There is uncertainty about the development of airway tolerance to beta-agonists and the phenomenon of rebound bronchoconstriction on beta-agonist withdrawal. We have recently completed a study of the regular terbutaline and budesonide treatment in asthma. We report our observations on the effect of starting and stopping terbutaline treatment on morning and evening peak flows. The study was a randomized four-way, double-dummy, cross-over comparison of regular inhaled terbutaline (500-1000 microg four times daily), budesonide, combined treatment and matching placebo. Each treatment was given for 6 weeks following a 4 week single-blind placebo washout. Ipratropium was used for symptom relief. No other asthma medication was permitted during either the treatment or wash-out periods. Evaluable data were obtained from 52 subjects for both placebo and terbutaline treatment. Changes in mean morning and evening peak flows during terbutaline treatment were compared to the baseline peak flows during the last 2 weeks of the preceding washout. The peak flow changes on stopping terbutaline were also analysed. Mean morning peak flow was not significantly different during terbutaline treatment when compared to either baseline or placebo treatment. Evening peak flows were significantly higher during terbutaline treatment [mean increase 23.1 l min(-1) (95% CI = 18.8, 27.4)]. Analysis of the peak flow changes on a day-by-day basis revealed an initial increase in morning peak flows for the first 2 days of treatment of 19.2 and 13.41 min(-1) [increases of 25.0 and 17.31 min(-1) in comparison with the corresponding values during placebo (P<0.01)] followed by a return to baseline. The increase in evening peak flows was also greater for the first 2 days of treatment than for the remainder of the treatment period (P<0.01). On ceasing terbutaline treatment there was a fall in mean morning peak flow below the baseline on the following morning of 21.6 l min(-1) (P<0.05 compared to placebo). The temporary increase in morning peak flows and greater than expected rise in evening peak flows for the first 2 days of treatment suggest the development of tolerance to the bronchodilator effect of terbutaline. Similarly, the fall in morning peak flows on treatment withdrawal suggests rebound bronchoconstriction. These effects are likely to be mediated by downregulation of the beta-receptor during treatment. The clinical significance of these changes is uncertain in view of the stability of overall asthma control during terbutaline treatment, but sudden withdrawal of beta-agonist treatment could conceivably lead to a deterioration in asthma control.