Stripped, proximal bullfrog duodenum was mounted in an Ussing chamber between HCO3--buffered nutrient (serosal) and unbuffered secretory (luminal) solutions. This preparation showed stable electrical parameters and caused alkalinization of the secretory solution at a rate of 0.95 +/- 0.03 mueq.cm-2.h-1 (mean +/- SE; n = 100). Anoxia and 2,4-dinitrophenol each reduced alkalinization by 50-60%, but acetazolamide (5 X 10(-4)M) had no effect. Removal of nutrient HCO3- and CO2 reduced alkalinization by over 90%, whereas increasing nutrient [HCO3-] at constant partial pressure of CO2 (PCO2) or increasing nutrient PCO2 at constant [HCO3-] each caused saturable increases in alkalinization, despite opposite effects on nutrient pH. Dibutyryl adenosine 3',5'-cyclic monophosphoric acid, but not dibutyryl guanosine 3',5'-cyclic monophosphoric acid, increased luminal alkalinization to 167 +/- 21% of control. Removal of nutrient, but not secretory, Na+ reduced alkalinization by 74%. Changes in the rate of alkalinization were accompanied by corresponding changes in potential difference and short-circuit current. Removal of Cl- or nutrient K+ or addition of histamine, thiocyanate, or catecholamines had no effect on electrical or secretory characteristics. We conclude that a) the amphibian duodenum transports alkali from nutrient to secretory solutions by both active and passive processes, b) there is a small secretion of endogenous HCO3-, c) alkaline secretion is electrogenic, d) Cl- does not contribute to the short-circuit current, e) alkaline secretion is partially dependent on nutrient Na+ that acts in a facilitatory, not cotransport, role, f) there is no Cl--HCO3- exchange, and g) alkaline secretion is independent of nutrient pH.