Expanding retrieval practice (T. K. Landauer & R. A. Bjork, 1978) is regarded as a superior technique for promoting long-term retention relative to equally spaced retrieval practice. In Experiments 1 and 2, the authors found that expanding retrieval practice of vocabulary word pairs produced short-term benefits 10 min after learning, conceptually replicating Landauer and Bjork's results. However, equally spaced retrieval produced superior retention 2 days later. This pattern occurred both with and without feedback after test trials. In Experiment 3, the 1st test occurred immediately or after a brief delay, and repeated tests were expanding or equally spaced. Delaying the first test improved long-term retention, regardless of how the repeated tests were spaced. The important factor for promoting long-term retention is delaying initial retrieval to make it more difficult, as is done in equally spaced retrieval but not in expanding retrieval. Expanding the interval between repeated tests had little effect on long-term retention in 3 experiments.