During the process of deriving oral Reference Dose (RfDs) values for chemical warfare agents, several issues arose regarding the identification of adverse effect levels and the application of uncertainty factors. For those agents that function as cholinesterase inhibitors (e.g., agents VX, GA, GB, and GD), these issues included the following: (1) Is the endpoint of blood cholinesterase inhibition an indicator of toxicity or a biomarker of exposure? (2) Can an experimental animal species be more sensitive than humans, thereby eliminating the need for an animal-to-human uncertainty factor? (3) Can the uncertainty factor that is used to extrapolate from a lowest-observed adverseeffect-level (LOAEL) to a no-observed-adverse-effect-level (NOAEL) be less than the default value of 10? (4) Can an oral RfD be derived from non-oral toxicity data? (5) Can an uncertainty factor of less than 10 be used to extrapolate from subchronic to chronic exposure (i.e., is the critical effect adequately described by the subchronic exposure data)? (6) What constitutes an adequate data base for organophosphate cholinesterase inhibitors, and what uncertainty factor should be used for an incomplete data base? Young, et al.Analysis of relevant data resulted in the following selection and justifications of uncertainty factors. For uncertainty associated with intraspecies extrapolation (UF H ), physiologic and pathologic conditions affecting cholinesterase activity levels justified maintaining a UF H of 10 for all of the nerve agents. Because available data indicated that humans tended to be more sensitive than rats regarding anticholinesterase effects, an interspecies variability (UF A ) factor of 10 was retained for agents GA, GB, and GD. For agent VX, however, the available data revealed that the domestic sheep test species exhibited sensitivity equivalent to or greater than that of humans thereby justifying a UF A of 1. For uncertainties regarding extrapolation from subchronic-to-chronic exposure data, consideration of information on the physiology of cholinergic systems and the available toxicity data for the nerve agents and other cholinesterase inhibitors indicated that a UF S of 3 was justified for all four of the nerve agents. For uncertainties regarding LOAEL-to-NOAEL extrapolation (UF L ), the selection of agent GB, GD, and VX doses resulting in cholinesterase inhibition in the absence of clinical signs of toxicity (biomarker of exposure) justified this endpoint as a minimal LOAEL and a UF L of 3. For agent GA, a NOAEL was used, and therefore no UF L was required. The uncertainty factor for data base completeness (UF D ), was based upon several considerations. Of primary concern was the fact that chronic toxicity studies are not considered an essential component of the data base requirements for cholinesterase inhibitors because of the unlikelihood that the endpoint will change with an increase in exposure time beyond that defined as a subchronic exposure. Additionally, limited data regarding reproductive and developmental toxicity ...