Trying to keep my post size down...
Ok, I see that isolated experience can be a problem. As far as I can see
there will be a level of complexity best averted by empirical testing, so
long as I don't miss something really fundamental. Which I am trying not to
do.. I found a USMC guide to RF and antenna selection, a 5MB PDF file, that
has a Marine's habit of cutting to the chase, it looks like I found a good
guide there.
http://www.armymars.net/ArmyMARS/Ant...antenna-hb.pdf

The topolgy I describe WILL carry imbalance through, I know. I thought it
might at least be a start if it didn't add any of its own. I realise you were
using the amps as an illustration of how imbalance can arise between those
toroids, I was thinking that fewer parts means less to go wrong. Precise
symmetry in topology might be a lot easier if I didn't have to use those.
Again, I wonder why they're even there. Isn't it easier to put a gain stage
after the second toroid?

Point taken on the costs of the selector. It was a long time ago, but I think
I saw something intended to do this task, a design that built it from wire
wound on a bit of plastic water pipe or something. I started Googling for
designs earlier but got waylaid. One bit of waylaying involved a remark that
tuning might offer more problems that a fixed narrowband filter, but so
little context was offered that I don't know how to judge the remark.

If I put up conspicuous loops here I might get people bothering me about
planning permission or some other means of negative compulsion. If I can
do this with a vertical whip it will be much less awkward.