The present work was designed to determine the potential genotoxicity at the vicinity of a solid waste incinerator in the metropolitan area of São Paulo, using the Tradescantia stamen-hair bioassay. Experiments were carried out between December 1998 and April 1999 in four regions (40 pots of plants per site) selected on the basis of their pollution levels predicted by theoretical modeling of the dispersion of the incinerator's plume. The exposure sites were defined as follows: highest level (incinerator); a high level (museum) located 1.5 km from the emission point; a moderate level (school, at a distance of 3.5 km from the incinerator); and a control (at Jaguariúna countryside). The difference in genotoxicity among the groups was statistically significant (p < 0.001). The frequency of mutations observed in the countryside was significantly lower [2.25 +/- 1.55, mean +/- SD (standard deviation)] than that of the sites close to the incinerator. The frequency of mutations measured at the school (3.70 +/- 1.36) was significantly lower than that measured at both the museum (4.89 +/- 1.12) and the incinerator (5.69 +/- 1.34). In conclusion, we found a positive correlation between the spatial distribution of the emissions of the incinerator located in an urban area and the mutagenic events measured by the Tradescantia stamen-hair assay. The in situ approach employed in this study was simple, efficient, and of low cost. No air or chemical extraction of pollutants was necessary for genotoxicity testing as required by other assays.