The primary secretion formed in various exocrine glands has a [K+] 2-5 times that of plasma. In this study we measured the transepithelial flux of 36Cl-, 22Na+ and 42K+ across the frog skin and applied the single-channel patch-clamp technique to the apical membrane of frog skin gland acini to investigate the pathway taken by K+ secreted by the glands. Transepithelial K+ secretion was active and was driven by a larger force than the secretion of Na+. When driving Na+ through the epithelium by clamping the transepithelial potential to 100 mV (apical solution reference), blockers of cellular secretion (apical 5-nitro-2-(3-phenylpropylamino)benzoate or basolateral quinine or furosemide) decreased K+ secretion but left Na+ secretion unaffected. We conclude that K+ follows a transcellular pathway across the epithelium. Patch-clamp analysis of the apical membrane of microdissected gland acini revealed a population of voltage- and calcium-activated K+ channels of the maxi K+ type. In cell-attached patches these channels were activated by membrane potential depolarisation or exposure to prostaglandin E2 and had a permeability of 3.6 +/- 0.3 x 10(-13) cm3 s-1, giving a calculated conductance of 170 pS with 125 mM K+ on both sides of the membrane. In inside-out patches the channels were activated by increasing intracellular [Ca2+] from 10(-7) to 10(-6) M and were blocked by Ba2+ added to the cytoplasmic side. Exposure of inside-out patches containing the maxi K+ channel to ATP on the inside activated cystic fibrosis transmembrane conductance regulator (CFTR) Cl- channels, confirming that both channels are co-localised to the apical membrane. We interpret these findings in terms of a model where transepithelial NaCl secretion can be supported in part by an apical K+ conductance.