A method is described for the extraction and concentration of oxytocin from plasma which is simpler and more rapid than other available procedures. It has proved satisfactory for both radioimmunoassay and bioassay of circulating oxytocin. The recovery of added oxytocin was 60 ± 5·5% and showed no significant variation between plasma from pregnant and non-pregnant women, or plasma from other species. The sensitivity of the assay is related to the volume of plasma extracted. With a 10 ml plasma sample and the radioimmunoassay method described previously, the maximum sensitivity under optimal conditions is 0·75 μu. (1·5pg)/ml. In the second stage of human labour, oxytocin is not detectable in maternal plasma at this level of sensitivity. Of 36 cord venous plasma samples studied, 15 showed positive results in the range 1·5–20 μu. (3–40 pg)/ml; of 16 simultaneous cord arterial and venous plasmas, 12 showed positive results; the arterial samples showed a range of 8–145 μu. (7–290 pg)/ml with an average of 45 μu./ml; the venous samples showed a range of 0–100 μu. (0–200 pg)/ml with an average of 24 μu./ml. Plasma oxytocin levels during the second stage of labour in the goat averaged 120 μu. (240 pg)/ml by radioimmunoassay and 100 μu. (200 pg)/ml by bioassay. The half-life of infused oxytocin in the non-pregnant human subject as determined by radioimmunoassay was 5 min.