STUDY OBJECTIVE--The aim was to investigate the relation between passive smoking and childhood acute respiratory illness. DESIGN--The study involved an initial postal survey on a random sample of children followed by a case-control study based on the survey. A respiratory illness score was calculated from maternal reports of episodes of illness in the previous 12 months. SETTING--The study was a population survey based on Adelaide metropolitan area in South Australia. PARTICIPANTS--The reference population (n = 13,996) was all live born children registered in South Australia in 1983 whose parents lived in Adelaide metropolitan area. Of these, 4985 families were contacted by post and from 2125 respondents, 1218 (58%) gave consent for home interview. "Cases" were children with respiratory illness scores in the top 20%, controlling for age and time of year (n = 258); "controls" were taken in the bottom 20% (n = 231). MEASUREMENTS AND MAIN RESULTS--Maternal smoking in the first year of life was associated with a doubling in relative odds of respiratory proneness in the child (odds ratio = 2.06, 95% CI 1.25-3.39) after adjustment for confounding by parental history of respiratory illness, other smokers in the home, use of group child care, parent's occupation, and levels of maternal stress and social support. There was no evidence that this association was attributable to differences in the way smoking and non-smoking parents perceived or managed childhood acute respiratory illness. Maternal smoking in the first year, without smoking in pregnancy, was also associated with increased risk of respiratory proneness (odds ratio 1.75, 95% CI 1.03-3.0), showing an effect of passive smoking independent of any in utero effect. There was a strong negative effect modification by breast feeding: relative odds of respiratory proneness with maternal smoking were seven times higher among children who were never breast fed than among those who were breast fed. CONCLUSIONS--The results suggest a relatively small but real effect of passive smoking on childhood acute respiratory illness. Effect modification by breast feeding may be due to a combination of behavioural and biological mechanisms.