The Menuet of Haydn’s Symphony No. 100 is a case in point. In the first period (measures 1-8, which stand for the whole), the initial motion is strongly downward, but the final cadence produces a clear ascent from ^5 to ^8 in the upper-most part.Thinking of this in Schenkerian terms—as I was in 1987—the rising line is not workable in the theme's first presentation because it doesn't mesh well with the bass, especially in bars 5-6, where one would have to imagine a doubling of bass and soprano, never a good idea. It's much easier to build a line in this way: D5 initiates a fifth-line; to C in bar 4, recapture C in bar 6, B on the last beat of that bar, then A in bar 7, and an implied G in bar 8. The ascending scale in the cadence is boundary play. See this version here:
In the reprise, on the other hand, the chromatic passing tone D# in the bass (from m. 6) is gone, and a string of diatonic figures, all rising, take over the lower parts, directly linking the chromatic scale fragment to the diatonic scale fragment (see the arrows in the figure below). As a result, the rising line from ^5 to ^8 has a clear path and pitch design can be read as well-matched to the important aspects of expression.
Nevertheless, nowadays I think that octave shapes work just as well as lines to describe the frame of this theme. In the first phrase, the ornamented arpeggiation that brings G5 down to G4 is only answered meagerly by the rising chromatic scale in bars 3-4. The second phrase does better, as A5 to A4 is answered by the diatonic scale that brings the close back to G5.
