Glutaraldehyde (GA) has a wide spectrum of industrial, scientific and biomedical applications. Its potential to produce chronic toxic and/or oncogenic effects was investigated in Fischer 344 rats (100/sex/group) given GA in drinking water for a maximum of 104 weeks. GA concentrations were 0 (control), 50,250 and 1000 ppm, resulting in average daily GA consumptions, respectively, of 0, 4, 17 and 64 mg/kg for males and 0, 6, 25 and 86 mg/kg for females. Interim euthanasia (10/sex/group) was performed at 52 and 78 weeks. Parameters evaluated were clinical signs, body weight, food and water consumption, hematology, serum chemistry, urinalysis, organ weights, gross and microscopic pathology. There were no treatment-related effects on mortality. Absolute body weights and body weight gains of the 250 and 1000 ppm males and females were reduced over the study in a dosage-related manner. Food and water consumption by the 250 and 1000 ppm groups were decreased in a statistically significant dose-related manner over the study, and mean water consumption by the 50 ppm animals was slightly reduced but not with statistical significance. The 250 and 1000 ppm groups had a dose-related decrease in urine volume with increased osmolality, and pH was slightly reduced. Absolute kidney weights were increased in the 250 and 1000 ppm groups at the 52 and 78 week sacrifices, and decreased at 104 weeks. Relative kidney weights were increased at all sacrifice times for the 1000 ppm group, at 52 weeks for the 250 ppm group, and at 72 weeks for the 50 ppm group. The urinalysis and renal weight changes are compatible with a physiological compensatory adaptation to reduced water consumption. Gross and histological evidence for gastric irritation was observed principally in the 1000 ppm rats euthanized at 104 weeks and in animals that died during the study. Bone marrow hyperplasia and renal tubular pigmentation, seen in rats that died and the 104 week euthanasia animals, may have been secondary to a low grade hemolytic anemia in animals with large granular lymphocytic leukemia (LGLL). The only neoplasm that showed a statistically significant increase was LGLL, which occurred at a high incidence in both sexes and all groups, including the controls, for both animals that died and at the 104 week euthanasia. A few instances of LGLL were observed at 78 weeks. The overall incidence of LGLL in the spleen for the 0, 50, 250 and 1000 ppm groups was, respectively, 43, 51, 40 and 46% for males, and 24, 41, 41 and 53% for females. Statistical analyses indicated that the severity of LGLL was associated with the higher dosages of GA in female, but not male, rats. Due to the background and variable incidence of LGLL in the Fischer 344 rat, the finding of a statistical significance only for female rats, and because, there was no clear dose-response relationship, the biological significance of the LGLL findings is unclear. There is the possibility that the significance was a statistical artifact due to the low incidence of LGLL in the female control animals as a result of biological variability within the study. It is also considered to be possible that the chronic dosage of GA in the drinking water resulted in a modification of one or more of the factors responsible for the expression of this common and spontaneously occurring neoplasm in the Fischer 344 rat.