This study investigates the mechanism whereby the antioxidant 2-t-butyl-4-methoxyphenol (BHA) relaxes guinea pig gastric fundus smooth muscle. In circular smooth muscle strips, 10 microM cyclopiazonic acid, a specific inhibitor of sarcoplasmic reticulum Ca2+-ATPase, induced a prolonged rise in tension which depended on the presence of extracellular Ca2+. BHA (pIC50 = 5.83), sodium nitroprusside (6.85), isoproterenol (7.69) and nifedipine (8.02), but not 2,6-di-t-butyl-4-methoxyphenol (DTBHA) (up to 30 microM), relaxed muscle strips contracted with cyclopiazonic acid. Methyl-1,4-dihydro-2,6-dimethyl-3-nitro-4-(2-trifluoromethylphenyl)-pyri dine-5-carboxylate (Bay K 8644) (1 microM) antagonised the nifedipine- but not the BHA-induced relaxation. Nifedipine and isoproterenol (10 microM) caused a decrease in spontaneous tone, but did not counteract the subsequent rise in tension elicited by 10 microM cyclopiazonic acid. Conversely, 100 microM BHA and 100 microM sodium nitroprusside not only significantly reduced spontaneous tone but also markedly impaired the response of the muscles to cyclopiazonic acid. DTBHA failed to show either effect. When added to preparations completely relaxed by 100 microM BHA, 10 mM tetraethylammonium still elicited nifedipine-sensitive tonic and phasic contractions in the presence or absence of 10 microM cyclopiazonic acid. BHA and DTBHA inhibited, in a concentration-dependent manner, the Ca2+-promoted contraction of strips depolarised by 10 mM tetraethylammonium. The BHA antagonism showed a non-competitive profile while that of DTBHA was competitive. In muscle strips at rest, 10 microM BHA caused a significant increase in tissue cAMP concentration, leaving cGMP unmodified. To conclude, the myorelaxant action of BHA on gastric fundus smooth muscle appears to be mediated partly by an increase in cAMP levels and partly by inhibition of Ca2+ influx from the extracellular space.