We have shown previously that inspired CO2 (3-5%) improves ventilation-perfusion (Va/Q) matching but with the consequence of mild arterial hypercapnia and respiratory acidosis. We hypothesized that adding CO2 only late in inspiration to limit its effects to the conducting airways would enhance Va/Q matching and improve oxygenation without arterial hypercapnia. CO2 was added in the latter half of inspiration in a volume aimed to reach a concentration of 5% in the conducting airways throughout the respiratory cycle. Ten mixed-breed dogs were anesthetized and, in a randomized order, ventilated with room air, 5% CO2 throughout inspiration, and CO2 added only to the latter half of inspiration. The multiple inert-gas elimination technique was used to assess Va/Q heterogeneity. Late-inspired CO2 produced only very small changes in arterial pH (7.38 vs. 7.40) and arterial CO2 (40.6 vs. 39.4 Torr). Compared with baseline, late-inspired CO2 significantly improved arterial oxygenation (97.5 vs. 94.2 Torr), decreased the alveolar-arterial Po2 difference (10.4 vs. 15.7 Torr) and decreased the multiple inert-gas elimination technique-derived arterial-alveolar inert gas area difference, a global measurement of Va/Q heterogeneity (0.36 vs. 0.22). These changes were equal to those with 5% CO2 throughout inspiration (arterial Po2, 102.5 Torr; alveolar-arterial Po2 difference, 10.1 Torr; and arterial-alveolar inert gas area difference, 0.21). In conclusion, we have established that the majority of the improvement in gas exchange efficiency with inspired CO2 can be achieved by limiting its application to the conducting airways and does not require systemic acidosis.