Well, I was wrong about one thing - I discovered that the normals for the back wing surfaces were indeed wrong (pointing inward). Because I could see a dot where the normal was, I didn't realize that you can see the normals on either side of the surface.
I found another thing that made things better - I 'optimized' surfaces and vertices - and 3 or 4 things were removed. Then, doing the subdivision was much more symmetric, and did not have holes.
However, it still isn't quite right. Will post more pictures tomorrow, if you aren't totally sick of me by then
.
Thanks again,
Stan