We compared qualitative and quantitative aspects of perceived exertional dyspnea in patients with interstitial lung disease (ILD) and normal subjects and sought a physiological rationale for their differences. Twelve patients with ILD [forced vital capacity = 64 +/- 4 (SE) %predicted] and 12 age-matched normal subjects performed symptom-limited incremental cycle exercise tests with measurements of dyspnea intensity (Borg scale), ventilation, breathing pattern, operational lung volumes, and esophageal pressures (Pes). Qualitative descriptors of dyspnea were selected at exercise cessation. Both groups described increased "work and/or effort" and "heaviness" of breathing; only patients with ILD described "unsatisfied inspiratory effort" (75%), "increased inspiratory difficulty" (67%), and "rapid breathing" (58%) (P < 0.05 patients with ILD vs. normal subjects). Borg-O2 uptake (VO2) and Borg-ventilation slopes were significantly greater during exercise in patients with ILD (P < 0.01). At peak exercise, when dyspnea intensity and inspiratory effort (Pes-to-maximal inspiratory pressure ratio) were similar, the distinct qualitative perceptions of dyspnea in patients with ILD were attributed to differences in dynamic ventilatory mechancis, i.e., reduced inspiratory capacity, heightened Pes-to-tidal volume ratio, and tachypnea. Factors contributing to dyspnea intensity in both groups were also different: the best correlate of the Borg-VO2 slope in patients with ILD was the resting tidal volume-to-inspiratory capacity ratio (r = 0.58, P < 0.05) and in normal subjects was the slope of Pes-to-maximal inspiratory pressure ratio over VO2 (r = 0.60, P < 0. 05).