1. We aimed to study whether passive smoking is associated with respiratory ill health in primary school children in Hong Kong. 2. In this cross-sectional study, a standardised structured questionnaire was used to obtain information on respiratory symptoms, smokers in the household and smoking habit of the children. A random sample of 30 primary schools was included. Four classes were randomly selected for each school, one class each from each school year of primary 3-6. A total of 3964 children aged 8 - 13 years completed the questionnaire during October 1995-May 1996. The response rate was over 94%. 3. In 3480 children who had never smoked, after adjusting for gender, age, place of birth and living district, the odds ratios (95% confidence interval) were: (a) throat problems, 1.35 (1.08-1.68); (b) cough, 1.54 (1.28-1.84); (c) phlegm, 1.43 (1.21-1.70); (d) wheezing, 1.21 (0.89-1.41); (e) nose problems, 1.17 (1.02-1.35). The odds ratios increased with increasing number of smokers at home (P for trend < 0.001; except for wheezing, P=0.3). The odds ratios for having any of the above symptoms for one, two and three smokers at home were respectively 1.15 (0.99-1.34), 1.48 (1.16-1.88) and 2.03 (1.47 - 2.81); P for trend < 0.001. 4. We conclude that the results provide further evidence that passive smoking is a cause of respiratory ill health in school children in Hong Kong.