Review of the case notes of 373 admissions to a neonatal intensive care unit during one year revealed 30 babies with pulmonary air leak. These babies had the uniform characteristics of being between 26 and 34 weeks' gestation and receiving intermittent positive pressure ventilation during the course of respiratory distress. A control group was formed of the remaining 27 infants with these characteristics who did not sustain an air leak. No significant difference was found between the groups in any clinical factor, nor in any maximum ventilator setting other than a longer maximum inspiratory time in the study group. A more dynamic evaluation of variables was performed using standard deviation scoring, and the findings of slightly elevated peak inspiratory pressure and mean airway pressure in the study group prior to air leak were overshadowed by much greater prolongations of inspiratory time. Excessive prolongation of inspiratory time may be a major risk factor in the pathogenesis of pulmonary air leak.