The effects of four test chemicals [2-acetylaminofluorene (2-AAF), D,L-ethionine (ethionine), butylated hydroxyanisole (BHA), and catechol] were compared in medium- and long-term in vivo systems. In the medium-term assay, animals were sequentially treated with N-diethylnitrosamine (100 mg/kg body weight, i.p., single injection), N-methylnitrosourea (20 mg/kg body weight, i.p., 4 times during weeks 1 and 2), N-butyl-N-(4-hydroxybutyl)nitrosamine (0.05% in the drinking water during weeks 1 and 2), 1,2-dimethylhydrazine (40 mg/kg body weight, s.c., 4 times during weeks 3 and 4) and dihydroxy-di-N-propylnitrosamine (0.1% in the drinking water during weeks 3 and 4) for multi-organ initiation, and then treated with one of the four test chemicals for 24 weeks, and killed at week 28 (group 1). In the long-term assay, animals were treated in the same manner and then given basal diet and tap water (group 3) or test chemical continuously (group 4) for the remainder of the lifespan. Animals receiving multi-organ initiation and then maintained on basal diet for 24 weeks (group 2) or their lifespan (group 5) served as controls. Detailed histopathological examinations were performed on all rats. Hepatocellular carcinoma incidences in the long-term assay were found to reflect closely the respective medium-term results. Induction of proliferative forestomach or glandular stomach lesions by BHA and/or catechol, and bladder lesions by 2-AAF and BHA in the medium-term assay also correlated with tumor development in the long-term. Furthermore, inhibition of thyroid proliferative lesions by all test chemicals corresponded with low thyroid tumor incidences in the long-term assay. The observed strong correlation between medium- and long-term results confirms the applicability of our medium-term multi-organ carcinogenesis bioassay system for detection of modifying effects of test chemicals in different organs.