ExtractThe arterial blood gas tensions in the right radial artery and the lower abdominal aorta were compared in 2 1 infants suffering from respiratory distress syndrome. Simultaneous sampling was carried out using a radial artery catheter inserted by a cut-down technique and an umbilical arterial catheter inserted into the abdominal aorta. One hundred thirty-nine simultaneous blood samples were analysed for paO,, pH, and pCO,. In no case was there any significant difference in p H and pCO, values in samples of blood drawn from the two sites. Differences in paO, were observed but owing to inherent errors in measurement, only differences of greater than 10 mm Hg were regarded as significant. In 27 out of the 139 determinations, paO, from the right radial artery exceeded that from the abdominal aorta by more than 10 mm Hg. The babies studied were divided into the following groups : Group 1, 6 babies with no significant paO, difference between the two sampling sites. Group 2(a), 6 babies with significantly higher oxygen tensions in blood drawn from the radial artery but only at aortic paO, levels of more than 140 mm Hg. I n this group there were no significant differences when the abdominal aorta paO, was less than 140 mm Hg. Group 2(b), 3 babies with slightly higher radial artery oxygen tensions only when they were being ventilated with a positive pressure respirator. Group 2(c), 6 babies with significantly higher paO, levels in the radial artery at arterial oxygen tensions of less than 100 mm Hg. I n two of these babies, the right radial paO, (and therefore presumably the retinal artery paO,) would have exceeded 160 mm Hg had the ambient oxygen concentration been raised sufficiently to produce an aorta paO, of 100 mm Hg and had the total shunt and the proportion passing through the ductus arteriosus remained constant. I n two of these babies, however, increasing ambient oxygen concentration decreased the total shunt and the proportion of the shunt passing through the ductus. In the remaining four babies in Group 2(c), the severity of the disease was such that it was impossible to raise the aorta paO, to 100 mm Hg.From these data it seems that most of the R -L shunt in respiratory distress syndrome must pass through the foramen ovale or the lungs. Only rarely does more than 15 % of the R -L shunt pass through the ductus arteriosus.The final quantitative differentiation between shunting in the lungs and through the foramen ovale is most likely to be made by dye dilution studies from peripheral vein or right atrial injections.