Our purpose was to determine the reliability of the currently used measurements of pulmonary vascular hydrostatic and oncotic pressures in monitoring pulmonary water after a major burn. The flow of pulmonary lymph, a reliable indicator of the rate of filtration of pulmonary transvascular fluid (Qf), was measured before and for 72 hours after a third-degree burn over 40 to 55 percent of the total body surface in sheep. Changes in Qf were correlated with measured changes in pulmonary microvascular (Pmv), plasma colloid osmotic pressure (pi p), and interstitial colloid osmotic pressure (pi i), as measured in the pulmonary lymph. The periods from 0 to 24 hours (resuscitation), from 24 to 48 hours (early mobilization of fluid), and from 48 to 72 hours (early recovery) were compared with baseline. Regression equations and correlation coefficients for Qf vs Pmv, Qf vs Pmv minus pi p, and Qf vs Pmv minus (pi p -ph i), the oncotic gradient, were calculated. There were significant differences in slopes between the periods of time, with all three comparisons indicating that the response of the microcirculation to the same changes in pressure was different in each of the periods. The correlation between Qf and the comparisons of pressures was clearly best in the period from 48 to 72 hours. The comparison of Pmv minus pi p was not better than Pmv alone, while the comparison of Pmv minus the gradient (pi p -- pi i) was significantly better than either of the other comparisons in predicting Qf.