Educators often face serious time constraints that impede multiple repetition lessons on the same material. Thus, it would be useful to know when to schedule a single repetition unit to maximize memory performance. Laboratory studies revealed that the length of the retention interval (i.e., the time between the last learning session and the final memory test) dictates the optimal lag between two learning sessions. The present study tests the generalizability of this finding to vocabulary learning in secondary school. Sixth-graders were retaught English-German vocabulary after lags of 0, 1, or 10 days and tested 7 or 35 days later. In line with our predictions, we found that the optimal lag depends on the retention interval: Given a 7-day retention interval, students performed best when relearning occurred after 1 day. When vocabulary was tested after 35 days, however, students benefited from lags of both 1 and 10 days. Model-based analyses show that enhanced encoding processes and stronger resistance to forgetting--but not better retrieval processes--underlie the benefits of optimal lag. Our findings have practical implications for classroom instruction and suggest that review units should be planned carefully by taking the time of the final test into consideration.