||In an extension of research demonstrating the testing effect, the present study examined three questions concerning the effect of tests on learning. What is the optimal distribution of test and study times? Is retrieval necessary to reap the benefits of the testing effect? Does testing reduce the rate of forgetting? Participants cycled through a list of Eskimo-English word pairs 2 or 4 times, then were given an initial test on those items. The initial test manipulated the time available for retrieval (and additional study) by 0-, 2-, 5-, 8-, or 10-seconds within an interval of 10 seconds. After a brief distracter task, the final test on the word pairs was presented immediately (5-minutes) or 48-hours later. Performance on the final test revealed differences at each retention interval, such that study appeared to be important for the shorter interval, whereas testing was better for the longer interval. Findings further revealed the need for future research studies to include shorter retrieval conditions that could disrupt retrieval processes and conditions that could tease apart the inverse test-study association. Overall, conditions that included some form of testing provided protection against forgetting.