We studied the effects of steady state increases in left atrial pressure on the albumin-excluded volume in the lungs of mongrel dogs. We estimated the tissue blood volume using51Cr-labeled red cells, the extracellular space using99mTc-DTPA (diethylene triamine pentacetic acid), and the albumin space using125I-labeled human serum albumin. An afferent lymphatic to the left tracheobron-chial node was cannulated for measurement of lymph flow, total protein and albumin concentration. Total extravascular water (Qw), extravascular "Tc-DTPA space, extravascular albumin content, and albumin space were calculated in lung tissue samples taken during a baseline period and during steady states, following increases in pulmonary capillary pressure. Lymph flow increased by 0.20-fold, and Qw by 0.05 g/g blood free dry weight for each cm H2O increase in capillary pressure over the range of capillary pressures used in this study. Using an extravascular albumin space based on the concentration of albumin in lymph, an excluded albumin volume of 36% of the extravascular99mTc-DTPA space was calculated for normally hydrated lungs. This excluded volume fraction decreased to approximatley 10% of the extravascular99Tc-DTPA space at capillary pressures above 30 cm H2O. Oncotic buffering increased the plasma to lymph colloid osmotic pressure gradient, and buffered approximately 29% of the increase in pulmonary capillary pressure. Excluded volume had a simple physicochemical relation-ship to tissue hydration. It also served as a "safety factor" against pulmonary edema because there was a greater increase in available volume than total interstitial volume during edema formation. Oncotic buffering then could occur with a smaller increase in interstitial fluid volume than would otherwise be possible.Circ Res 47: 866-875, 1980