The ventilatory response to hypoxia during sleep in neonates is biphasic with an initial increase followed by a decrease in ventilation. More recently it has been shown that i) this response consists only of a sustained decrease in ventilation in the very low birth weight preterm infant, andii) the response is also biphasic in the awake adult subject. However, neonates have been studied for 5 minutes only, whereas adult subjects have been studied for 20 minutes; also adults have not been studied during sleep. To make the groups comparable, we studied infants and adult subjects for 20 min. during quiet sleep. The hypotheses were that ventilation would not decrease further after 5 minutes of hypoxia in neonates, and that adults might decrease their ventilation to control levels by the end of 20 minutes. Subjects were 5 preterm infants [birthweight 1498 ± 197 g(mean±SE), study weight 1660±152 g, gestational age 31 ± 1 wk, postnatal age 25±3 d]; 6 term infants [birthweight 3020±250 g, study weight 2910±270 g, gestational age 37±1 wk, postnatal age 5±2 d]; and 5 adults [weight 67±6 kg, age 31±4 yrs]. Ventilation was measured using a flow-through system in all subjects. The experimental procedure consisted in giving the subjects 21% O2 for 2 min., decreased FiO2 to 80% saturation for 20 min., followed by 2 min. of recovery. The percent change in ventilation was as follows: 1 min [preterm: -0.8±8%; term 12±7; adult 67±17; p < 0.001, ANOVA], 5 min. [preterm: -1±9; term -10±66; adult 84±17; p < 0.001], 10 minutes [preterm: 10±10; term-8±4; adult 86±14; p < 0.001], 20 min. [preterm:-8±-4; term -15±7; adult 40±21; p < 0.002] recovery[preterm: -14±10; term -1±8; adult -5±7; p=0.55]. The changes were primarily frequency dependent in neonates and tidal volume related in adults. The findings suggest: 1) preterm infants do not hyperventilate initially but term infants do in response to hypoxia; 2) both groups do not significantly decrease their ventilation after 5 min of hypoxia; 3) sleeping adult subjects respond biphasically to hypoxia, but ventilation does not return to baseline. We speculate that hypoxia administered for periods longer than 20 min would not induce further decrease in ventilation in any of the subjects studied. Supported by the Canadian Lung Association and Children's Hospital of Winnipeg Research Foundation.