In footnote 32 to my "Ascending Urlinie" article, I included the Haydn Piano Sonatas in Eb and Ab—the slow movement of the former (Hob. XVI/52, II), the menuet of the latter (Hob. XVI/43, II)—among pieces that use one of the variants of the rising line: the form ^5-^6-(reg.)^7-^8. We've seen this version already in the menuets of Symphony no. 86 and no. 104.
In this post, I will use my holograph sketch of this piece; it's probably from 1982 (when I did most of the initial research on rising lines for the sake of a Schenker seminar). I have placed a facsimile in my public folder on Dropbox: link.
The opening is one of those designed to frighten beginning Schenker students, as it offers ^5, ^8, and ^3 as plausible starting points for an Urlinie. Although ^3 is weak, since it is over vi, not I, the move to ^2 in bar 4 has to be encouraging; and you can almost always read chord support backwards to the beginning if you really want to (true here), so that ^3 is understood to be supported by the initial tonic chord rather than the vi that prolongs that I.
In 1982, however, I read the Urlinie from ^5, not at all disturbed by its cover-tone quality, as ^5 very often sounds like that in its prolongations. The ^3 and its interruption, then, are placed in an alto voice. See the condensed version of my sketch below.
In the second strain's altered reprise, one can certainly be forgiven for wondering about ^3 again—note the prominent C6, then the double neighbor figure—but one is obliged to imply/invent the ^2 in the cadence. A line consisting of ^5-^6-(reg) ^7-^8 is more direct and also more musically satisfying.