This study investigated two issues: (1) whether availability of planning time affects learners' ability to notice and learn from recasts in the input; and (2) whether pre-task or online planning is more effective. Participants were randomly assigned to three groups that formed the treatment conditions: a no planning group (N = 13), a pre-task planning group (N = 14), and an online planning group (N = 14). All participants took a pre-test, an immediate post-test, and a treatment session, and 25 participants took a delayed post-test two weeks later. Results show that both planning conditions are more beneficial for the noticing and learning of recasts than no planning and that online planning may provide more optimal conditions for recasting than pre-task planning does. (Contains 14 notes, 8 tables, and 3 figures.)