Hi Jai,
" To be honest, I don't really like what I've seen of the scabious flower form. "
And you may be in for some more disappointments from the scabious zinnias. They have a reputation of not being true to form. Many years ago the first scabious zinnias I grew had maybe one in twenty that were true scabious, with the other nineteen being small semi-double zinnias. That is a dismal performance for a commercial zinnia variety.
I understand that in recent years that situation has improved some, but scabious strains still tend toward frequent off-type specimens. As an amateur zinnia breeder, I took the optimistic "the glass is one-twentieth full" approach, and crossed that one-in-twenty good one with cactus flowered zinnias. And I did get some very pleasing results.
" I am just growing them to see what happens. "
That is an excellent scientific attitude.
" On scabious zinnias does the pollen actually come from the petal-like structures in the center which have changed color? "
Yes indeed. Those central floral parts are actually a mound of petal-colored pollen florets, complete with internal pollen producing anther bundles. The classic scabiosa flowered zinnia has a single row of conventional petals, called "guard petals", topped with a mound of scabious colored florets. In some scabious zinnias the guard petals are so short that they are basically hidden by the mound of scabious florets.
The real value of scabiosa flowered zinnias is in what they can do in crosses. They can do a lot of stuff. One of my favorite scabious hybrids looked like a water lily.
Its scabious florets were yellowish and normal sized, but its guard petals were gigantic. Scabious hybrids with longer, slimmer guard petals can look rather like an echinacea.
Scabious genes can produce some surprising recombinations in zinnias. Another less dramatic feature is their plant habit, which tends to be much more branched than conventional zinnias. I think they have great potential for improving zinnia plants.
Be prepared to discard off-type scabious zinnias, but know that the on-type ones can participate in some interesting hybridizations.
ZM