Thanks Doug (am happy to learn from City as well as my own
institution!) and twocoolgliders.

So, if I understand you both correctly, the glider climbs on both
winch and aerotow because there is a force *pulling* it in (roughly)
the direction it is pointing, i.e. above horizontal. Once in a steady
climb, the lift generated by the wings balances the weight of the
glider + any other downward forces.

In a winch launch there are substantial downward forces from the
weight of the cable and the downward vector of the direction of pull.
Thus lift is higher than in steady free flight, and AoA is higher.

On aerotow the only additional downward force is from half the weight
of the towrope (pretty small), so the lift required is similar to that
in steady free flight (and in fact a little lower for other reasons).

_____________

This means that there are only two possible explanations for the
phenomenon on slow tow where the glider feels as if it is close to the
stall. Either:

1. It really is close to the stall, which means that the AoA is
greater than above, which means it must be flying in a continuous
downdraft (Andreas's explanation); or

2. Its AoA is as above, and the phenomenon has some other cause (such
as vortices acting on different parts of the wing) which replicate the
symptoms of approaching stall but do not in fact herald it.

Presumably we could test which is correct by taking a slow tow and
deliberately stalling the glider, monitoring the airspeed at which the
stall occurs. Volunteers to perform this experiment might be hard to
find!

Is there anyone who has actually stalled on tow unintentionally and
noted the airspeed when the stall occurred? I'd guess not, as the
pilot's attention would probably be elsewhere..