This preliminary study devised and tested a series of psychometric tests to facilitate selection of simultaneous and consecutive interpreter-trainee candidates. Twelve tests, based either on text materials, linguistic subskills or speed-stress, were correlated with judges' ratings of the final interpretation examination. Students who passed the exam had higher mean scores on all tests than those who failed. Text score were positively intercorrelated between completion-detection tests and simultaneous interpretation ratings and between recall tests and consecutive examination ratings. Text-based tests were more predictive than subskills or speed)-stress tests. The relationship between test type and interpreting is discussed together with possible reasons underlying differential performance on test of the same kind.