The availability of a sensitive and highly specific rabbit antiserum and the development of a peptide-extraction method employing glass beads permitted the evolution of a rapid reliable radioimmunoassay that measures the sum of the concentration of angiotensin II and its active metabolite, angiotensin III. At a dilution of 1:32,000 the antiserum is capable of measuring 1 fmol (1 pg) of angiotensin II. Cross reactivities of this antiserum, taking angiotensin II as 1.0, are: angiotensin III, 0.75; angiotensin-(3-8) hexapeptide, 0.11; angiotensin I, 0.006; angiotensin-(1-14) tetradecapeptide, 0.0001. The recovery of angiotensin II added to hormone-free plasma was 73 +/- 2% [mean +/- standard deviation (SD), n = 20]. When 0.9 ml of plasma was extracted, the minimal concentration of angiotensin II and III that could be quantified was 4 fmol/ml. When larger volumes of plasma were extracted, sensitivity was enhanced. Plasma blanks were zero. Intra-assay variability was 7.6% SD and interassay variability was 11.7% SD. Angiotensin II and III concentration in venous plasma of normal volunteers on an ad libitum diet was 15 +/- 8 fmol/ml (mean +/- SD, range less than 4 to 35 fmol/ml). The plasma of a patient with primary aldosteronism had an unmeasurable value (less than 4 fmol/ml). Posture, converting enzyme inhibition, and renal artery stenosis resulted in expected changes of angiotensin concentration.