In five experiments, college-age students of differing foreign language-learning abilities were asked to learn Latin word translations to determine the effectiveness of the keyword method of foreign language vocabulary learning. The Latin words were the types for which it has been argued that the keyword method effects would be maximized (the words had concrete meanings with easy-to-generate concrete keywords). Specifically, the experiments were designed to determine whether the keyword method would be effective with both more and less proficient foreign vocabulary learners, would be as helpful when students had to generate their own keywords as it was when students were provided with keywords, would increase performance on foreign items associated with more than meaning, and would have any incidental effect on foreign word spelling. Results indicated that (1) the method was equally effective for good and poor foreign vocabulary learners; (2) the method was effective for foreign vocabulary items associated with more than one English meaning; (3) for vocabulary items with easy-to-generate keywords, it made little difference whether the keywords were actually provided by the experimenter; and (4) providing keywords resulted in increased incidental spelling errors on items for which keywords and corresponding foreign word syllables did not match exactly. (Subjects, procedures, and results for each of the five experiments are discussed.) (HOD)