A method is described for the assay of insulin. It involves pre-incubation of insulin under assay with an excess of guinea pig anti-insulin serum and subsequent incubation of the partially neutralized serum with an excess of unlabeled and/or I-131-labeled bovine insulin. If no more than about half of the antibodies in the serum are neutralized during pre-incubation, then the amount of bovine insulin bound during incubation decreases linearly and in proportion to the amount of pre-incubated insulin. Such a linear relationship was obtained with nine samples of pooled serum from uve groups of guinea pigs and was shown to be applicable over a wide range of serum and insulin concentrations. Since neutralized acid-alcohol affected the assay only when present in relatively high concentrations, the method has been applied to neutralized acid-alcohol extracts of pancreatic tissue without removal or excessive dilution of the alcohol. It is applicable at concentrations of insulin to be found in blood and is therefore a potentially simple and rapid method which would not require the use of labeled insulin of high specific activity. The uses and limitations of the method are discussed.