Thomas and Wang in a series of recent papers (Thomas and Wang, 1996; Wang and Thomas, 1992; Wang, Thomas and Ouellette, 1992) demonstrate one condition under which the Keyword Method fails to enhance the longer‐term retention of foreign language vocabulary, namely when immediate testing of learned material is prevented. However, they then go on to use this demonstration to argue that the Keyword Method in general is inferior to rote learning in the longer‐term retention of vocabulary. They discount the considerable evidence, some of which they produce themselves, that where immediate retrieval is required of learned vocabulary, the longer‐term retention of vocabulary is considerably superior to any rote‐learning condition. In practical terms, where interaction with testing is allowed, therefore, Keyword learning is the manipulation which most enhances later retention. Again, at a practical level, in real‐life situations, interactions determine actual performance, not just one variable acting on its own. © 1998 John Wiley & Sons, Ltd.