Testing-effect literature has shown the benefits of retrieval practice to enhance retention and attenuate forgetting. However, research in memory reconsolidation has demonstrated that a memory trace can be rendered labile by retrieval and requires restabilisation to persist. This study investigated the effects of the initial test interval and feedback timing on vocabulary retention while controlling the retention interval in the naturalistic setting. Participants were 60 undergraduate students from four intact sessions of a lower-level ESL course. The four groups completed three spaced, repeated tests on a vocabulary list taken from their textbook glosses. The initial test interval (short and long) and feedback timing (immediate and delayed) were manipulated. It was revealed that the optimal schedule to promote retention was a short-interval initial test followed by a delayed feedback. The results showed a U-shaped pattern of a downward fluctuation in the intermediate test in the presence of interference but an upward fluctuation in the final test across groups. Groups in the delayed feedback condition produced better final test performance than groups in the immediate feedback condition. In line with reconsolidation theory, the results suggested that memory is susceptible to decay and disruption after retrieval. Testing might not insulate memory strength and content from modification, but repeated retrievals might reduce the susceptibility to interference. The research findings have classroom implications for teachers in regard to scheduling tests and delivering feedback for durable learning.