Addressing a common problem for students-limited study capacities and a fixed deadline-this study investigated how and when selective practice should be implemented to maximize recall of practiced and unpracticed items. In each of three experiments, participants studied a list of items and recalled the material 48 hr later. Immediately after study, 24 hr after study, or shortly before test, participants practiced some of the items through restudy (Experiment 1), retrieval practice without feedback (Experiment 2), or retrieval practice plus feedback (Experiment 3). Results showed that, on average, recall was best after retrieval practice plus feedback and worst after retrieval practice without feedback, and it improved from early to late practice. Consistently, recall was highest after late retrieval practice plus feedback. Results also point to pronounced differences between practice types regarding the effects of practice timing and the relative contributions of practiced and unpracticed material to test performance.