The present study investigated the effects of repetitions and questions (without feedback) at varying lags during self-paced learning from text. High school students read a series of unrelated paragraphs, each of which was repeated or tested after a variable lag. In a mixed condition, texts and repetitions of particular sentences were combined within paragraphs. Subjects were given free-recall and cued-recall tests one week later. In general, repetition did not improve performance. Correct answers to questions during the initial study task decreased as presentation-question lag increased. However, question-answering improved performance on the delayed tests substantially; the effect was constant over all lags studied. Question-answering produced better performance than repetition. Question-answering appeared to have operated primarily on memory traces of the information specifically tested. There was no evidence of any generalized attentional or strategy shifts or of any spread of effects to topically related information in memory. The effect of answering questions correctly appeared to have increased with increasing presentation-question lags. (Author)