Teaching and learning situations nowadays typically build on richly illustrated material or multimodal presentations. Under these circumstances, the transparency of images and models used for explaining various phenomena becomes central. The present study deals with 20 Swedish children, 9–12 years old, discussing an illustration meant to show the cause of the different appearances of the Moon in the sky. The students’ task was to place eight numbered moon phases in the lunar orbit in the image. The illustration in question was chosen (a) because it was of a kind frequently used to explain the lunar phases and (b) because the phenomenon is known to be difficult to understand for students of all ages. The analysis leans on historical and sociocultural approaches as well as on multimodal semiotics. The results show that a majority of students were able to make sense of the most central features of the illustration but that very few spontaneously reasoned in a way that could be interpreted as the intended meaning-making of the cause of the lunar phases. The results also indicate that the simultaneous adoption of two perspectives necessary for understanding the phenomenon was a stumbling block for most of the students.