Addition of HCO3- to the serosal side (S) of the isolated turtle bladder results in a HCO3- flow from S to the mucosal side (M) which markedly reduces the net rate of acid secretion. To characterize the driving forces for this downhill HCO3- flow, the effects of metabolic inhibitors and substrates were examined. In short-circuited bladders with the M pH lowered to the point of zero net H+ secretion, the rate of HCO3- entry into M in response to a 20-mM HCO3- gradient was measured by pH stat titration. Deoxygenation reduced the HCO3- flux from 1.24 plus or minus 0.1 mum/h/8 cm2 (SEM) to 0.50 plus or minus 0.1 muM/h with glucose (2 times 10-3 M) AND FROM 1.32 PLUS OR MINUS TO 0.47 PLUS OR MINUS 0.1 MUM/h without glucose. A similar reduction (61 per cent) was observed in the presence of 1 per cent C92. Dinitrophenol (10-4 M), cyanide (10-3 M), and deoxyglucose (10-2 M) inhibited the HCO3- flux by 39 per cent, 37 per cent, and 38 per cent, respectively. The combination of any of these inhibitors with N2 caused the same inhibition as N2 alone. In bladders depleted of substrate, pyruvate (5 times 10-3 M) increased the HCO3- flux from 0.36 plus or minus 0.05 to 0.58 plus or minus 0.01 muM/h (P smaller than 0.005); the increment was abolished by deoxygenation. The results indicate that the bulk of the downhill HCO3- flow in this system is dependent on metabolic energy derived primarily from oxidative sources, and that this energy-dependent flow approximates the electroneutral component of HCO3- secretion that is coupled to Cl- absorption.