I'm reading Alfred Mann's translation of Fux's Gradus (original available here). On page 35, footnote 9 (presumably Alfred Mann, certainly not the original author) says "The tritone is to be avoided even when reached stepwise [examples] if the line is not continued stepwise and in the same direction." (He goes on to talk about tritones arrived at "by the progression of two voices".) The examples he shows are on the treble clef: F4, G4, A4, B4, and another: F4, G4, B4.
However, on page 36, Figure 14, the bass part is: F3, E3, F3, A3, Bb3, G3. This violates the rule given in the footnote by moving from E3 to Bb3 without continuing stepwise in the same direction. I say to myself, "No problem. Nobody's perfect" and continue only to find the same situation in the next example. Page 37, Figure 15, soprano line, notes nine to the end are: G4, C5, A4, G4, F#4, G4. Again, from notes 10 to 13, we have a descending tritone without continuing stepwise in the same direction! Okay, two "flukes" is a bit much, but warily, I continue. The next example, on Page 39 -- Figure 21 -- tenor (bottom) part, notes 8 - 12 are B3, C4, A3, F#3, G3. Again, a descending tritone not followed by step in the opposite direction. Three examples in a row?
It seems that Fux and Mann simply disagree here, but which is correct? Did Fux not care about this "rule"? Did others care about this "rule"? Why would Mann write something like that just before 3 (!) counterexamples?