Recent efforts to improve risk assessment methodologies have sought to provide a fuller representation of the variability and uncertainty in risk estimates in order to provide risk managers with a more complete description of risks. Recently, we and others (Swartout et al. , 1998; Price et al. , 1997; Slob and Pieters, 1997; Baird et al. , 1996) have proposed approaches to characterize the uncertainty in the reference dose, (RfD) a key component of the non-carcinogenic risk estimation process. The operational definition of the RfD as the "lower-bound" estimate of the NOAEL in a sensitive human subpopulation (NOAEL HS ) is used along with information on the inter-chemical variation in ratios associated with the uncertainty factors used in setting the RfD to characterize the uncertainty in the NOAEL HS (Swartout et al. , 1998). This paper presents a description of how information on the uncertainty in the NOAEL HS can be used to characterize the uncertainty and variability in estimates of non-carcinogenic risks for individuals and populations. The paper also explores the impact of using alternative estimates of the inter-chemical variation in the ratios. The analysis suggests the following findings. First, the current method of estimating risks from mixtures of chemicals may overestimate the risks when two or more compounds contribute significantly to the estimate of the hazard index. Second, the uncertainty in the NOAEL HS of a mixture is affected by the number and size of the uncertainty factors used in setting the RfDs of the mixture's components. Third, jointly assessing both the uncertainty and variability in exposure and the uncertainty in the estimate of the NOAEL HS can have a significant impact on the characterization of noncarcinogenic risks for individuals and populations. Finally, these findings are independent of the choice of distributions for interchemical variation in ratios.