Comparing 21 rotogravure printers exposed to toluene (medians: time-weighted air level 150 mg/m3, blood toluene 1.6 mumole/l) and 21 unexposed controls (median blood toluene less than or equal to 0.01 mumole/l) there was a significant increase in the frequency of micronuclei (MN) in pokeweed mitogen (PWM)-stimulated peripheral blood lymphocytes in the printers, as compared to the controls (2.8% vs. 1.5%; p = 0.03; all p adjusted for age and smoking). The frequency of small MN (size ratio MN/main nucleus less than or equal to 0.03) in PWM-stimulated lymphocytes was associated with the exposure (1% vs. 0.3%; p = 0.05). Furthermore, among the exposed subjects there was an association between blood toluene and small MN (0.17% per mumole/l; p = 0.0005). Small MN in phytohemagglutinin (PHA) cultures displayed no association with any exposure parameter. However, in the printers, an estimated cumulative exposure index was weakly correlated with the frequency of total MN in PHA-stimulated cells (0.00003% per mg/m3 x year; p = 0.07). Among the printers, chromosomal breaks in PHA-stimulated cells were associated with the duration of earlier benzene exposure (0.03% per year; p = 0.01). The results of this study strongly indicate that toluene causes a clastogenic effect on the B-cells even at low exposure levels. Further, earlier benzene exposure seems to have caused chromosomal breaks in T-cells.