Cancer risk assessment methods for chemical mixtures in drinking water are not well defined. Current default risk assessments for chemical mixtures assume additivity of carcinogenic effects, but this may not represent the actual biological response. A rodent model of hereditary renal cancer (Eker rat) was used to evaluate the carcinogenicity of mixtures of water disinfection by-products (DBPs). Male and female Eker rats were treated with individual DBPs or a mixture of DBPs for 4 or 10 months. Potassium bromate, 3-chloro-4-(dichloromethyl)-5-hydroxy-2(5H)-furanone, chloroform, and bromodichloromethane were administered in drinking water at low concentrations of 0.02, 0.005, 0.4, and 0.07 g/l, respectively, and high concentrations of 0.4, 0.07, 1.8, and 0.7 g/l, respectively. Low and high dose mixture solutions comprised all four chemicals at either the low or the high concentrations, respectively. Body weights, water consumption, and chemical concentrations in the water were measured monthly. All tissues were examined macroscopically for masses and all masses were diagnosed microscopically. Total renal lesions (adenomas and carcinomas) were quantitated microscopically in male and female rats treated for 4 or 10 months. A dose response for renal tumors was present in most treatment groups after 4 or 10 months of treatment. Treatment with the mixture produced on average no more renal, splenic, or uterine tumors than the individual compound with the greatest effect. This study suggests that the default assumption of additivity may overestimate the carcinogenic effect of chemical mixtures in drinking water.