Clinical evaluation of the continuous, simultaneous measurement of fetal scalp surface oxygen and carbon dioxide partial pressures by mass spectrometry was undertaken for 52 labors. The mass spectrometer (MM8-80, V.G. Gas Analysis, Winsford, England) was easy to operate and had good long-term stability. The mean drifts for both oxygen and carbon dioxide over the study periods were less than 2 mm Hg. The mean (+/- SD) cervical dilatation at the time of transducer application was 6.1 (+/- 1.9) cm and the mean (+/- SD) duration of the studies was 169 (+/- 122) minutes; 10.5% of the transducer applications were unsuccessful. Falls in fetal scalp surface oxygen levels and rises in carbon dioxide levels were more frequent with late than with variable and with variable than with early fetal heart rate decelerations and with increasing severity and frequency of decelerations. Fetal scalp surface pressure changes also occurred with fetal heart rate variability changes, including some related to behavioral state changes. There was not a constant reciprocal relationship between oxygen and carbon dioxide changes, and fetal heart rate patterns were not related to actual blood gas levels. Fetal scalp surface measurements were related to both fetal blood sample and umbilical artery results. Trends in both oxygen and carbon dioxide levels during the course of labor were compared and related to other fetal variables, and most of the time the scalp surface measurements were an accurate guide to systemic blood gas levels. Maternal oxygen administration resulted in significant increase in fetal scalp surface oxygen levels, and on two of eight occasions it also led to decreases in fetal carbon dioxide levels. Scalp surface gas measurement by means of mass spectrometry is a powerful new method of intrapartum fetal monitoring, which should increase the precision of fetal surveillance as well as allow the accurate assessment of both established and new methods for optimizing labor and delivery.