We assessed the contributions of particulate matter with aerodynamic diameters < or =10 and < or =2.5 microm (PM2.5 and PM10) and ozone (O3) to peak expiratory flow (PEF) and respiratory symptoms in 40 schoolchildren 8-11 years of age for 59 days during three periods in 1991 at a school in southwest Mexico City. We measured peak expiratory flow in the morning on the children's arrival at school and in the afternoon before their departure from school. Separately for morning and afternoon, we normalized each child's daily measurement of peak flow by subtracting his or her mean peak flow from the daily measurement. Child-specific deviations were averaged to obtain a morning and afternoon mean deviation (APEF) for each day. Mean 24-hour O3 level was 52 parts per billion (ppb; maximum 103 ppb); mean 24-hour PM2.5 and PM10 were 30 microg/m3 (maximum 69 microg/m3) and 49 microg/m3 (maximum 87 microg/m3), respectively. We adjusted moving average and polynomial distributed lag multiple regression analyses of APEF vs pollution for minimum daily temperature, trend, and season. We examined effects of PM2.5, PM10, and O3, on deltaPEF separately and in joint models. The models indicated a role for both particles and O3 in the reduction of peak expiratory flow, with shorter lags between exposure and reduction in peak expiratory flow for O3 than for particle exposure (0-4 vs 4-7 days). The joint effect of 7 days of exposure to the interquartile range of PM2.5 (17 microg/m3) and O3 (25 ppb) predicted a 7.1% (95% confidence interval = 11.0-3.9) reduction in morning peak expiratory flow. Pollutant exposure also predicted higher rates of phlegm; colinearity between pollutants limited the potential to distinguish the relative contribution of individual pollutants. In an area with chronically high ambient O3 levels, school children responded with reduced lung function to both O3 and particulate exposures within the previous 1 to 2 weeks.