This study was conducted in Switzerland between May and October 1989 to assess possible decrements in lung function occurring as a result of 10 minute exposure to ambient air containing different ozone concentrations. Once a month, 128 children in two different areas of Southern Switzerland (Chiasso and Aurigeno) had a pulmonary function test before and after a standardized 10 minute exercise (pulse rate, 170/min) on a cycle ergometer, outdoors. Ozone concentrations were similar in both areas, ranging from 40 to 157 micrograms/m3 (1/2h means) during the exercise tests. The two communities differed with respect to long-term average pollution levels. The mean NO2 concentration over the six months study period was 70 micrograms/m3 in Chiasso and 18 micrograms/m3 in rural Aurigeno. Of the eligible children 85% participated and attended 4-6 tests. Parents completed a standardized questionnaire on family background, home characteristics and the child's early and present illness history. A total of 500 acceptable pairs of spirograms and corresponding ozone concentrations (average 3.8 per child) were available for analysis. Regressions of each individual's pre-post differences of FVC, FEV1 and peak flow on ozone concentrations measured during the outdoor exercise indicated that elevated ozone levels significantly reduced peak flow values. Adjustments for temperature and relative humidity increased the magnitude of the peak flow slopes. The average adjusted regression coefficient for delta-peak flow on ozone was -2.28 mL/s/micrograms/m3 (95% CI, -0.57 -3.99). It is noteworthy that the observed relationships occurred at ozone concentrations below 160 micrograms/m3 and after an exercise duration of only 10 minutes.