Attempts have been made to use the hypoxanthine-guanine-phospho-ribosyl-transferase-assay as a method for automated screening of agent-induced phenotypic variants of human peripheral lymphocytes reflecting 6-thioguanine resistance and assumed to indicate genotoxic action. Different protocols of the hypoxanthine-guanine-phospho-ribosyl-transferase-system were used in this study in order to investigate whether the system can be a candidate for a short-term test for a rapid and reliable identification of biological systems exposed to agents. The current protocols were based on: 1) fluoresceinated monoclonal antibodies against bromodeoxyuridine-DNA for labelling of 6-thioguanine-resistant human lymphocytes and direct flow-cytometric enumeration of bromodeoxyuridine-positive events and: 2) indirect flow-cytometric enrichment of 6-thioguanine-resistant cells labelled with 3H-thymidine followed by autoradiographic enumeration of positive events. Both the direct and the indirect enumeration method yielded similar results down to the range 10(-4) with respect to frequency of variants. For the less time-consuming direct enumeration method the resolution was limited due to non-specific binding of the antibody and false positives. It was, nevertheless, sufficient to score variants induced in vitro with the mutagens EMS, MMC and TT in the same range as e.g. that of cancer patients during and after chemotherapy or radiotherapy, or that of psoriasis patients during the after PUVA (8-methoxypsoralen and long range UV light)-therapy. We conclude that the direct enumeration protocol can be used for a rapid screening of so called outliers, but a more sensitive test, such as the more time-consuming enrichment protocol based on autoradiography, must be used in order to score variants in the range 10(-5)-10(-6).