Résumé : In exercising humans, cardiac output (CO) increases, with minor increases in pulmonary artery pressure (PAP). It is unknown if the CO is accommodated via distention of already perfused capillaries or via recruitment of nonconcomitantly perfused pulmonary capillaries. Ten subjects (9 female) performed symptom-limited exercise. Six had resting mean PAP (PAPm) <20 mmHg, and four had PAPm between 21 and 24 mmHg. The first-pass pulmonary circulatory metabolism of [3H]benzoyl-Phe-Ala-Pro (BPAP) was measured at rest and at peak exercise, and functional capillary surface area (FCSA) was calculated. Data are means ± SD. Mean pulmonary arterial pressure rose from 18.8 ± 3.3 SD mmHg to 28.5 ± 4.6 SD mmHg, CO from 6.4 ± 1.6 to 13.4 ± 2.9 L/min, and pulmonary artery wedge pressure from 14 ± 3.3 to 19.5 ± 5 mmHg (all P ≤ 0.001). Percent BPAP metabolism fell from 74.7 ± 0.1% to 67.1 ± 0.1%, and FCSA/body surface area (BSA) rose from 2,939 ± 640 to 5,018 ± 1,032 mL·min-1·m-2 (all P < 0.001). In nine subjects, the FCSA/BSA-to-CO relationship suggested principally capillary recruitment and not distention. In subject 10, a marathon runner, resting CO and FCSA/BSA were high, and increases with exercise suggested distention. Exercising humans demonstrate pulmonary capillary recruitment and distention. At moderate resting CO, increasing blood flow causes principally recruitment while, based on one subject, when exercise begins at high CO, further increases appear to cause distention. Our findings clarify an important physiologic question. The technique may provide a means for further understanding exercise physiology, its limitation in pulmonary hypertension, and responses to therapy.