The relative contribution of the pulmonary and bronchial circulatory systems to heat and water exchange in normal lungs was evaluated in 20 isolated, in situ perfused dog lungs and in four patients undergoing elective cardiopulmonary bypass. In isolated dog lungs, if the pulmonary artery was perfused at a nominal flow rate (0.5 l/min), bronchial artery perfusion (up to 70 ml/min) did not significantly affect the expired gas temperature. When the lungs were not perfused through either system, 8 min of ventilation with cool, dry gas decreased the temperature of the expired gas by 6.2 ± 1.4°C. Selective perfusion of bronchial arteries at 68 ± 10 mmHg resulted in a mean flow rate of 28 ± 16 ml/min and increased the average temperature of the expired gas by 0.6°C. An increase in the rate of bronchial arterial perfusion to 55 ± 14 ml/min increased the average temperature of the expired gas by 1.3°C. The time constant for equilibration of tritiated water between the perfusate and the lung parenchyma was 130 ± 33 min for pulmonary arterial perfusion and 35 ± 13 min for combined bronchial and pulmonary perfusion, which indicated that filtration of water from high-pressure bronchial vessels facilitated water exchange in the lung interstitium. The rate of tracer equilibration was similar between the perfusate and gas in both variants of perfusion, but the ratios of tracer gas to perfusate were different (0.42 ± 0.06 for pulmonary, 0.98 ± 0.07 for combined), which indicates that bronchial vessels contribute mainly to the hydration of the bronchial mucosa. In humans, the bronchial blood flow was capable of maintaining heat supply after the initiation of cardiopulmonary bypass. Before bypass, when both pulmonary and bronchial blood flow were present, the mean time constant of the temperature decay after a switch to ventilation with cool, dry gas was 35 ± 12 s. The average temperature difference between the blood and expired gas was 2.4 ± 0.50°C. After 5 min of dry gas ventilation, the temperature difference between the expired gas and initial blood temperature decreased an average of 3.8 ± 0.06°C (P < 0.05). The time constant of temperature decay increased to 56 ± 14 s (P < 0.05). We conclude that bronchial perfusion has a less important role in the temperature balance of the respiratory tract compared with pulmonary arterial perfusion because heat flux is "flow limited" but is important in providing water for hydration of the mucosal surface and interstitial compartments of peribronchial tissues.