In 21 studies on 15 infants an additional dead space tube produced a significant rise in end-tidal PCO2 and fall in end-tidal PO2, associated with a rise in minute ventilation (228 +/- 77 mL/kg/min at zero, 348 +/- 85 mL/kg/min at one, and 437 +/- 128 mL/kg/min at two anatomical dead spaces). The differences between end-inspiratory and end-expiratory PCO2 and PO2 did not change significantly, suggesting an increase in dead space, but not in alveolar ventilation. In a further 9 babies the rise in ventilation was unchanged when measurements were repeated in 30% oxygen (361 +/- 65 vs. 340 +/- 54 mL/kg/min at one anatomical dead space). Studies on 8 babies, with the added tube ventilated by a fan, showed that a mean 28% of the rise in minute ventilation was due to increased resistance. Although the response to tube breathing in neonates is complex, carbon dioxide appears to be the major factor producing increased ventilation.