This study on the in vitro micronucleus assay, comprising 11 laboratories using human lymphocytes, was coordinated by an organizing committee supported by the SFTG (the French branch of the European Environmental Mutagen Society). Nine coded substances were assessed for their ability to induce micronuclei in human lymphocytes in vitro, mitomycin C being used as a positive control. Cultures were exposed to the test substances for a short (early or late) time or for a long time, followed by a short or long recovery period, in the presence of cytochalasin B. Each chemical was evaluated, generally in two laboratories, using three treatment schedules at least twice. The data were assessed for acceptability, and then classified as negative, positive or equivocal. Two of seven genotoxic compounds, namely colchicine and bleomycin, clearly induced micronuclei. Reproducible results were difficult to obtain for some substances, which tended to be those acting at specific stages of the cell cycle. Cytosine arabinoside, diethylstilboestrol and 5-fluorouracil were classified as equivocal. Urethane and thiabendazole were classified as negative. The two presumed non-genotoxic compounds, mannitol and clofibrate, did not induce micronuclei. Repeat testing, exposing cells at both an early and late time after mitogenic stimulation, was needed to detect substances classified as equivocal. These results show the importance of achieving sufficient inhibition of nuclear division to avoid the possibility of missing an effect. The evaluation of micronuclei in mononucleated as well as binucleated cells was particularly useful to detect aneugens. There were no false positive results using lymphocytes, indicating a high specificity. It is concluded that the clastogenic or aneugenic potential in vitro of the substances tested was correctly identified in this study, but that refining the protocol to take into account factors such as the stages of the cell cycle exposed to the compound, or the duration of recovery would be likely to improve the sensitivity of detection using lymphocytes.