Diets containing 2.5 or 5% D-mannitol were fed to groups of 50 F344 rats and 50 B6C3F1 mice of each sex for 103 wk. Similar groups served as controls. There were no significant differences in survival between treated and control rats or between treated and control mice. Mean body weights were similar in treated and control male rats and in treated and control female mice. Throughout the study, the mean body weights of female rats on the 5% diet were slightly (less than 10%) lower than those of the controls, and by the end of the study the mean body weights of treated male mice were slightly (c. 10%) higher than those of the controls. Feed consumption by treated and control animals was approximately the same in rats and mice of either sex. The incidence of dilation of the gastric fundal gland was higher (46%) in treated female rats than in the controls (12%). A mild nephrosis, characterized by focal vacuolization of the renal tubular epithelium, showed an increased incidence in treated mice of both sexes and was considered to be related to the administration of D-mannitol. There were no statistically significant increases in tumour incidence in any of the treated groups when compared with the corresponding controls. Under the conditions of this bioassay, D-mannitol was not shown to be carcinogenic for F344 rats or B6C3F1 mice of either sex.