Diagrams in medieval manuscripts of Greek mathematical and astronomical works can seem peculiar for a modern reader, given their persistent and widespread tendency to represent more geometric regularity than the argument requires and their usual visual inaccuracy in depicting the mathematical objects discussed in the text. Although most scholars believe that these tendencies go back to the original Greek authors, in a recent paper I argued that these odd features should not be attributed to Greek authors, but to transmission. My (...) main argument there is based on experiments made with university students who played the role of copyists. In this paper, I show that the vestiges of the changes caused by transmission are present in the very diagrams of the extant manuscripts. I analyze four diagrams taken from Aristarchus's On Sizes. (shrink)