The lungs of sensitized horses were exposed to aerosolized ovalbumin. Some horses (n = 4) were given ovalbumin in 1 lung only, whereas in others (n = 7), ovalbumin or vehicle were inoculated in the cranial, ventral, and caudal regions of the caudal lung lobe. Horses were exercised 5 hours after ovalbumin exposure. Immediately before exercise, endoscopy failed to reveal any abnormality. After exercise, endoscopic examination of horses subjected to unilateral ovalbumin exposure revealed extensive blood in airways leading to the exposed lung in all horses. Blood was not observed in the airways leading to the control lung. Mean (+/- SEM) minimum volume of the exposed and control lungs was 9.5 +/- 1.5 and 5.5 +/- 1.6 L, respectively; this difference was statistically significant (P < 0.05). Bronchoscopy of horses subjected to regional ovalbumin or vehicle exposure and exercise revealed a small amount of blood-tinged fluid in the bronchi serving the regions of the lung inoculated with ovalbumin. Minimum volumes of such regions were not significantly different from one another. However, their minimum volume was significantly (P < 0.05) larger than that of vehicle-inoculated regions. Gross and histologic examination confirmed inflammation and hemorrhage in the ovalbumin-exposed, but not the control lungs or lung regions. Thus, exercise can cause blood from an injured region of lung to appear in the larger airways. Regional differences in lung structure and function do not influence the appearance of blood in the airways.