts and procedures incompatible with the biological environment, is illustrated on the discrepancy between short-term laboratory tests predictions that wider use of diesel engines on our roads will increase the risk of respiratory cancer and the widely negative epidemiological evidence. Mutagenic activity of diesel particulates was minimal or negative when tested in extracts obtained with biological fluids, was substantially dependent on the presence of nitroreductase in the microbial tester strain, and disappeared completely 48 hr after the diesel particles had been phagocytized by alveolar macrophages. Similarly, long-term animal inhalation exposures to high concentrations of diesel particles did not induce the activity of hydrocarbon metabolizing enzymes or specific adverse immune response unless organic solvent extracts of diesel particles were administered intratracheally or parenterally in doses that highly exceed the predicted levels of public exposure even by the year 2000. Furthermore, the suspected cancer producing effects of inhaled diesel particles have thus far not been verified by experimental animal models or available long-term epidemiological observations. It is concluded that unless the biological accessibility of the active component on the pollutant as well as its biotransformation and clearance by natural defense mechanisms are considered, lung cancer risk assessment based solely on laboratory microbial tests will remain an arbitrary and unrealistic process and will not provide meaningful information on the potential health hazard of a pollutant.