"This report describes the implications of exposure variability upon the assessment of occupational exposures to chemicals. It begins by tracing the development of occupational health standards from the Threshold Limit Values to the current permissible exposure limits (PELs) issued by the Occupational Safety and Health Administration in order to demonstrate that these limits implicitly define overall levels of exposure. The notion of the exposure distribution is then introduced and the lognormal model is used to illustrate the characteristics of occupational-exposure data. Three strategies for evaluating occupational exposures are explored. The first@ based upon a PEL never to be exceeded@ represents the current regulatory situation; this strategy is flawed because compliance outcomes depend critically upon the number of exposure measurements. The second strategy@ based upon an allowable level of exceedance@ allows rigorous statistics to be applied (tolerance limits) but has limited power and cannot identify specific exposure distributions. The third strategy@ termed a ""limiting distribution"" constrains both the mean and the variance of the exposure distribution; thus@ it overcomes the difficulty of the exceedance strategy in differentiating various situations. One possible method is discussed which uses the elimination half time of the chemical to determine the parameters of the limiting distribution and which tests exposures in a hypothesis-testing framework ."