# How does a siphon work
**URL:** https://boards.straightdope.com/t/how-does-a-siphon-work/423969
**Category:** Cecil's Columns/Staff Reports
**Created:** [October 25, 2007, 3:19pm UTC](https://boards.straightdope.com/t/how-does-a-siphon-work/423969 "2007-10-25T15:19:34Z")
**Posts on this page:** 20
**Page:** 3
### Author:  [@PMWinn](https://boards.straightdope.com/u/PMWinn)
#### Post date: [May 17, 2013, 4:17pm UTC](https://boards.straightdope.com/t/how-does-a-siphon-work/423969/41 "2013-05-17T16:17:40Z")
Learjeff you are mixing up two different questions: The first is “What makes a siphon work?”; The second is “What limits the height it can work?”. What makes a siphon work was answered nicely by Irishman, it’s gravity (pressure head); a siphon can work in a vacuum with a non volatile fluid, no air pressure required, as shown in the earlier video.
The limit in height depends on the fluid you’re trying to pump because the flowing liquid can vaporize if its temperature gets too high or the pressure in the pipe gets too low (vapor pressure). As an example: A higher water temperature and low barometric pressure conditions limits the height of siphons in cooling tower condenser applications to somewhere between 26 and 28 feet. You effectively create a barometer on both sides with a vacuum at the top of the loop if it gets too high.
---
### Author:  [@PMWinn](https://boards.straightdope.com/u/PMWinn)
#### Post date: [May 17, 2013, 5:03pm UTC](https://boards.straightdope.com/t/how-does-a-siphon-work/423969/42 "2013-05-17T17:03:21Z")
Opps, typed too fast, correction: manometer not barometer.
---
### Author:  [@Irishman](https://boards.straightdope.com/u/Irishman)
#### Post date: [May 17, 2013, 7:15pm UTC](https://boards.straightdope.com/t/how-does-a-siphon-work/423969/43 "2013-05-17T19:15:31Z")
Okay, I was doing some poking around based on comments, and ended up reading the wikipedia page on [siphons](https://en.wikipedia.org/wiki/Siphon). It is very informative and explanatory.
In particular, they discuss the various limiting cases we have been exploring - both looking at the contribution of air pressure and the contribution of liquid cohesion.
> [@](#):
>
> Specifically, Pascal demonstrated that siphons work via [atmospheric pressure](https://en.wikipedia.org/wiki/Atmospheric_pressure) (as Torricelli had advocated), not via _horror vacui,_ via the following experiment.[[12]](https://en.wikipedia.org/wiki/Siphon#cite_note-CalSiph-12) Two beakers of mercury are placed in a large container, at different heights. The beakers are connected with a three-way tube: a regular siphon (U-shaped tube), with an additional tube extending upward from the hook in the tube: one end of the tube goes down into each beaker (as in a normal siphon), while the third end faces upward, and is open to the air. The large container is slowly filled with water (the tube remains open to the air): as water goes into the container, the weight of the water forces the mercury up into the tube (water being denser hence heavier than air) – as the water level increases, the level of mercury rises because the pressure increases – and once the mercury enters the top of the siphon, the mercury flows from the higher beaker to the lower, as in a standard siphon. As the mercury had been open to the air at all time, there was never a vacuum – it was instead the pressure of the water.
See the corresponding diagram. This experiment (and the description of how a barometer works) demonstrates the role of atmospheric pressure. Atmospheric pressure is what allows the liquid to go up the tube. That column could not be supported without the atmospheric pressure.
