OBJECTIVE: To assess the natural history of respiratory symptoms not labelled as asthma in primary schoolchildren. DESIGN: Repeat questionnaire survey of subgroups identified from a previous questionnaire survey after a two year delay. SUBJECTS: The original population of 5321 Sheffield children aged 8-9 years yielded 4406 completed questionnaires in 1991(82.8%). After excluding children with a label of asthma, there were 370 children with current wheeze, 129 children with frequent nocturnal cough, and a random sample of 222 children with minor cough symptoms and 124 asymptomatic children. RESULTS: Response rates in the four groups were 233 (63.0%), 77 (59.7%), 160 (72.1%), and 90 (72.6%) respectively. Of those who initially wheezed, 114 (48.9%) had stopped wheezing and 42 (18.0%) had been labelled as having asthma. Those with more frequent wheezing episodes (p < 0.02) and a personal history of hay fever (p < 0.01) in 1991 were more likely to retain their wheezy symptoms. In the children with frequent nocturnal cough in 1991, 20.1% had developed wheezing, 42.9% had a reduced frequency of nocturnal coughing, and 14.2% had stopped coughing altogether two years later. One sixth had been labelled as having asthma. Children with nocturnal cough were more likely to develop wheezing if they had a family history of atopy (p = 0.02). Only 3.8% and 3.3% of those with minimal cough and no symptoms respectively in 1991 had developed wheeze by 1993 (1.9% and 1.0% labelled as asthma). CONCLUSIONS: Most unlabelled recurrent respiratory symptoms in 8-10 year olds tend to improve. Unlabelled children who have persistent symptoms have other features such as frequent wheezing attacks and a family or personal history of atopy. If a screening questionnaire were to be used to identify such children, a combination of questions should be employed.