The effects of selective restriction of rib cage (Res,rc) and abdominal wall (Res,ab) movements on endurance of short-term constant-load heavy exercise and on diaphragmatic function during such exercise were examined in five normal young men. An inelastic surgical corset was used to achieve Res,rc and Res,ab. Subjects exercised on a cycle ergometer at 80% of their maximum power output to exhaustion on three occasions: with Res,rc, with Res,ab, and without restriction of chest wall movements (control). Transdiaphragmatic (Pdi), esophageal, and gastric pressures were measured. Electromyogram of the diaphragm was recorded by an esophageal electrode, and the ratio of the power content of a high-frequency to low-frequency band (H/L ratio) was measured. In addition, maximum Pdi (Pdimax) pre- and immediately postexercise was recorded. Res,rc was associated with a shorter endurance time, a progressive decline of the H/L ratio, and a significant reduction of Pdimax postexercise, whereas no such changes were found with Res,ab. We conclude that diaphragmatic function was well defended with abdominal wall loading, whereas limitation of rib cage expansion reduced diaphragmatic endurance during exercise. The diaphragmatic tension-time index (TTdi) in exercise was always less than the critical value of 0.15 found by Bellemare and Grassino (J. Appl. Physiol. 53: 1190–1195, 1982) when subjects inspired against large resistive loads at normal minute ventilations. We suggest that the higher inspiratory flow rate (P less than 0.05) and breathing frequency (P less than 0.05) account for the occurrence of diaphragmatic fatigue in exercise with Res,rc when the TTdi was 0.06 +/- 0.02.