Erythrocyte micronucleus frequencies in wild fish from two riverine environments and in fish reproduced and reared under controlled conditions (control group) were compared, with the aim to evaluate the suitability of the MN test for the in situ detection of mutagens in freshwaters. Fish were caught in different months in two rivers of central ItaIy which have different pollution levels. As indicator species, the barbel (Barbus plebejus) was chosen because of its ecological significance. Blood samplings wen performed on wild fish immediately after capture and repeated at different time intervals on the same individuals, which were maintained in controlled conditions after capture. A total of 10,000 erythrocytes per specimen were scored. No significant differences in micronucleus frequencies were observed between the control group and fish from the unpolluted river (Mignone). A significantly higher frequency of micronuclei was observed in fish caught in the polluted river (Tibet), in comparison to both the controls and the Mignone river fish. No significant seasonal differences were observed. Barbels examined 50 and 100 days after capture presented a remarkable decrease in micronucleus frequency in comparison with the frequency observed in barbels at capture. The micronucleus test in fish erythrocytes was shown to be a sensitive bioassay for detecting mutagenic pollution in fresh water environments.