How Does Water Climb Over the Edge? An Explanation That Was Wrong for Centuries

A siphon is a tube filled with liquid, with one end in a reservoir and the other outside it below the level of the liquid surface. Once started, liquid flows continuously up the short arm, over the highest point and down the long arm until the reservoir empties to the level of the tube's inlet or until air enters. No external energy is supplied, and the flow continues against gravity in the rising section, which is what makes it look like something for nothing. The energy actually comes from the liquid descending, since the outlet is below the surface in the reservoir, and the net effect is liquid falling from a higher level to a lower one by an indirect route.

The traditional account attributes everything to atmospheric pressure and cannot be complete:

• •It holds that air pressure pushes liquid up the short arm to replace what has fallen down the long one
• •That predicts a maximum height set by the pressure the atmosphere can support, around ten metres for water
• •Siphons have been operated at greater heights than atmospheric pressure allows using degassed liquid
• •Siphons have been demonstrated to work in a vacuum chamber, where there is no atmosphere to push
• •The account also fails to explain why the liquid column does not simply separate
• •The height limit it predicts is real for ordinary water and is not the whole story

The better account relies on the liquid behaving as a continuous connected body. Molecules in a liquid attract each other, and in a column of liquid within a tube that cohesion transmits tension, so the descending column on the long side effectively pulls the liquid behind it over the top, in the same way a chain hanging over a pulley pulls the shorter side up. That is why the vacuum experiments work, and it is why degassed water, which has no dissolved gas to form bubbles, can be siphoned to heights far beyond the atmospheric limit. Ordinary water contains dissolved gas that comes out of solution under tension, forming bubbles that break the column, which is where the practical ten-metre limit comes from.

The device is used far beyond emptying a fish tank and several applications are substantial. Inverted siphons carry water and sewage beneath a river or a road by dipping the pipe down and back up, which is a standard civil engineering arrangement and is misnamed since it works by pressure rather than by tension. Reservoir spillways use siphons that start automatically when the water rises above a set level and stop when it falls, which requires no moving parts, no power and no operator. Toilets use a siphon to produce the sudden complete discharge that a simple valve would not, which is why the bowl empties in a rush rather than draining steadily. Irrigation systems distribute water over bunds with small portable siphon tubes.

The episode is instructive about how explanations persist. The atmospheric account appeared in dictionaries and textbooks for a very long time, including in a major dictionary whose definition was corrected in 2010 after a physicist wrote to point out the error, which received press attention at the time. It is not a case of a subtle refinement, since the two accounts make different predictions and one of them fails experimentally. It persisted because it is simple, because it gives the right answer for ordinary conditions, and because nobody with authority had reason to question it. That pattern applies well beyond siphons, and the useful lesson is that an explanation giving correct answers in the usual case may still be the wrong explanation.

Liquid flows up and over because the descending column pulls it, with cohesion between molecules transmitting tension through the tube, and the energy comes from liquid falling to a lower level by an indirect route. The atmospheric account fails because siphons work in a vacuum and above the atmospheric height limit with degassed liquid. Dissolved gas forming bubbles is what limits ordinary water.