Sucrose acetate isobutyrate (SAIB), a food additive used as a flavour emulsion stabilizer in citrus-based soft drinks, was evaluated for chronic toxicity in B6C3F1 mice and Fischer 344 rats. SAIB dissolved in acetone was blended into NIH07 rodent diet at concentrations that were adjusted weekly during the first 12 to 18 months of the studies so that ingested dose levels per kg body weight were constant. Groups of 20 rats per sex were given dose levels of 0.0, 0.0, 0.5, 1.0 and 2.0 g SAIB/kg body weight for 1 yr, and groups of 50 rats per sex were given dose levels of 0.0, 0.0, 0.5, 1.0 and 2.0 g SAIB/kg body weight for 2 yr. Mice were fed dose levels of 0.0, 0.0, 1.25, 2.5 and 5.0 g SAIB/kg body weight for 2 yr. The highest doses fed, equivalent to dietary concentrations of approximately 5%, were considered to be the maximum concentrations that could be fed without risk of nutritional deficiencies. Depressions in body weight gain were noted, particularly in female rats during the first 12 to 18 months of the studies. Recovery during the last quarter of the 2-yr study suggests that the reduced body weight gain was nutritional rather than SAIB-related. There were no differences in survival between SAIB-treated rats or mice and controls. Decreased body weight gains, primarily in females, but less consistent than those in the rat, were noted in the 2-yr mouse study. No signs of toxicity were observed in clinical chemistry, haematology, organ weights, gross necropsy findings or light microscopy studies in the 1- or 2-yr rat studies. Electron microscopic examinations of liver sections from high dose level rats from the 1-yr study also revealed no effects of SAIB treatment. There were no significant increases in benign or malignant tumours in the long-term rat or mouse carcinogenicity studies. The lowest no-observed-adverse-effect level (NOAEL) was 2 g SAIB/kg body weight derived from the 1- and 2-yr chronic toxicity studies in the rat.