AbstractTaking its point of departure in the origin of the notions of “diagram” and “iconicity” in Peirce’s philosophy of logic, this paper reviews and discusses a series of dimensions along which such diagrams may be compared, measured and subdivided: diagrams versus images and metaphors, operational versus optimal iconicity in diagrams, diagram tokens versus diagram types, diagrams as general signs; corollarial versus theorematic diagram reasoning; pure versus applied diagrams; logic diagrams versus diagrams facilitating logical inferences; continuous versus discontinuous diagrams; diagrams in non-deductive reasoning. Most of these developments occur in the mature Peirce after the turn of the century and thus form an important part of his mature semiotics – yet, they do not relate in any simple or straight-forward manner to his attempts at enlarging his combinatorial semiotics from its bases in the three-trichotomy theory of the 1903 Syllabus over the six-trichotomy theory of 1904–1906 to the sketchlike ten-trichotomy version of 1908, where diagrams rarely figure in the names of sign-types discussed – why?