One-hundred nebulizer trials were performed in 98 adult patients with chronic airflow limitation who had remained symptomatic despite regular use of bronchodilators by metered dose or dry powder inhalation. Mean baseline FEV1 was 0.9 (SD = 0.41) 1, FVC -2.0 (0.74) 1 and PEFR 169 (77) 1 min-1. After laboratory measurements of reversibility to inhaled salbutamol (5 mg) and ipratropium bromide (0.5 mg), patients were supplied with a compressor and a peak flow meter to make twice daily measurements at home for 3 weeks. They nebulized normal saline, salbutamol (5 mg) and salbutamol (5 mg) ipratropium (0.5 mg) mixture 6-hourly, each for 1 week. A positive nebulizer trial was defined as a 15% increase in mean PEFR over a week on an active treatment compared to the week on saline. Twenty-eight patients had positive trials based on these criteria. Of these nine responded to both active treatments and 16 to the salbutamol/ipratropium mixture. Although no laboratory measurements predicted a positive domiciliary trial, the patients' subjective assessment of benefit had a 93% sensitivity and 87% negative predictive value. We conclude that an appropriate protocol for assessing the value of long-term nebulized bronchodilators is for patients to measure their PEFR during a week of nebulized saline and a week of nebulized beta-agonist/ipratropium mixture. Those with an increase of 15% in mean PEFR in the week on active treatment and who experienced subjective benefit should be supplied with a compressor. Had we conducted our 100 trials in this way we would have started 25 of our 98 patients on long-term home nebulized bronchodilators.