A chronic feeding study was carried out in the F1a generation of dosed Fischer 344 rats of both sexes with gentian violet (GV). The test substance was administered in the diet to 570 male and 570 female rats at dose levels of 0 (control), 100, 300 and 600 ppm for 24 months. Rats were killed and necropsied after 12, 18 and 24 months of continuous dosing. Measurements of body weights, food consumption (and dose rate) and mortality and the results of histopathological examination were analysed statistically. Male and female rats fed 600 ppm GV for 24 months showed a decrease in body weights. Average food consumption based on g food/kg average body weight was essentially equal in all groups. Mortality at the end of the study (24 months) was approximately 33% in the controls for both males and females and approximately 66% in females of the high-dose group and 48 and 39% in males of the mid- and high-dose groups, respectively. All dose-related neoplastic pathology was noted at the final necropsy. Following 24 months of dosing, there was a significant difference from the controls in the incidence of follicular cell adenocarcinoma of the thyroid gland for both males (600 ppm GV) and females (300 and 600 ppm GV). Although the incidences were very low, statistical analysis showed a significant difference from the controls for hepatocellular adenomas in the mid-dose group of the females and the mid- and high-dose groups of the males. A dose-time-related incidence of mononuclear cell leukaemia was also noted in the females. There was high background incidence of the leukaemia. Several non-neoplastic dose-related lesions were observed in both males and females, principally in the 18- and 24-month necropsies. Almost all of these lesions were focal changes in the liver, many of which were probably related to the mononuclear cell leukaemia.