In an effort to improve risk assessments for ozone (O3) we compared the incorporation of inhaled oxygen-18-labeled O3 (18O3) into the lungs of humans and laboratory rats. Cells and fluids obtainable through bronchoalveolar lavage (BAL) were examined after exposure to 18O3 to determine whether excess 18O concentrations (presumed to be reaction products of 18O3) could be detected and equated to the O3 dose to the lung. Three O3 effect measurements (increased BAL protein and neutrophils and decreased BAL macrophages) were also made in subjects or animals exposed in parallel to determine whether there was a correspondence between dose and effect measurements. Eight human male volunteers 18 to 35 yr of age were exposed to 18O3 (0.4 ppm for 2 h) with 15-min alternating periods of heavy treadmill exercise and rest. Rats (F344) were exposed identically, except without exercise. 18O3 was generated directly from pure 18O2. BAL cells and centrifugally separable surfactant material were freeze-dried and analyzed by mass spectrometer for excess 18O. Results showed that the exercising humans had four- to fivefold higher 18O concentrations in all of their BAL constituents than did the rats. The humans also had significant increases in all of the effects markers after 0.4 ppm O3, whereas the rats did not. Rats that were exposed to higher concentrations of 18O3 (2.0 ppm) had levels of 18O in BAL that were more comparable to but still lower than those of exercising humans. Changes in all of the effects markers in these rats were comparable or higher than in exercising humans.(ABSTRACT TRUNCATED AT 250 WORDS)