The literature on testing effects is vast but supports surprisingly few prescriptive conclusions for how to schedule practice to achieve both durable and efficient learning. Key limitations are that few studies have examined the effects of initial learning criterion or the effects of relearning, and no prior research has examined the combined effects of these 2 factors. Across 3 experiments, 533 students learned conceptual material via retrieval practice with restudy. Items were practiced until they were correctly recalled from 1 to 4 times during an initial learning session and were then practiced again to 1 correct recall in 1-5 subsequent relearning sessions (across experiments, more than 100,000 short-answer recall responses were collected and hand-scored). Durability was measured by cued recall and rate of relearning 1-4 months after practice, and efficiency was measured by total practice trials across sessions. A consistent qualitative pattern emerged: The effects of initial learning criterion and relearning were subadditive, such that the effects of initial learning criterion were strong prior to relearning but then diminished as relearning increased. Relearning had pronounced effects on long-term retention with a relatively minimal cost in terms of additional practice trials. On the basis of the overall patterns of durability and efficiency, our prescriptive conclusion for students is to practice recalling concepts to an initial criterion of 3 correct recalls and then to relearn them 3 times at widely spaced intervals. (Contains 14 figures, 4 tables and 4 footnotes.)