I suspect the second half is illustrating the fact that the seventh sometimes has an ornamented resolution. I guess it’s kind of similar to the point about how a seventh can be transferred to another voice, because the seventh is introduced at the beginning and then seems to go away, but is then reintroduced on the upbeat of beat 2, reestablishing the seventh before moving to the Eb chord on beat 3. It doesn’t shift voices, but it does have a similar status since it looks on the surface like it isn’t resolving down like a seventh should. Indeed, there’s a further delay of the standard downward resolution as the seventh hangs on as a suspension in the Eb chord, which is then itself further delayed by a decorated resolution. The ultimate resolution of the seventh introduced on beat 1 doesn’t happen until the final note of the measure, after many changes of embellishing function and several surface embellishments of the embellishment.
Aside from the similarity that both voice shifting and extensive decoration cause what looks like a surface violation of the downward resolution of all sevenths, I can’t explain why they’re put in the same example. I’m very surprised there isn’t some text explaining these details, which are an important additional example of the underlying principle of delayed resolution. Perhaps there was a breakdown during the editing process.