Three experiments investigated word frequency and age of acquisition (AoA) effects in recognition and recall. Experiments 1 and 2 used the "remember-know" procedure developed by J. M. Gardiner (1988). In Experiment 1, recognition performance was higher for low-frequency words than for high-frequency words and higher for late-acquired words than for early-acquired words, but only in "remember" responses. Experiment 2 replicated the AoA effect by using a different set of early- and late-acquired words. Experiment 3 found advantages for low-frequency and late-acquired words in recall, but only when words were presented in mixed lists. The frequency effect was reversed, and the AoA effect was eliminated, when participants studied pure lists. Findings were attributed to the more distinctive encoding of low-frequency and late-acquired words.