There have been several failures to obtain learning-to-learn (LTL) in immediate recall experiments, while powerful LTL effects are commonly observed in other learning tasks. 2 experiments compared LTL in immediate Free and Ordered Recall as a function of number of presentations (1, 3, or 5) before a single test of recall (Experiment I), or on Trials 1, 3, and 5 when Ss were tested for recall after each of 5 presentations. With a single recall test, only Ordered Recall Ss showed LTL, while with repeated trials both Free and Ordered Recall Ss improved. In each experiment there is a shift in the direction of relatively greater recency effects with practice, but this in itself does not seem to be sufficient to explain all of the LTL observed.