This study demonstrated that the modality principle applies to multimedia learning of regular science lessons in school settings. In the first field experiment, 27 Dutch secondary school students (age 16-17) received a self-paced, web-based multimedia lesson in biology. Students who received lessons containing illustrations and narration performed better on subsequent transfer tests than did students who received lessons containing illustrations and on-screen text. In the second field experiment, 55 Dutch secondary school students (age 16-17) received similar multimedia programs that allowed more self-pacing and required students to record the time to learn. The illustrations-and-narration group outperformed the illustrations-and-text group on subsequent transfer tests for students who required less time to learn but not for students who required more time to learn. The interaction of learning time spent with modality of presentation on post-test scores was studied. Implications for testing of the robustness of cognitive theory of multimedia learning are discussed.