SummaryIn this paper, several aspects of the radioimmunoassay (RIA) method for digitalis glycosides are discussed. Although this method is used routinely, the results are not always accurate. Kits for RIA of digitalis glycosides are commercially available, but antibodies, e.g. against digoxin, can easily be obtained in rabbits : high titers are reached within a few months.The general procedure is described ; results can be calculated from a calibration line. The coefficient of variation«within run»and«between run»for the assay of digoxin in patient serum is approximately 10%.The RIA method has a limited specificity, and cross-reaction between different glycosides is present; this makes it possible to measure several glycosides, including digitoxin, using the antibodies against digoxin. Metabolites of digoxin and digitoxin interfere with the determination of the parent molecules.Important discrepancies exist between the results obtained with the different commercial kits : literature data and personal experience are cited.In a final paragraph, known sources of error are discussed. When performing RIA of digitalis glycosides, one should never forget the possibility that smaller or larger errors of measurement are possible.