The herbicide metolachlor was evaluated for genotoxic potential. Metolachlor did not induce micronuclei in mice, however at 40 mg/kg it significantly decreased the percentage of polychromatic erythrocytes, which is a cytotoxic effect. Metolachlor did not induce chromosomal aberrations in human lymphocytes in vitro, but 2.0 mug/ml culture medium resulted in cytotoxicity, decreasing the mitotic index significantly. The indirect exposure test was carried out by adding plasma from metolachlor-pretreated rats to the human lymphocyte cultures. There was no indication of clastogenicity by metolachlor metabolites. On the other hand, plasma of cyclophosphamide-pretreated rats had a significant clastogenic effect