Just like in the similar situations in b. 1 and 3, we believe that it is the top arm of the mark in A that is more reliable. In FC Fontana averaged the length of the mark, which is one of possible solutions. We consider the mark in GE, slightly shorter than in FC, to be compliant with our interpretation of A. The mark of FE, stretched out, so that it covers an entire bar (and inaccurately reproduced in EE), is most probably a revision.
Compare the passage in the sources »
category imprint: Graphic ambiguousness; Differences between sources
issues: Scope of dynamic hairpins, EE inaccuracies, FE revisions
notation: Pitch