AES wrote:

You guys are into some interesting physics here. Do the plane wings
frost even when the surrounding air is above 32 F because their net heat
transfer to the surrounding above-freezing air is not as strong as their
net radiative heat transfer to the cold sky?

No, it's not necessary for the heat transfer to the night sky to be
"stronger", just for it to be a significant effect. According to:
http://www.efunda.com/formulae/heat_..._enclosure.cfm
the radiative temperature of the clear night sky is about -150 F, so
if the wing surface got no heating from its surroundings and reached
thermal equilibrium with the night sky then it would cool to -150 F.
On the other hand, if it didn't lose any radiational heat to the sky
but was in complete equilibrium with the ambient air temperature that
dropped to a low of 40F, then that would have been its lowest
temperature. In reality, both effects occur. It gets some heating
from the surrounding air and also conduction from other parts of the
wing and plane, but it also loses some heat through radiation to the
cold night sky. The resulting temperature is therefore somewhere
between +40 F and -150 F and in practice probably around 25F - 30F on
a perfectly clear night and very close to 40F on a cloudy night.

(which should also imply
equally strong net heat connectivity to the ground below, is that not
so?)

Yes, for the bottom surface of the wing. I wouldn't expect ice to form
on that surface on nights where neither the air nor ground temperatures
drop below freezing.

Or do they acquire and retain frost because some colder and therefore
heavier below-freezing air just fell down on them at some earlier time?

In that case a thermometer registering the ambient air temperature
should also record this below-freezing temperature sometime during
the night. But the lowest temperature recorded on the night in
question was reported to be +40F.

I'll wait for an authoritative answer, but bet on the latter explanation.

How much?