The antitussive effect of several antitussive agents has been objectively evaluated in patients with chronic stable cough due to bronchial carcinoma, pulmonary tuberculosis or chronic obstructive lung disease. The patients received the active antitussive drugs or placebo in a double-blind, randomized crossover design. The preparations were administered at 10 p.m. and 2 a.m. on 7 consecutive nights and no antitussive was given for the following 20 hours. Cough frequency and intensity were recorded from 10 p.m. until 6 a.m. The active medications were noscapine (30 mg), dextromethorphan (20 mg), dihydrocodeine (30 mg) and codeine (20, 30 and 60 mg) at 10 p.m. and 2 a.m. Cough frequency and intensity were objectively assessed with a pressure transducer placed over the trachea and recorded on a chartrecorder. Statistical analysis was performed with analysis of variance and multiple range testing. Noscapine, dextromethorphan, dihydrocodeine and codeine (60 mg) significantly (p less than 0.001) reduced the cough frequency compared to placebo. They also produced a greater reduction of cough intensity than placebo, codeine (20 mg) and codeine (30 mg) (p less than 0.001). The duration of action of low-dose codeine (6 hours) was unsatisfactory. Subjective preference for dextromethorphan indicates a psychotropic central nervous action of this drug not assessed by the measuring device. Noscapine was equally well tolerated but more neutral psychologically.