Subjects residing in the warm, dry climate of Tucson, Arizona showed the same relationship of productive cough to smoking as noted elsewhere, but they less regularly reported seasonal or diurnal exacerbations of their symptoms and they less regularly reported phlegm production along with chronic cough. In this locale, even nonchronic dry cough was associated with an increased frequency of other respiratory symptoms and of ventilatory impairment. In comparing characteristics of subjects with a clinical diagnosis of chronic bronchitis with characteristics of subjects who had only questionnaire-detected chronic productive cough, it was found that smoking habits were more closely related to the presence of chronic cough than to the clinical diagnosis. The clinical diagnosis was associated with other features, such as wheezing, at least as often as with productive cough. This was especially true in children, in whom a suprisingly high prevalence of both chronic cough and clinically confirmed chronic bronchitis was noted. These observations led to questions concerning the appropriateness of using a common clinical term such as chronic bronchitis to describe subjects whose only known abnormality is an affirmative answer to a direct question concerning phlegm production.