We determined the incidence of pneumothorax in 295 infants (mean birthweight, 1,917 gm) with the respiratory distress syndrome (RDS) treated according to the same protocol. Fifty-five infants (mean birthweight, 1,594 gm) developed pneumothorax (incidence, 19%); incidence varied with severity of RDS and intensity of respiratory assistance. Pneumothorax occurred in 3.5% (2 of 58) of infants who received no assisted ventilation and in 11% (14 of 124) of infants who received continuous positive airway pressure (CPAP) as the only form of assisted ventilation; the difference between these two groups is not significant. Forty-nine infants initially treated with CPAP later required mechanical ventilation with positive end-expiratory pressure (PEEP). Pneumothorax occurred in 12 of the 49 (24%) and in 21 of 64 (33%) of those infants initially treated with PEEP; the incidence of pneumothorax for both these groups was significantly higher than for those treated with no assisted ventilation or CPAP only. To assess the value of frequent measurement of vital signs, blood gas tensions, and pH in the recognition of pneumothorax, we analyzed these variables by the cumulative sum statistical technique. We noted the following significant changes associated with pneumothorax: arterial blood pressure, heart rate, and respiratory rate decreased in 77% of cases; pulse pressure narrowed in 51% of cases; Po2 decreased in 17 of 20 cases in which ventilatory settings were constant for at least three hours prior to pneumothorax. However, pH and Pco2 showed consistent changes. Frequent measurements of vital signs and Po2 aid in the early diagnosis of pneumothorax.