Contrary to the assertion of Arthur Schmidt (PHYSICS TODAY, April 2011, page 11), atmospheric pressure is not necessary to force fluid flow in a siphon. Force, after all, depends on pressure gradient, and the atmospheric pressure gradient in an ordinary siphon is insignificant. Schmidt’s comparison of a siphon to a drinking straw fails to account for the large difference in air pressure gradient for the two devices.
In a common argument, the siphon is said to be driven by the weight of the liquid on the longer leg. That is true, but it would be untrue if the liquid had no tensile strength. You cannot siphon sand.
If the tensile strength is sufficient to hold the long-leg water column together, then it is certainly sufficient to do the same for the short-leg water column. That the tensile strength is indeed sufficient is demonstrated by the Geissler mercury vacuum pump that Thomas Edison used for his early lamps.
If there is an air bubble in the siphon column, then atmospheric pressure is also required. Otherwise, atmosphere is needed only to prevent the fluid from evaporating.
