Abstract

We argue that diagrams are not just a communicative tool but play important roles in the reasoning of biologists: in characterizing the phenomenon to be explained, identifying explanatory relations, and developing an account of the responsible mechanism. In the first two tasks diagrams facilitate applying visual processing to the detection of patterns that constitute phenomena or explanatory relations. Diagrams of a mechanism serve to guide reasoning about what parts and operations are needed and how potential parts of the mechanism are related to each other. Further they guide the development of computational models used to determine how the mechanism will behave. We illustrate each of these uses of diagrams with examples from research on circadian rhythms.