Whether excessive ventilatory response to exercise is related to the maldistribution of pulmonary blood flow was examined in 23 patients with chronic heart failure and nine age-matched normal subjects. With the use of technetium 99m macroaggregated albumin, the resting distribution of pulmonary blood flow was assessed by the scintigraphic counts ratio of upper to lower lung fields. The ventilatory response to exercise was assessed by the slope of the relationship between minute ventilation and carbon dioxide production during exercise. Eight patients (group A) had slope less than 33, the upper limit of the normal range, and 15 patients had slope of 33 or greater (group B). In group B pulmonary blood flow was distributed more to the upper lung, which made the counts ratio (60%) higher than in normal subjects (34%) or in patients in group A (38%). There was no significant difference in pulmonary flow distribution between normal subjects and patients in group A. In group B tidal volume did not increase during exercise as much as it did in normal subjects and in patients in group A; therefore, the respiratory pattern was rapid and shallow. Although the ratio of physiologic dead space to tidal volume fell by 20% during exercise in normal subjects and by 23% in patients in group A, it failed to decrease in patients in group B (-1%), which indicates a relative increase in dead space respiration during exercise. These data indicate that decreased lung compliance and regional ventilation-perfusion mismatch caused by pulmonary vascular and parenchymal abnormalities would play an important role in the excessive exercise ventilation in chronic heart failure.