Within a cohort of 846 inner-city asthmatic children aged 4 to 9 yr, we looked for subgroups that were more susceptible to the effects of summer ozone. Daily diaries were compared with ambient ozone levels to evaluate effect modification by demographic and environmental characteristics. Children born > 3 wk prematurely or weighing < 5.5 lb. had greater declines in morning % peak expiratory flow rate (PEFR) (1.8% versus 0.3% per 15 ppb ozone, p < 0.05) and a higher incidence of morning symptoms (odds ratio = 1.42 versus 1.09 per 15 ppb ozone, p < 0.05) than did children who had been full-term infants of normal birthweight. Among children who had been of low birthweight (LBW) or had been premature infants, greater declines were seen among those whose reported baseline medication category was "no medication" (3.2% decline) or "steroids" (2.7%) as opposed to beta agonists or xanthines (0.8%) or cromolyn without steroids (0. 1%). Among the children who had been normal birthweight and full-term infants, the cromolyn without steroids group had the greatest declines in %PEFR (1.3%, versus < 0.5% in each of the other three groups). Nonatopic children also had greater responses to ozone. We conclude that among an asthmatic cohort, children who had had an LBW or a premature birth showed the greatest responses to ozone.