Insights into muscle energetics during exercise (e.g., muscular efficiency) are often inferred from measurements of pulmonary gas exchange. This procedure presupposes that changes of pulmonary O2 (VO2) associated with increases of external work reflect accurately the increased muscle VO2. The present investigation addressed this issue directly by making simultaneous determinations of pulmonary and leg VO2 over a range of work rates calculated to elicit 20-90% of maximum VO2 on the basis of prior incremental (25 or 30 W/min) cycle ergometry. VO2 for both legs was calculated as the product of twice one-leg blood flow (constant-infusion thermodilution) and arteriovenous O2 content difference across the leg. Measurements were made 3-5 min after each work rate imposition to avoid incorporation of the VO2 slow component above the lactate threshold. For all 17 subjects, the slope of pulmonary VO2 (9.9 +/- 0.2 ml O2.W-1.min-1) was not different (P greater than 0.05) from that for leg VO2 (9.2 +/- 0.6 ml O2.W-1.min-1). Estimation of "delta" efficiency (i.e., delta work accomplished divided by delta energy expended, calculated from slope of VO2 vs. work rate and a caloric equivalent for O2 of 4.985 cal/ml) using pulmonary VO2 measurements (29.1 +/- 0.6%) was likewise not significantly different (P greater than 0.05) from that made using leg VO2 measurements (33.7 +/- 2.4%). These data suggest that the net VO2 cost of metabolic "support" processes outside the exercising legs changes little over a relatively broad range of exercise intensities. Thus, under the conditions of this investigation, changes of VO2 measured from expired gas reflected closely those occurring within the exercising legs.