College students numbering 206 were examined on their beliefs of the movement of sunflowers, and 112 students who had the false belief participated also in the experiment. The subjects were asked to read the science text which explained the facts that contradicted their beliefs in the following three conditions: (a) the photosynthetic rule was instructed, and the contradictory facts were referred to as examples of the rule; (b) the photosynthetic rule was instructed, but the facts were referred independently from the rule; and (c) only the facts were presented. The subjects were then put to some reading comprehension tests. The frequencies in the occurrence of belief-dependent misreading (BDM) on the tests were analysed. The following results were obtained: (1) there were less BDMs in the condition of the rule and example than in the other two conditions; (2) there were no less BDMs in the condition of the rule and facts than in the condition of the facts only. There findings suggested that the instruction in the relation of the rule and example was useful in order to avoid BDM.