A radioimmunoassay technique has been developed for the measurement of cortisol in a single methylene chloride extract of human plasma without chromatography. The antiserum, obtained by immunizing rabbits with cortisol-3-carboxymethyl-oxime conjugated to bovine serum albumin, had a high affinity (KA = 1.8 X 10(9) 1/mole) and capacity (2.3 X 10(-6) moles/L undiluted serum) for cortisol. The minimum detectable amount determined at the lower 95% confidence limit of the buffer control tubes was 8.3 +/- 4.7 pg/tube and a log dose - logit response standard curve was linear between 20 pg and 20 ng/tube. The antiserum was highly specific for cortisol with only corticosterone, cortisone, 11-deoxycortisol and 21-deoxycortisol showing significant cross-reaction (12.4, 6.6, 3.8 and 3.7%, respectively). The cross-reaction for the other tested naturally occurring and synthetic steroids did not exceed 1%. Regression analysis of cortisol concentration estimates obtained on 20 samples before and after Sephadex LH-20 column chromatography gave a coefficient of correlation (r) of 0.995 and a regression coefficient (b) of 1.04. Recovery of cortisol added to plasma samples was quantitative. The intra-assay error was 8.5% and the inter-assay error averaged 5.7%. The method is simple requiring a single solvent extraction of plasma, therefore permitting large numbers of samples to be handled efficiently by a single technician.