A procedure is presented for screening foods for radionuclide contamination. It was developed by the U.S. Food and Drug Administration (U.S. FDA) to be an option for augmenting analytical capability following a major radiological contamination event involving beta- or alpha-emitting radionuclides. The expected application of this procedure would be during late-phase monitoring, after initial monitoring suggests an area is contamination-free or levels are negligible but additional confirming data are desired. When food is taken from multiple samples and a composite analytical portion is prepared and analyzed using a quantitative method, it is possible to show that radionuclide activity levels are below a regulatory limit for a number of samples simultaneously. Although radionuclide activity levels are not obtained for individual samples, the number of samples that can be processed can be increased dramatically. In application, a limited number of selected samples would be analyzed using the usual unmodified quantitative method while the screening method would be used to provide supporting information for bulk quantities of samples. The procedure involves combining equal-mass portions from a number (n) of food samples to make a composite analytical portion, which is then analyzed using a quantitative method. Instead of comparing results with the regulatory limit, they are compared with a screening level equal to 1/n of the regulatory limit. If the observed activity concentration for the composite analytical portion is below the screening level, then the radionuclide levels are below the regulatory limit for all samples represented by it. Screening throughput will therefore depend on n. For n = 2, 3, etc., sample throughput will double, triple, etc., respectively. The maximum number of samples that may be combined is subject to limitations such as those associated with sample nonhomogeneity, detection capability, and the need to be able to discern abnormal radioactivity from that typical for a potentially-affected area.