Hey guys, thanks for all of the commentary/concern ... and sorry I've
been remiss in responding until now.

Thanks in particular to Bruce for going the extra mile and a half in
actually hooking up and testing to see if you can pump backwards thru
a typical Facet pump (nope).

A simple "gear motor" mentioned elsewhere is what I need, I suppose,
so that the single pump could pump both ways (with polarity reversed).
I could add an on/off valve (manual or electric solenoid) to complete
the single-pump-and-single-tube hookup. Such gear motors are
typically found in cheap hardware-store siphoning pumps that let you
drain or fill your auto crankcase via the dipstick tube (turned by an
electric hand drill). I wonder if I can find one of aircraft quality
somewhere.

I did find the ppavionics site mentioned. That's the link I mentioned
having found (and then temporarily misplaced) earlier. They've got a
Facet pump model that contains internal valving to prevent flow in
either direction when off. So they could be used for a two-line and
two-pump setup for "fill" and "drain". I found some other stuff on
their site that seems "interesting", and wrote them (no response as
yet). (And no, I have no financial or other interest.)

I just want to assure you all that even when the aux tank is full,
I'll still be within rear CG limits ... barely. I wouldn't really
CHOOSE to take-off or land (or do intentional stalls) at that rear
limit, but it WILL be possible to do so safely.

I'm currently leaning towards the obvious two-line and two-pump
arrangement, but am still on the lookout for a lighter/simpler
alternative using a single line and single pump.

As I mentioned earlier, it's easily possible to use a single line and
a single pump to fill and drain (with the pump always running in its
normal mode) by using a stacked selector valve as used with a
fuel-injected engine, and a bit of fancy cross-plumbing of the valve
ports. This is basically trading the expensive selector valve for the
second pump and line. It might be a little lighter and somewhat less
complicated.

Someone else suggested much earlier using a single pump to fill and a
"bypass valve" to drain by gravity. That's an excellent and simple
solution -- if I could trust that gravity alone will drain the tank
within a reasonable length of time. I need to do a simple experiment
to find out.

Thanks again for all the comments,
Greg