A collaborative study was performed in 15 laboratories to validate a monoclonal antibody-based enzyme immunoassay (EIA) for determination of gluten in foods. The study included 13 samples: maize starch, "gluten-free" baking mixes, wheat flours, cookies, cooked meats, and a soup. Gluten was present in these samples at either zero or 0.02 to 10% by weight, i.e., over almost 3 orders of magnitude. The mean assay values for the foods varied from 88 to 105% of the actual amounts. The assay was quantitative for cereal products and the soup with repeatability (RDS-r, relative standard deviation) and reproducibility (RSD-R) of 16-22% and 24-33%, respectively. The assay was semiquantitative for the processed meat products (RSD-r 14 and 26% and RSD-R 46 and 56%), probably because gluten was unevenly distributed in the small (1 g) samples that were analyzed. The ELISA method produced no false positive results, and false negatives obtained with tannin-containing foods could be avoided by use of a modified sample extractant. None of the collaborators reported problems in following the protocol. The method has been adopted official first action by AOAC for determination of wheat gluten in foods.