We have studied the effects of acute respiratory alkalosis (ARALK, hyperventilation) and acidosis (ARA, 8% CO2), chronic respiratory acidosis (CRA; 10% CO2 for 7–10 days), and subsequent recovery from CRA breathing air on loop of Henle (LOH) net bicarbonate flux ([Formula: see text]) by in vivo tubule microperfusion in anesthetized rats. In ARALK blood, pH increased to 7.6, and blood bicarbonate concentration ([[Formula: see text]]) decreased from 29 to 22 mM. Fractional urinary bicarbonate excretion ([Formula: see text]) increased threefold, but LOH[Formula: see text]was unchanged. In ARA, blood pH fell to 7.2, and blood [[Formula: see text]] rose from 28 to 34 mM; [Formula: see text] was reduced to <0.1%, but LOH[Formula: see text]was unaltered. In CRA, blood pH fell to 7.2, and blood [[Formula: see text]] increased to >50 mM, whereas[Formula: see text]decreased to <0.1%.[Formula: see text]was reduced by ∼30%. Bicarbonaturia occurred when CRA rats breathed air, yet LOH[Formula: see text]increased (by 30%) to normal. These results suggest that LOH[Formula: see text]is affected by the blood-to-tubule lumen [[Formula: see text]] gradient and[Formula: see text] backflux. When the usual perfusing solution at 20 nl/min was made[Formula: see text] free, mean[Formula: see text]was −34.5 ± 4.4 pmol/min compared with 210 ± 28.1 pmol/min plus [Formula: see text]. When a low-NaCl perfusate (to minimize net fluid absorption) containing mannitol and acetazolamide (2 × 10−4 M, to abolish H+-dependent[Formula: see text]) was used,[Formula: see text]was −112.8 ± 5.6 pmol/min. Comparable values for[Formula: see text]at 10 nl/min were −35.9 ± 5.8 and −72.5 ± 8.8 pmol/min, respectively. These data indicate significant backflux of[Formula: see text] along the LOH, which depends on the blood-to-lumen [[Formula: see text]] gradient; in addition to any underlying changes in active acid-base transport mechanisms, [Formula: see text]permeability and backflux are important determinants of LOH[Formula: see text]in vivo.