mented with data from several other studies, were considered adequate for derivation of the three AEGL classifications for five time periods.

The AEGL-1 was based on the observation that adult volunteers could tolerate exposure to 10 ppm for 15 min without irritant effects (Keplinger and Suissa 1968). Although this value is below the definition of an AEGL-1 (slight irritation), it provides the longest controlled exposure duration for which no irritation in humans was reported. An intraspecies uncertainty factor of 3 was applied because fluorine is highly corrosive to the tissues of the respiratory tract and effects are not expected to vary greatly among individuals, including susceptible individuals (NRC 2001). Although no data on asthmatics were found, the uncertainty factor of 3 was considered adequate to protect this sensitive subpopulation because the value was a NOAEL and because shorter-term, repeated exposures produced no substantially greater effects in healthy individuals. The value is supported by a second study in which volunteers “tolerated” exposure to 10 ppm for an undefined period of time (Belles 1965). A modifying factor of 2 was applied based on a limited data base and short exposure durations. The resulting value of 1.7 ppm was used across all AEGL-1 exposure durations because, at mildly irritating concentrations, adaptation to slight sensory irritation occurs. As noted, this value is supported by limited workplace monitoring data: workers exposed to fluorine at average yearly concentrations up to 1.2 ppm (range, 0.0-17 ppm) over a four-year period reported fewer incidences of respiratory complaints or diseases than a similar group of nonexposed workers (Lyon 1962). The workers are assumed to encompass a small range of sensitivity; the additional intraspecies uncertainty factor of 3 was considered sufficient to protect sensitive individuals.

Mild lung congestion was selected as the threshold for irreversible, long-lasting effects as defined by the AEGL-2. The AEGL-2 was based on an animal study in which mild lung congestion was observed in mice at 67 ppm for 30 min and 30 ppm for 60 min (Keplinger and Suissa 1968). Effects were slightly less serious in three other species. Although concentrations causing irritant effects or lethality in three other species for the same time periods suggested similar species sensitivity, the mouse data, because of slightly lower values, were chosen as the basis for developing the AEGL-2 and AEGL-3. Similar sensitivity was observed among all species in the key study; therefore, an interspecies uncertainty factor of 1 was applied to address interspecies variability. Fluorine is a highly corrosive gas that reacts directly with the tissues of the respiratory tract, with no pharmacokinetic component involved in the toxicity; therefore, there is likely to be little difference among individuals in response to fluorine at concentrations that define the AEGL-2. The 30- and 60-min values for the mouse were divided by an intraspecies uncertainty factor of 3 to protect sensitive individuals, since effects are not likely to differ greatly among individuals, and by a modifying factor of 2, based on a limited data base. The 30-min value was time scaled to the 10-min AEGL-2, and the 60-min value was time scaled to the 4-h AEGL-2 value. Time scaling was based on the C^1.77 × t = k relationship. The value of n was derived from regression analysis of the mouse lethality data in the key