V igorous exercise causes a marked increase in cardiac output with only a minimal increase in measureable pulmonary vascular pressures. These changes in pulmonary hemodynamics should affect lung water and solute movement. On nine occasions, we measured the effect of normoxic exercise on lung lymph flow in four sheep and two goats with chronic lymph fistulas (wt = 15-25 kg). In addition, lymph flow was also measured on five occasions in sheep during exercise at reduced barometric pressures (430 and 380 mmHg). During normobaria, the animals ran at 3-5 km/h with 0-10% elevation of the treadmill for 15 to 85 min. Exercise on average caused a 100% increase in cardiac output, a 140% increase in lung lymph flow, and a slight but significant reduction in lymph to plasma concentration ratio (l/p) for total protein and albumin (mol wt = 70,000). There was a significant linear correlation between lymph flow and cardiac output (r = 0.87, P less than 0.01). There was no change in l/p for IgG (mol wt = 150,000) or IgM (mol wt = 900,000) and no significant change in mean pulmonary arterial (Ppa) or mean left atrial (Pla) pressures. Transition from normobaria to hypobaria caused an increase in Ppa but no change in Pla, cardiac output, or lymph flow. Exercise during hypobaria caused increases in lymph flow that were qualitatively similar to changes observed during normobaric exercise: there was a 60% increase in cardiac output, a 90% increase in lymph flow, and an 11% reduction in l/p for total protein. There was no change in l/p for albumin, IgG, or IgM, and no further change in Ppa. The increased lymph flow during normoxic and hypobaric exercise is best explained by an increase in pulmonary vascular surface area for fluid and protein exchange. Our results suggest that the normal ovine lung has the potential to nearly triple the amount of perfused microvascular surface area. This speculation is relevant to the interpretation of lymph flow data from other experiments.