Commercially available regular instant coffee was given in the diet to barrier-maintained, specified pathogen-free Swiss mice for 2 yr. Groups of 150 males and 150 females were fed diets containing 10, 25 or 50 g instant coffee powder/kg. The animals had already been exposed to coffee in utero. Coffee increased the energy expenditure of the animals as shown by increased daily calorific intake and depressed growth. The overall tumour incidence was inversely correlated to the coffee intake, and no unusual tumour or site of origin was found. The most frequent neoplasms were lymphosarcomas, bronchiolo-alveolar adenomas and adenocarcinomas, as well as hepatocellular adenomas. The incidence of total neoplasms (benign and malignant) decreased from 70.6 and 56.8% in control males and females, respectively, to 34.8 and 36.2%, respectively, in the high-dose group. This decrease, which was essentially due to a reduction in the number of lymphosarcomas and hepatocellular adenomas, was associated with a slower growth rate. The number of leiomyomas in the uterus was slightly increased due to coffee intake as shown by the analysis of positive trend (P less than or equal to 0.05). However, the incidence of this benign tumour was very low; 2.72% of mice affected in the high-dose group, 1.37% in the low-dose group and 0% in the control and medium-dose groups. From this study it is concluded that instant coffee did not increase the incidence of malignant neoplasms in mice when fed at dietary levels of up to 5% for 2 yr.