It is logically possible that there are two mecchanism for siphon -
effect:
1. Atmospheric pressure minus hidrostatic pressure in the arms of
different LENGHTS give rise to two different PRESSURES at the mouths of
short and long arms, and the liquid moves in the direction of the LOWER
PRESSURE.
The comment of Hero from Alexandria shows that it is not necessarly the
direction of the BIGGER WEIGHT.
2. In the absence of external pressure, the possible mecanism is that
there are two arms of different WEIGHT and that the liquid moves
in the direction of the BIGGER WEIGHT. This mechanism, of course, depends
on the existence of, at least, of one liquid being resistent to boiling in
vacuum and able to sustain positive stress necessary for pulling.
It would be nice to know if anybody has seen a published result of an
experiment which show that such a siphon does work.
If the existence of two mechanisms for siphoning were an experimental
fact, than we should teach siphon-effect more carefully, avoiding to
suggest that there is only one mechanism (either atmospheric
pressure based or gravity-cohesion based).