We examined the recent proposition (Circ. Res. 57: 903-905, 1985) that the interstitium-to-luminal transport of albumin is an active phenomenon. Studies were made using cultured bovine and sheep pulmonary-artery endothelial cells. The transendothelial 125I-albumin flux from the luminal-to-abluminal side was compared with the flux from the abluminal-to-luminal side. The endothelial cells were grown to confluence on gelatinized-polycarbonated filters separating the abluminal from the luminal compartments. The albumin concentration in each compartment was 1 g/100 ml to equalize the oncotic pressure gradients. The effect of hydrostatic pressure was eliminated by maintaining equal fluid levels in both compartments. The transendothelial albumin flux across the monolayer was measured by adding the 125I-albumin tracer either on the luminal or the abluminal side. A double-isotope method was also used to study bidirectional transendothelial flux of albumin at the same time for the same cultured endothelium. The results indicated that albumin flux from the luminal-to-abluminal side was equal to the flux from the abluminal-to-luminal side. Both bovine and sheep pulmonary artery endothelial cells in culture behave symmetrically for albumin, suggesting that albumin is not actively transported from the interstitium to the lumen.