Interest in mathemagenics (adjunct aids used in text to enhance learning) has most likely resulted from the indirect effects of inserted postquestions, whereby the learning of unquestioned material is enhanced as well as that of questioned material. However, since most mathemagenics research has occurred in laboratory settings in which subjects were prevented from turning back to a text segment once they had encountered a postquestion, the effects of mathemagenics in naturalistic settings remain problematic. A study was conducted to examine the nature of the nonreview constraints and to compare the results to those obtained in a naturalistic setting in which review was allowed. Subjects were 99 secondary school students who were placed in either an experimental (review) or a control group (nonreview) and given a 1,700 word passage to read. Those passages used by the control groups did not contain inserted questions; however these subjects were told that they would be asked to recall as much as possible about the passage after reading. All subjects were given free recall and cued recall tests following the experiment. Results showed that, contrary to expectations, relevant learning was not enhanced by inserted questions in the text. In addition, incidental learning was also slightly depressed by the use of inserted questions. This was true for both review and nonreview groups. (Materials used in the study are appended.) (FL)