> [@](#):
>
> A siphon works because gravity pulling down on the taller column of liquid causes reduced pressure at the top of the siphon (formally, [hydrostatic pressure](https://en.wikipedia.org/wiki/Hydrostatic_pressure)). This reduced pressure means gravity pulling down on the shorter column of liquid is not sufficient to keep the liquid stationary so it flows from the upper reservoir, up and over the top of the siphon.[[6]](https://en.wikipedia.org/wiki/Siphon#cite_note-RichertHawaii-6)[[13]](https://en.wikipedia.org/wiki/Siphon#cite_note-micolich-13)
> In more detail, one can look at how the [hydrostatic pressure](https://en.wikipedia.org/wiki/Hydrostatic_pressure) varies through the siphon, considering in turn the vertical tube from the top reservoir, the vertical tube from the bottom reservoir, and the horizontal tube connecting them (assuming a U-shape). At liquid level in the top reservoir, the liquid is under atmospheric pressure, and as one goes up the siphon, the hydrostatic pressure decreases (under [vertical pressure variation](https://en.wikipedia.org/wiki/Vertical_pressure_variation)), since the weight of atmospheric pressure pushing the water up is counterbalanced by the column of water in the siphon pushing down (until one reaches the maximum height of a barometer/siphon, at which point the liquid cannot be pushed higher) – the hydrostatic pressure at the top of the tube is then lower than atmospheric pressure by an amount proportional to the height of the tube. Doing the same analysis on the tube rising from the lower reservoir yields the pressure at the top of that (vertical) tube; this pressure is lower because the tube is longer (there is more water pushing down), and requires that the lower reservoir is lower than the upper reservoir, or more generally that the discharge outlet simply be lower than the surface of the upper reservoir. Considering now the horizontal tube connecting them, one sees that the pressure at the top of the tube from the top reservoir is higher (since less water is being lifted), while the pressure at the top of the tube from the bottom reservoir is lower (since more water is being lifted), and since liquids move from high pressure to low pressure, the liquid flows across the horizontal tube from the top basin to the bottom basin. Note that the liquid is under positive pressure (compression) throughout the tube, not tension.
This discusses the role of hydrostatic pressure, and how it is the elevation that causes the pressure differential that drives the siphon in the direction it flows.
While certain fluids with high tensile strength can siphon in a vacuum, under most conditions the siphon operates by compression, not tension. I.e. the fluid is pushed up the tube by atmosphere, the lower hydrostatic pressure at the top of the long column directs the flow to that side, but the total fluid is still under positive pressure. At negative pressure, the column breaks.
So high tension fluids can be made to siphon without air pressure, but most fluids do not have enough tensile strength to operate in this manner. Both extreme limiting cases can be made to siphon - no air pressure, no cohension. One or the other must be present, either alone is sufficient.
Most fluids do not have sufficient cohesion, so air pressure is required for those cases.
> [@](#):
>
> A further problem with the chain model of the siphon is that siphons work by a _gradient_ of [hydrostatic pressure](https://en.wikipedia.org/wiki/Hydrostatic_pressure) within the siphon, not by absolute differences of _weight_ on either side. The weight of liquid on the up side of the siphon can be greater than the liquid on the down side, yet the siphon can still function. For example, if the tube from the upper reservoir to the top of the siphon has a much larger diameter than the section of tube from the lower reservoir to the top of the siphon, the shorter upper section of the siphon may have a much larger weight of liquid in it, yet the siphon can function normally[[18]](https://en.wikipedia.org/wiki/Siphon#cite_note-Planinsic-18) – this is because hydrostatic pressure depends on height (reduces as one goes up a column), but does not depend on diameter of the tube.
This demonstrates that the driving force is _head_ or hydrostatic pressure from altitude, not the weight difference of fluid in the tubes.
---
### Author:  [@nolaman](https://boards.straightdope.com/u/nolaman)
#### Post date: [May 17, 2013, 9:09pm UTC](https://boards.straightdope.com/t/how-does-a-siphon-work/423969/44 "2013-05-17T21:09:13Z")
At least, I’ll try to make it simple.
Air pressure first. So we have three siphons running. One is siphoning water, one new motor oil and the third mercury. All three are in a big vessel. We begin to evacuate the air. Once the pressure in the vessel goes below the vapor pressure of the water, the water boils and that siphon fails. The oil siphon and the mercury siphon continue merrily siphoning. We continue to reduce the pressure, and at some point the oil begins to boil, that siphon fails. The mercury siphon is still working. In fact, a mercury siphon will work just fine in a pretty much total vacuum.
So air pressure is not required. I could go the other way and point out that at some increasing pressure the water would turn solid so the siphon would fail. I could point out that if the ambient temperature were 215 deg. F the siphon would fail at atmospheric pressure because the water would boil.
Think about this one. Take two tubes. For laffs lets make one 6" in diameter. That tube is open on one end, closed on the other. Let’s take another tube, this one 1/4" in diameter. It similarly is closed on one end. Let’s install valves at the top of each tube. We insert the bottoms of the tubes in water and, using the valves, suck out all the air until the tubes are full of water.
Suppose the top of the 6" tube is 3’ above the water level. What is the pressure at that point? It is atmospheric pressure minus 3’ of water head (in round numbers 14.7 PSI - 1.32 PSI, or about 13.38 PSI).
Suppose the top of the 1/4" tube is 4’ above the water level. The pressure at the top of that tube is atmospheric pressure minus 4’ of water head (app. 14.7 PSI - 1.76 PSI, about 12.94 PSI).
