A chronic feeding study was conducted to determine the carcinogenic potential of oxolinic acid, an antimicrobial agent, in rats and mice. Oxolinic acid was administered in the diet to Wistar rats (0, 30, 100, 300 or 1000 ppm; 50 rats/dose/sex) for 104 wk and to ICR mice (0, 50, 150 or 500 ppm; 50 mice/dose/sex) for 78 wk. Clinical signs, body weight, food consumption, autopsy findings and histopathological data were noted. Mortality was unaffected by oxolinic acid administration in neither species. In rats, body weight gain was suppressed in both sexes at 1000 ppm. Histopathological examinations conducted after autopsy at 104 wk revealed a slight increase in benign Leydig cell tumours of the testis at 1000 ppm, which did not appear until late in the lifetime of rats. No other treatment-related neoplastic lesions were observed in rats. Non-neoplastic lesions in males at 1000 ppm included Leydig cell hyperplasia and tubular atrophy of the testes. In mice, decreased body weight gain was observed in both sexes at 500 ppm, but no non-neoplastic or neoplastic lesions attributable to the treatment with oxolinic acid occurred in either sex. In conclusion, oxolinic acid induced benign Leydig cell tumours of the testis in rats at the highest dose level tested (1000 ppm). The no-effect level for tumour induction was confirmed to be 300 ppm (10.9 mg/kg/day) in rats. None was induced in mice.