We performed the following experiments to evaluate the accuracy of our newly developed radioimmunoassay for somatostatin: (1) Recovery of synthetic somatostatin added to human, porcine, and canine plasma with or without extraction with 67% acetone or 76% ethanol, using 3 different region-specific antibodies and, where applicable, 125I-labelled Tyr-1- or Tyr-11-substituted somatostatin or 125I-N-Tyr-somatostatin as tracers. The recovery of somatostatin corrected for losses inherent in the extraction procedure was close to 100%, and independent of species, antibody and tracer. Somatostatin 1-28 was extracted slightly less efficiently. Unextracted plasma interfered massively in the assay. (2) Pharmacokinetic experiments with infusion of somatostatin into 14 pigs and determination of metabolic clearance rate (MCR) and T-1/2. MCR was 27-38 ml/kg per min, independent of infusion rate (6.1 or 13 pmol/kg per min), extraction procedure or tracer. T-1/2 was 1.9 min. The infused somatostatin was not measurable in unextracted plasma. (3) Characterization of endogenous and exogenous, labelled and unlabelled somatostatin 1-14 in human plasma, using Sephadex G-50 columns at pH 7.5 and 9.0. Human plasma showed excess immunoreactivity eluting at the void volume whereas synthetic somatostatin was recovered quantitatively at the position of marker somatostatin when added to the plasma. The immunoreactivity of the tracers was decreased (125I-Tyr-11-somatostatin) or abolished (125I-N-Tyr- or 125I-Tyr-1-somatostatin) after incubation with plasma or void volume fractions of plasma subjected to gel filtration. Extracted plasma did not contain void volume immunoreactivity, but like whole plasma, small amounts of components which coeluted with intact somatostatin.