The Community Bureau of Reference (BCR) is preparing a series of animal feed reference materials to provide a basis for analytical quality assurance for aflatoxin B1analysis, a problem of particular importance in view of Community legislation. Before reference values can be assigned to the reference materials the major errors in the underlying measurements must be identified and reduced. This paper presents the results of two intercomparison exercises involving some 20 European laboratories who applied a wide variety of analytical methods. It is shown that the major source of error and discrepancy is connected with incomplete extraction and/or losses during clean‐up and that, provided correction for recovery/background interference is made, many methods can achieve acceptable accuracy. Sources of error and their control are discussed, and essential details of the methods used are presented. It is concluded that analytical QA is more important than the use of standardized methods when a high degree of accuracy and comparability are required.