The temporal dating (knowledge of an item's position in a list) of subjects who had been told they would be tested by free recall was then assessed by a reconstruction task, after they had received either one list or six. Other subjects were tested by free recall after the same amounts of practice. Free recall showed a drop in primacy with practice whereas reconstruction did not, which suggests that temporal tags per se may be relatively unimportant as retrieval cues for the recall of early list members.