The relationship between different grades of laryngeal function, as assessed by endoscopy at rest, and the measurements of indices of gas exchange and exercise capacity was assessed during a standardised treadmill exercise test in 149 horses. Horses with abnormalities other than idiopathic laryngeal hemiplegia (ILH) were excluded from the study and laryngeal function was graded according to an established system. There were no significant differences in age, weight, maximum oxygen uptake, maximum carbon dioxide production, maximum respiratory exchange ratio, maximum oxygen pulse and run time between the grades. Blood lactate concentration at 10 m/s was greater (P<0.01) in horses with grade 5 laryngeal function than other grades. Minimum PaO2 (P<0.001) and SaO(2) (P<0.01) were lower and maximum PaCO2 (P<0.001), higher in horses with grades 4 and 5 laryngeal function than other grades. Horses with grade 4 function had a lower minimum CaO2 (P<0.01) than horses with other grades. Minimum PAO(2) decreased from grades I and 2 to grades 4 and 5 (P<0.05). The minimum alveolar ventilation was lower (P<0.05) in horses with grades 4 and 5 laryngeal function compared to other grades. The results of this study indicate that endoscopic assessment of laryngeal function at rest, using a simple grading system, provides an indication of dynamic changes in ventilation and the effects on blood gases during exercise. From the data, we suggest that horses that have some movement of the left arytenoid cartilage but are unable to achieve full abduction have similar ventilatory effects and blood gas responses during maximal exercise to those with complete paralysis, Some horses with grade 3 laryngeal function had blood gas results similar to those of horses with grades 4 and 5 laryngeal function, indicating that discrepancies may occur between the resting assessment and laryngeal function during strenuous exercise.