Chronic toxicity and carcinogenicity studies were conducted on D-mannitol and propyl gallate in F344 rats and B6C3F1 mice. Groups of 50 rats and 50 mice of each sex were maintained on diets containing either 0, 2.5 or 5.0% D-mannitol or 0, 0.6 or 1.2% propyl gallate for 103 wk. D-Mannitol had no effect on survival or mean body weight of rats and mice, and feed consumption was approximately the same in control and treated groups in each species. Gastric fundal gland dilation occurred at a higher incidence in treated female rats than in controls. A mild nephrosis characterized by focal vacuolization of the renal tubular epithelium was observed in an increased incidence in treated mice. No significant increase in tumour incidence was observed in any of the treated groups in comparison with the corresponding controls. Survival of rats and mice given propyl gallate was similar to that of the controls. Mean body weights were lower in chemically exposed animals, and more so for females. Male rats exposed to propyl gallate showed an increased incidence of hepatic cytoplasmic vacuolization and suppurative inflammation of the prostate gland. Tumours of the preputial gland, islet-cell tumours of the pancreas, and phaeochromocytoma of the adrenal gland occurred at a significantly (P less than 0.05) higher incidence in the low-dose male rats. Malignant lymphoma occurred with a positive trend in male mice (control 1/50, low dose 3/49 and high dose 8/50), and the incidence in the high-dose group was significantly (P less than 0.05) higher than in the control group. However, since the incidence in the control group was much less than the historical control rate (36/398 or 9%) in this laboratory, this apparent increase was not considered to be related to propyl gallate administration. Under the conditions of these studies, neither D-mannitol nor propyl gallate was considered to be carcinogenic to F344 rats or B6C3F1 mice of either sex.