1. Twelve healthy subjects (33 +/- 3 years) with a variety of fitness levels (maximal oxygen uptake (VO2, max) = 61 +/- 4 ml kg-1 min-1, range 40-80), exercised at 95 and 85% VO2, max to exhaustion (mean time = 14 +/- 3 and 31 +/- 8 min, expired ventilation (VE) over final minute of exercise = 149 +/- 9 and 126 +/- 10 l min-1). 2. Bilateral transcutaneous supramaximal phrenic nerve stimulation (BPNS) was performed before and immediately after exercise at four lung volumes, and 400 ms tetanic stimulations were performed at 10 and 20 Hz. The coefficients of variation of repeated measurements for the twitch transdiaphragm pressures (Pdi) were +/- 7-10% and for compound muscle action potentials (M wave) +/- 10-15%. 3. Following exercise at 95% of VO2, max, group mean Pdi twitch values were reduced at all lung volumes (range -8 +/- 3 to -32 +/- 5%) and tetanically stimulated Pdi values were reduced at both 10 and 20 Hz (-21 +/- 3 and -13 +/- 2%, respectively) (P = 0.001-0.047). Following exercise at 85% VO2, max, stimulated Pdi values were reduced at all lung volumes and stimulating frequencies, but only significantly so with the twitch at functional residual capacity (-15 +/- 5%). Stimulated Pdi values recovered partially by 30 min post-exercise and almost completely by an average time of 70 min. 4. The fall in stimulated Pdi values post-exercise was significantly correlated with the percentage increase in diaphragmatic work (integral of Pdi min-1) from rest to end-exercise and the relative intensity of the exercise. 5. The integral of Pdi min-1 and the integral of Po min-1 (Po, esophageal pressure) rose together from rest through the fifth to tenth minute of exercise, after which integral of Pdi min-1 plateaued even though integral of Po min-1, VE and inspiratory flow rate all continued to rise substantially until exercise terminated. Thus, the relative contribution of the diaphragm to total respiratory motor output was progressively reduced with exercise duration. 6. We conclude that significant diaphragmatic fatigue is caused by the ventilatory requirements imposed by heavy endurance exercise in healthy persons with a variety of fitness levels. The magnitude of the fatigue and the likelihood of its occurrence increases as the relative intensity of the exercise exceeds 85% of VO2, max.