Four methods for the identification of Enterobacteriaceae in cooked frozen meat have been appraised. Although these methods could be modified to improve efficiency they could be clearly classified into two categories: (i) those that allow identification to genus level and which often cannot discriminate between two genera without the help of additional reactions; (ii) those that allow identification to species level and which only occasionally require additional reactions to differentiate one species from another. A different evaluation criterion was followed for each of these categories. In the first, as methods lead to final identification by means of data furnished by the manufacturer or via a number code which does not indicate probability, reactions were compared with those obtained by classical methods. In the second group, every number code admits three possibilities and indicates a probable biotype for each, thus reducing the risk of faulty interpretation on the part of the operator while also compensating for deficiencies of certain reactions. Evaluation was therefore based on the percentage of correctly identified cases per species. Efficiencies of the main biochemical reactions are also discussed.