Twenty-one asthmatics underwent treadmill exercise to exhaustion at a workload that elicited ~90% of each subjects maximal O₂ uptake [(VO_2max) EX1]. Following EX1, 12 subjects experienced significant exercise-induced bronchospasm [(EIB⁺), % decrease in forced expiratory volume in 1.0 s = -24.0 ± 11.5%; pulmonary resistance (RLi) at rest vs. post-exercise = 3.2 ± 1.5 vs. 8.1 ± 4.5 cmH₂O^.L^-1.sec^-1] and nine did not (EIB-). The alveolar-to-arterial PO₂ difference (AaDO₂) was widened from rest (9.1 ± 6.7 Torr) to 23.1 ± 10.4 and 18.1 ± 9.1 Torr at 35 min after EX1 in subjects with and without EIB, respectively (P < 0.05). Arterial PO₂ (PaO₂) was reduced in both groups during recovery (EIB⁺, -16.0 ± -13.0 Torr vs. baseline; EIB^-, -11.0 ± 9.4 Torr vs. baseline, P 0.05). Forty minutes following EX1, a second exercise bout was completed at VO_2max (EX2). During EX2, RLi decreased to baseline levels in the EIB⁺ group and the AaDO₂ and PaO₂ returned to match the values seen during EX1 in both groups. Sputum histamine (34.6 ± 25.9 vs. 61.2 ± 42.0 ng^.ml^-1, pre vs. post-exercise) and urinary 9, 11- prostaglandin F₂ (74.5 ± 38.6 vs. 164.6 ± 84.2 ng^.mmol creatinine^-1, pre vs. post-exercise) were increased after exercise only in the EIB⁺ group (P < 0.05), and post-exercise sputum histamine was significantly correlated with the exercise PaO₂ and AaDO₂ in the EIB⁺ subjects. Thus, exercise causes gas exchange impairment during the post-exercise period in asthmatics independent of decreases in forced expiratory flowrates after the exercise; however, a subsequent exercise bout normalizes this impairment secondary in part to a fast acting, robust exerciseinduced bronchodilatory response.