Many substances that produce largely chronic health effects should have an LTA OEL. However, the LTA OEL should be in addition to a TWA OEL. I am not aware of any chronic disease agent that does not have short term effects that can be ignored. For a chronic disease agent, short term exposures could refer to a period as short as 15 minutes for those agents with a short term exposure limit in addition to a TWA OEL, or as long as days, weeks, or months, depending upon the particular toxicology of the substance. Indeed, many have short term exposure limits which immediately render the long term interpretation invalid. Others for example, silica and cadmium can produce disease due to excessive exposures over a period of days, weeks, months, or a few years. Crystalline silica produces chronic, accelerated, and acute silicosis, depending on the frequency, duration, and severity of exposure. Cadmium at lower levels is a chronic toxicant producing lung cancer and kidney disease, but excessive exposure within a single shift can cause life threatening pneumonitis and pulmonary edema. Rappaport and colleagues interpretation of the OELs assumes that dose rate is unimportant and that Haber's rule the notion that toxic effect is directly proportional to the product of concentration and time applies without exception. A legitimate LTA OEL is useful for gauging whether or not the true average exposure of each employee is actually being limited to a reasonable value. 14 In the absence of an official LTA OEL, the AIHA EASC 3 recommendation of one third the TWA OEL is an excellent first approximation. IHs should, as they accumulate exposure data, calculate long term averages to determine if the long term goal implicit or explicit in each TWA OEL is being met. However, it is improper to create an LTA OEL that is numerically equal to the TWA OEL. TWA OELs were never meant to be interpreted as multi month or multi-year averages. In light of the fact that most of OSHA's Z-table PELs are nearly 30 years old and may not be sufficiency protective, even when properly interpreted as a single shift limit, the insistence of Rappaport and colleagues that this can be done is particularly unsettling. The ANOVA based methodology espoused by Rappaport and colleagues may point the user in a promising direction 9 with regard to determining whether intervention should be focused on factors that affect the entire group or factors associated with each individual. Such procedures have already been used effectively in research studies to analyze the impact of workplace variables: for example, see Woskie et al. 18 Whether or not this is the most efficient and effective use of limited resources for distinguishing between group related and task work practice related determinants of exposure is another issue, and is not addressed by Rappaport and colleagues. Their procedure should be seriously considered for determining compliance with a legitimate LTA OEL. The notion that such procedures can be used to determine compliance with existing OELs is unequivocally wrong. The adoption of new or revised OELs that are expressed in terms of both long term average exposures and single shift limits is preferable. There are limitations to the more mainstream approaches suggested by AIHA, 3 CEN 13 and others.14 Because of limited resources, some IHs may evaluate group exposure profiles where the groups are less than reasonably homogeneous. Other IHs may have been provided resources that permit no more than one or a couple measurements per exposure group per year. In the latter case, a strict compliance testing strategy may be the best analysis technique for managing risk. These limitations have been acknowledged in the industrial hygiene literature and can be mitigated through the use of professional judgment, judicious use of available resources, identification of critical exposure groups, prioritization between exposure groups, task and work practice analysis, periodic reassessment of exposure group definitions, and so on. We should continually press for additional resources and continually seek to control exposures through followup surveys and work practice analysis. Toward this end, the publications of Dr. Rappaport and colleagues have assisted in identifying areas where improvement can be made. However, exposure assessment is an ongoing, iterative process that cannot be reduced to the mechanistic collection and analysis of a large number of repeat measurements. In future publications, Rappaport and colleagues should consider the following recommendations: 1. Provide examples i.e., exposure limits for specific gases, vapors, and particulates where the LTA OEL interpretation truly applies. 2. Demonstrate through case studies that the ANOVA based methodology is, given fixed resources, more efficient and/ or more effective than conventional methods in identifying work environments in need of intervention. 3. Acknowledge that the LTA OEL interpretation does not apply to OSHA PELs either Z table or 6b standards or to the majority if not all of the ACGIH TWA TLVs. 4. Carefully explain that their definition of overexposure completely ignores dose rate, and therefore applies to few, if any, chronic disease agents.