The paper focuses on student-teachers' geometric diagrams to mediate the emergence of different proofs for a geometric proposition. For Peirce, a diagram is an icon that explicitly and implicitly represents the deep structural relations among the parts of the object that it stands for. Geometric diagrams can be seen as epistemological tools to understand explicit and hidden geometric relations. The systematic observation of and experimentation with geometric diagrams triggers abductive, inductive, and deductive reasoning which allows for the understanding of the conditions given for a geometric construction and its necessary logical consequences. We adopt Stjernfelt's model of diagrammatic reasoning to analyze two proofs for a geometric task posed to student-teachers who participated in a four-month classroom teaching experiment. [For the complete proceedings, see ED583608.]