The effect of preparative trauma on lung microvascular permeability during lung lymph fistula preparation in sheep was investigated. Two groups of anesthetized sheep were used. In group I (controls: n = 3) the lymph fistula was prepared 3 to 4 days before the experiment. In group II (experimental group: n = 13): left atrial and lymph catheters were introduced on the day of the experiment. All animals had catheters in the aorta and pulmonary artery. The stable prostanoid metabolites thromboxane B2 (TxB2) and 6-keto-PGF1 alpha (6-keto) were measured in lymph. In group I systemic (Psa) and pulmonary pressures (Ppa) remained constant. Leukocytes and platelets were unchanged as was pulmonary lymph flow (QL) and the lymph to plasma total protein concentration ratio (L/P). In group II Psa and Ppa remained constant. All sheep developed a temporary leukopenia and thrombocytopenia. TxB2 and 6-keto were low and stable during anesthesia in group I but were elevated immediately following surgery in group II but values normalized during a 4-hr observation period. QL during the first half hour was 1.9 +/- 0.3 ml/30 min and increased over 5 hr to 3.1 +/- 0.7 ml/30 min corresponding to 60 +/- 15% over baseline (P less than 0.01). L/P did not change. These changes could be due either to changing permeability or surface area. These effects of preparative trauma indicated that the model was not in a steady state. This observation should be kept in mind when evaluating studies using the lung lymph fistula in acute experiments.