Cough is an important symptom of many respiratory disorders. We determined the frequency and diurnal variation of cough in normal subjects and in patients with asthma or with persistent cough of unknown cause. We used a portable, solid-state, multiple-channel recorder to record cough sounds over a 24 h period. The audio-signal was recorded from a unidirectional microphone strapped over the chest wall, and electromyographic (EMG) signals from the lower respiratory muscles were simultaneously registered with surface electrodes. The recorded digital data were examined on an IBM-compatible computer, and the typical signals induced by cough (as assessed by voluntary or experimentally-induced cough) were counted. In 12 normal subjects, only 0-16 coughs were recorded over 24 h. In 21 stable asthmatics with a history of chronic cough ("asthma") the median number was 282 (ranges: 45-1,577), and in 14 patients with the predominant symptom of daily dry coughs ("chronic coughers") the median number was 794 (64-3,639). In both groups of patients, there was a diurnal variation of coughs, such that the least numbers occurred between 2 and 5 a.m. (< 3% of total). In the asthma group, there was no significant correlation between forced expiratory volume in one second (FEV1) (% predicted) or diurnal variation of peak expiratory flow and cough frequency. In the chronic coughers, there was a significant correlation between daytime cough numbers and daytime cough symptoms scores but not for the night-time values. Our data show that cough frequency is not determined by the severity of asthma in relatively stable asthmatics on inhaled steroids, and is reduced during sleep in both asthmatics and chronic cough patients. This portable cough recorder may be useful in the assessment of drug therapy for chronic cough.