The opening bars (shown below) establish a three-part texture with great clarity; the topmost voice charts a neighbor-note figure across mm. 1-4 and even embellishes itself with little neighbor note figures along the way. The bass is a pedal point G2, and the middle voice walks sturdily on a path from ^5 to ^1 (^8): D3-E3-F#3-G3. As we shall see, the relationship of the upper two voices is simply reversed in the final cadence, the pedal point bass there being V or D3. That the topmost voice is the principal one is confirmed a few measures later, when a stereotypical (^2)-^2 brings ^3 over I down to ^2 over V (see the rectangle frame in the example). Along with this, the role of D4 is plainly identified as a cover tone, or focus of an auxiliary (secondary) voice above the principal voice.
Disruptive cadenzas like those in the Eb Major Prelude are lacking in the G Major Prelude, but there some dramatic moments before the final flourish. In the first system below, a registrally expansive gesture runs quickly across the strings from the open C2 to our cover tone D4. The latter is pushed one half tone higher to Eb4 two measures later -- that's the highest note before the run up to G4 at the end. The effect is immediately vitiated, however, by a move downward and resolution to B3 as ^3 (end of the rectangular frame), and D4 is heard again it's obviously a cover tone once more (circled in the last system).
At the end, the formation of A3 (^2) above V could hardly be clearer, and the sudden chromatic rush up to G4 is a surprise -- although the chromaticism itself is a marker of the cadenza, and that is apparently how the figure is meant to function here. The diagonal line marked in the score suggests that the figure outlines (unfolds) a sixth from A3 to F#4.
I suppose one could argue that the stretched-out chromatic scale changes the relations of the voice leading strands to the point that the middle voice replaces the upper voice as primary, and therefore one gets a rising Urlinie from the cover tone D -- picked up in the middle of the chromatic scale -- up to G4. I am wary of these sudden reversals, however, just as I am of Urlinie ^3s that show up just a few bars away from the end of a piece. The cover tone D4 never has the kind of significance earlier that would predict such a change of role -- the rising line, then, seems an arbitrary choice. Thus, I would go with Schachter's "equals" here, the inner voice being "first among" them, and would probably notate using Channan Willner's "polyphonic Ursatz" (see yesterday's post). But there is a caveat: a descent to ^1 is plainly as forced and arbritary as a rising Urlinie would have been. The Prelude, then, ends as it began: with ^3 in the principal voice (and probably then a background shape involving neighbor notes -- see my comment on Arthur Komar's reading of the WTC, C Major Prelude in the MTS article, 291).
Postscript: Should D779n13 not be read in a similar way? Have I not said that the rising cadence gesture is a surprise? And does not the set of parallel sixths force B4 (^2) back up to C#5 (^3) rather than down to A4 (^1)? I could settle for the latter reading, but as to the rising gesture, it is only a surprise in terms of the clichéd formulas of the cadence (and, in the waltz repertoire, therefore, rather less of a surprise than it would be in most other genres). Unlike the cover tone D4 in the G Major Prelude, in D779n13 the F#5 (^6) that appears almost immediately and is touched on repeatedly thereafter forces constant attention to the "space above ^5" and sets all the conditions needed for a move further up at the end.