Abstract This experiment was conducted to study the effect of changes in the retention interval (RI) on spontaneous recovery within an acquisition-test interference task. College students learned a reversal conditional discrimination to solve a task involving conflicting phases across two training phases. When the test was conducted immediately after training, participants’ performance revealed recency, behaving according to the information received during the last phase. Performance after retention interval averaged the information received across phases, regardless of the length of the RI (1.5, 3, 24 or 48 h). These results are not in agreement with traditonal theories of spontaneous recovery as they predict a recency to primacy shift effect of the RI. An interpretation of spontaneous recovery based on a temporal weighting rule (TWR) is discussed.