1. In guinea-pig tracheal preparations precontracted with 1 mumol l-1 carbachol, formoterol, procaterol, fenoterol, salmefamol, salbutamol and terbutaline (in that order of potency) caused a concentration-dependent and almost complete, relaxation. However, under these conditions, the maximum relaxation by salmeterol was approximately 30% of the maximum attainable relaxation. 2. We have therefore explored the ability of salmeterol to inhibit the relaxant response to beta 2-adrenoceptor agonists of different chemical structure and relatively higher efficacy in smooth muscle preparations from guinea-pig trachea and human bronchus. 3. With 1 mumol l-1 salmeterol in the organ bath, the concentration-effect curves for the other agonists were shifted to the right in a variable way by 1.8-2.8 log units, fenoterol and salbutamol being the extremes. 4. When 20 mumol l-1 sulfonterol, another low efficacy beta 2-adrenoceptor agonist, was substituted for salmeterol, the difference in the magnitude of the rightward shift between fenoterol and salbutamol was eliminated. 5. In the human bronchus, formoterol and terbutaline had a higher apparent efficacy than salmeterol. With 1 mumol l-1 salmeterol in the organ bath, the concentration-effect curve for formoterol was shifted 2.7 log units to the right. 6. Salmeterol inhibits, competitively, relaxant responses to beta 2-adrenoceptor agonists with higher efficacy. The degree of inhibition seems to be dependent on the agonist used. This contrasts with results obtained with sulfonterol and suggests that salmeterol interacts with the beta 2-adrenoceptor in a complex way.