Brief end-expiratory airway occlusions were performed in 22 preterm babies, 17 with and 5 without clinical apnea, and 4 full-term babies, 1 with Pierre-Robin syndrome. Airway stability was evaluated by comparing pressures measured simultaneously in the chest and nasal passages during occluded inspiratory efforts. The airway remained patent throughout all 301 trials in 20 babies during rapid-eye-movement (REM) and quiet sleep. Airway closure occurred during 31/102 trials in 6 babies (5 preterm and 1 term with Pierre-Robin syndrome), more commonly in quiet than in REM sleep. Overall and within individuals, mean closing pressures were significantly lower than the mean maximum falls in airway pressure recorded during occlusions without closure. Mixed-obstructive and obstructive apnea was significantly more frequent in babies with airway closure than in those without (5.3 +/- 4.0 vs. 0.4 +/- 0.8 episodes/h). Pauses in breathing greater than or equal to 3 s occurred during 28% of occlusions in preterm infants and 2% of occlusions in full-term babies. There was no significant difference between the mean frequency of pauses during occlusion and during the preceding control period or in the incidence of pauses in occlusions with vs. those without closure. It is concluded that the airway of most preterm and full-term babies is remarkably stable under load. Intermittent closure occurs in certain infants and may be related to airway muscle dysfunction.