The A section is in two segments, the first closing with a PAC on V--see the cadence chord in bar 9 below. The set of circled notes in the piano, right hand, show steady step-wise progress upward from Eb5 to Ab5 in the cadence that closes the A section.
The voice part works quite differently. I have isolated it below. The Eb5 at (a) is the seventh in an F7 chord, dominant of Bb. It is touched on again and then resolves downward, first to D-nat5, then Db5, and we would expect the line to continue downward to C5, but instead the voice returns to Eb5 at (d) and the piano left hand takes over to resolve the Db (as seventh of Eb7)--see arrow. At (b) an inner voice moves up by step reaching A-nat4, which joins the upper Eb5 in an unfolded diminished fifth -- at (c1). This resolves as expected to a third at (c2); I have shown the upper note as Db5, but to match harmony in the strictest way it would have been D-nat5. In any case, as the upper voice Eb5 is recovered at (d), the lower voice moves on from Bb4 to the lower note of the unfolded fifth Ab4-Eb5 at (e), and cadential closure follows -- with that framing interval, Ab4-Eb5, intact (see final bar). (The two asterisks, btw, point to admittedly obvious register play.)
The detail in the discussion above is of interest because of the way Wolf rewrites the reprise of the A section. The first half, bars 59-67, is shown below. The piano carries its left-hand tenor melody again, and the voice winds its way within the interval frame Eb4-Ab4-Eb5. The cadence, however, is most easily approached from ^3, as shown beginning in bar 64.
The second half of the reprise suggests the possibility of the voice part moving onward from that ^3 (see bar 69) and steadily upward to the cadence on Ab5--see bars 77-79. The piano part goes even further, expanding its right-hand movement from earlier to move step-by-step to overtop the voice at ^3 (as C6).