This study examined the effects of reading purpose and working memory span on reading comprehension. Based on the reading purpose and reading span test (RST) score, 42 Japanese native speakers were divided into 6 groups with 7 members each. A 3×2 factorial design was used with the reading purpose (writing a summary of the text/ writing an introduction about the text/writing a reaction on the text) and RST score (high/low) as independent variables. There were 4 major findings: (a)the subjects who were given instruction to introduce the text to an elementary student paid much attention to the difficulty of the words and insignificant parts of the text, (b)there was no difference on the scores of the reading comprehension test among subjects with different reading purposes, (c)there was a difficulty on disregarding irrelevant information for subjects with low working memory span than those with high span scores, (d)the subjects with low working memory span who were given the instruction to write a reaction after summarizing the text obtained a lower score on the comprehension test. These results were discussed from a viewpoint of both textbase representation and situation model.