Ten neonates with respiratory distress requiring mechanical ventilation and supplemental oxygen were studied during a continuous 24-h period to determine the value of continuous transcutaneous oxygen (PtcO2) monitoring. All 10 infants were continuously monitored during the study with a Clark-type skin electrode (Litton) and 5 of the 10 also had a catheter-tip oxygen electrode in place in the umbilical artery to measure umbilical artery O2 (PuaO2). The results of these two forms of monitoring were not available for the care of the infant during the study period. Hypoxia, as defined by a PO2 of less than 50 torr, occurred for an average of 237 +/- 51 min/24 h from continuous PtcO2 monitoring as compared with 146 +/- 33 min/24 h by estimation from arterial blood gas (PaO2) (p less than 0.05). Hyperoxia, a PO2 of greater than 75 torr, occurred 69 +/- 16 min/24 h in the continuous group and 113 +/- 26 min/24 h from PaO2 estimations. Severe hypoxia, a PO2 of less than 30 torr, was not observed from PaO2 estimations, but was seen for an average 32 +/- 15 min/24 h from the PtcO2 monitoring. These latter differences were not significant. Correlation between PaO2 and PtcO2 values (r = 0.93) was greater than the correlation between PaO2 and PuaO2 (r = 0.81). PtcO2 = 19.7 +/- 0.74 X PuaO2, and the correlation coefficient between PtcO2 and PuaO2 was 0.64. Continuous oxygen monitoring revealed significantly longer periods of hypoxia than that observed from blood gas estimations alone and its use in the low birth weight infant should result in more rational ventilatory therapy and in fewer episodes of hypoxia.