Hyperpnea with exercise or hypercapnia causes phasic contraction of abdominal muscles, potentially lengthening the diaphragm at end expiration and unloading it during inspiration. Muscle efficiency in vitro varies with load, fiber length, and precontraction stretch. To examine whether these properties of muscle contractility determine diaphragm efficiency (Eff(di)) in vivo, we measured Eff(di) in six healthy adults breathing air and during progressive hypercapnia at three levels of end-tidal Pco(2) with mean values of 48 (SD 2), 55 (SD 2), and 61 (SD 1) Torr. Eff(di) was estimated as the ratio of diaphragm power (Wdi) [the product of mean inspiratory transdiaphragmatic pressure, diaphragm volume change (DeltaVdi) measured fluoroscopically, and 1/inspiratory duration (Ti(-1))] to activation [root mean square values of inspiratory diaphragm electromyogram (RMS(di)) measured from esophageal electrodes]. At maximum hypercapnea relative to breathing air, 1) gastric pressure and diaphragm length at end expiration (Pg(ee) and Ldi(ee), respectively) increased 1.4 (SD 0.2) and 1.13 (SD 0.08) times, (P < 0.01 for both); 2) inspiratory change (Delta) in Pg decreased from 4.5 (SD 2.2) to -7.7 (SD 3.8) cmH(2)O (P < 0.001); 3) DeltaVdi.Ti(-1), Wdi, RMS(di), and Eff(di) increased 2.7 (SD 0.6), 4.9 (SD 1.8), 2.6 (SD 0.9), and 1.8 (SD 0.3) times, respectively (P < 0.01 for all); and 4) net and inspiratory Wdi were not different (P = 0.4). Eff(di) was predicted from Ldi(ee) (P < 0.001), Pg(ee) (P < 0.001), DeltaPg.Ti(-1) (P = 0.03), and DeltaPg (P = 0.04) (r(2) = 0.52) (multivariate regression analysis). We conclude that, with hypercapnic hyperpnea, 1) approximately 47% of the maximum increase of Wdi was attributable to increased Eff(di); 2) Eff(di) increased due to preinspiratory lengthening and inspiratory unloading of the diaphragm, consistent with muscle behavior in vitro; 3) passive recoil of the diaphragm did not contribute to inspiratory Wdi or Eff(di); and 4) phasic abdominal muscle activity with hyperpnea reduces diaphragm energy consumption.