Human health effects due to chronic exposure to ozone (O3) have not been established due to problems with exposure assignment and the use of measures of lung function which may not reflect the site of O3 toxicity in the lung. We investigated the feasibility of retrospective assessment of O3 exposure-relevant covariates and derived lifetime "effective exposure" to ozone. Mid- and end-expiratory flows (FEF25-75%, FEF75%) were regressed against effective exposure and ecological lifetime exposure. A convenience sample of 130 UC Berkeley freshmen, ages 17-21, participated twice in the same tests (residential history, questionnaire, pulmonary function), 5-7 days apart. Students had to be lifelong residents of Northern (SF) or Southern (LA) California. Monthly ambient O3 concentrations (OZ) were assigned based on the lifetime residential history. An "effective time" (T) spent in OZ environments was derived for each residence and age stratum (0-2, 3-5, 6-11, 12+) with the use of questions about "total time spent outdoors" and time spent in "moderate" and/or "heavy" activity. Effective exposure was calculated over the lifetime (OZ x T) of each subject. Ozone metrics used were 8-hr averages (10 AM-6 PM) and "hours above 60 ppb." FEF25-75% and FEF75% decreased with both effective exposure and ecologic assignment of O3 exposure. For a 20 ppb increase (interquartile range) in 8-hr O3, FEF75% decreased 334 ml/sec (95%Cl:11-657 ml/sec), which corresponds to 14% (1.0-28.3%) of the population mean FEF75%. The corresponding effect on FEF25-75% was -420 ml/sec (95%Cl: +46 to -886, P = 0.08) or 7.2% of the mean. Use of time-activity data to define exposure had no impact on estimates. Negative confounding factors were region (SF vs LA), gender, and ethnicity. Lifetime 8-hr average O3 concentrations ranged from 16 to 74 ppb with little overlap between regions. There was no evidence for different O3 effects across regions. Effects were independent of lifetime mean PM10, NO2, temperature, or humidity. Effects on FEV1 tended to be negative whereas those for FVC, although negative in some models, where inconsistent and small. The strong relationship of lifetime ambient O3 on mid- and end-expiratory flows of college freshmen and the lack of association with FEV1 and FVC are consistent with biologic models of chronic effects of O3 in the small airways. Since the present study was designed as a pilot study, these findings have to be confirmed in a larger sample that is representative of the target population.