The aim of this study was to discover the subsequent histories and ultimate diagnoses in 182 patients studied in a chest clinic because of unexplained cough. Twenty-nine patients (16%) developed asthma during the mean follow-up time of 4.4 years. Chronic bronchitis was diagnosed in 18 patients (10%) and sporadic instances of some other diseases were also seen. The development of asthma was studied both by calculating simple relative risks and using a multiple logistic model. Circadian changes in peak expiratory flow (PEF)-values and total blood eosinophil count were found to be the best variables for predicting the risk of developing asthma in a patient with unexplained cough, whereas mild bronchial hyperreactivity had only a low predictive value. If the risk of developing asthma in a patient presenting with unexplained cough is to be estimated, attention should be paid simultaneously to a number of variables. Even mild abnormalities in several variables can be significant when they occur together.