The purpose of this study aims at understanding the effects of reading texts by oneself with different layouts on 7th graders' mathematical performance about ”Pythagorean Theorem.” This study designed ”proof-first” and ”application-first” texts, and instruments for testing mathematical performance about ”Pythagorean Theorem.” After collecting pre-test and post-test data, the statistical method of ANCOVA was mainly used to analyze experimental data. The results showed that the interaction between text layouts and students' mathematical abilities was statistically significant, and that students, who read a mathematical text by themselves as testing in mathematics related to the text, did not necessarily perform better than students, who did not read the mathematical text by themselves as testing.