Abstract Readers′ comprehension and eye-fixations are monitored as they read descriptions of simple machines, pulley systems. The comprehension data indicate that readers′ comprehension depends on both the medium of instruction and the ability of the reader. The conjunction of text and diagrams particularly facilitated the understanding of how the pulley system moved, whereas either medium alone was sufficient for conveying the system configuration. The eye-fixation data indicate that subjects integrate the information in the text and diagram at the level of individual pulley-system components or groups of connected components. They read the text in increments, often rereading the information about a component or group of components before constructing a spatial mental model of these components with the aid of the diagram. Subjects′ diagram inspections vary from local inspections concerned with encoding the relations between two or three components to global inspections concerned with integrating the relations between many components.