In vitro guinea pig tracheal contraction to 5-hydroxytryptamine (5-HT) and to KCl increased significantly with time, an effect specific to the trachea because contraction of the rat aorta to KCl changed only modestly with time. The possibility that with time, contractile agonists could release endogenous contractile substances was considered and ruled out. Neither the neurotoxin tetrodotoxin, the muscarinic antagonist atropine, the thromboxane synthase inhibitor (E)-3-[4-(1-imidazolylmethyl)phenyl]-2-propanoic acid hydrochloride monohydrate (OKY 046), the leukotriene D4 and E4 antagonists 7-(3-hydroxy-2-prophylphenoxy[2-hydroxypropoxyl]-4- oxo-8-propyl-4H-1-benzopyran-2-carboxylate) (FPL55712) and 4-(5-cyclopentyloxycarbonylamino-1-methylindol-3-ylmythyl )-3-methoxy-N-o-toylsulfonyl benzamide (ICI 204,219), or the lipoxygenase inhibitor N-(3-phenoxycinnamyl)-acetohydroxamic acid (BWA4C) markedly affected the tracheal increase in contraction to KCl over time. The possibility that bronchodilatory substances were released and opposed contractility at early times, but not later was also considered. However, adenosine was not likely to be involved because 8-(p-sulfophenyl)theophylline, an inhibitor of adenosine, did not modify agonist-induced tracheal contractility. Propranolol (10(-6) M) also had no effect on tracheal improvement, suggesting that beta receptor activation of relaxation was not involved. Interestingly, incubation of trachea with the cyclooxygenase inhibitors indomethacin (10(-5) and 10(-6) M) or piroxicam (10(-6) M) increased the initial tracheal contraction to KCl, permitting the trachea to reach maximal force sooner, thereby minimizing the apparent improvement in contractility.(ABSTRACT TRUNCATED AT 250 WORDS)