Note that the pressure at the top of the 1/4" tube is lower than the pressure at the top of the 6" tube. Note that the weight per foot of water in the 6" tube is 576 times the weight of water per foot in the 1/4" tube. The total weight of the water in the 6" tube is 432 times the total weight of the water in the 1/4" tube. (This, by the way, is why the chain analogy is incorrect).
Suppose that the tubes are in separate containers so we can arrange the two tubes so their tops are at the same level. If we connect the tubes at the top the water will flow from the 6" tube to the 1/4" tube driven by a differential pressure of about 0.44 PSI, that 1’ of differential head.
Suppose we have identically the same setup with a liquid that doesn’t vaporize at vacuum. Note that the same differential pressure would still exist at the top of the tubes, so the siphon would work just fine in a vacuum.
---
### Author:  [@nolaman](https://boards.straightdope.com/u/nolaman)
#### Post date: [May 17, 2013, 9:32pm UTC](https://boards.straightdope.com/t/how-does-a-siphon-work/423969/45 "2013-05-17T21:32:56Z")
Fluid cohesion is not necessarily a requirement. I mentioned above a cold air siphon.
A cold air siphon works quite well if you can get about 3’ of head and a good supply of cold air. For example, have a largish tube (about 2" should do) stuck in the back of the freezer and running to the bottom of a bucket. Have another tube (1/2" seems to work well) running from the bucket down about 3’. If you hold a thermometer at the end of the 1/2" tube you will find that cold air siphons nicely in a warm room.
Pascal was wrong for the same reason so many of the posts above are wrong: It’s hydraulics that make a siphon work. In Pascal’s time hydraulics were not rigorously understood, although the principles were used widely (fountains, for example).
Also, cavitation and boiling are the same phenomenon. The vapor pressure of the liquid is higher than the pressure on the liquid. As the liquid head increases the pressure at the top of the siphon (or closed pipe) decreases. At some point (atmospheric pressure - liquid head) \< vapor pressure of the liquid. The liquid boils (or cavitates or vaporizes, your choice).
---
### Author:  [@Quercus](https://boards.straightdope.com/u/Quercus)
#### Post date: [May 20, 2013, 9:08pm UTC](https://boards.straightdope.com/t/how-does-a-siphon-work/423969/46 "2013-05-20T21:08:48Z")
> [@nolaman](#):
>
> Think about this one. Take two tubes. For laffs lets make one 6" in diameter. That tube is open on one end, closed on the other. Let’s take another tube, this one 1/4" in diameter. It similarly is closed on one end. Let’s install valves at the top of each tube. We insert the bottoms of the tubes in water and, using the valves, suck out all the air until the tubes are full of water.
>
> Suppose the top of the 6" tube is 3’ above the water level. What is the pressure at that point? It is atmospheric pressure minus 3’ of water head (in round numbers 14.7 PSI - 1.32 PSI, or about 13.38 PSI).
>
> Suppose the top of the 1/4" tube is 4’ above the water level. The pressure at the top of that tube is atmospheric pressure minus 4’ of water head (app. 14.7 PSI - 1.76 PSI, about 12.94 PSI).
>
> Note that the pressure at the top of the 1/4" tube is lower than the pressure at the top of the 6" tube. Note that the weight per foot of water in the 6" tube is 576 times the weight of water per foot in the 1/4" tube. The total weight of the water in the 6" tube is 432 times the total weight of the water in the 1/4" tube. (This, by the way, is why the chain analogy is incorrect).
>
> Suppose that the tubes are in separate containers so we can arrange the two tubes so their tops are at the same level. If we connect the tubes at the top the water will flow from the 6" tube to the 1/4" tube driven by a differential pressure of about 0.44 PSI, that 1’ of differential head.
Totally agree with all of this. We’ve got a working siphon, with the long tube 1/4" wide and the short tube 6" wide (obviously flow would be much faster through the 1/4" tube).
> [@nolaman](#):
>
> Suppose we have identically the same setup with a liquid that doesn’t vaporize at vacuum. Note that the same differential pressure would still exist at the top of the tubes, so the siphon would work just fine in a vacuum.
Wait, this is where I’m confused. If there was no atmosphere above the containers that the tubes are submerged in, the pressure at the top of the tubes would be what, exactly?
---
### Author:  [@Quercus](https://boards.straightdope.com/u/Quercus)
#### Post date: [May 20, 2013, 9:28pm UTC](https://boards.straightdope.com/t/how-does-a-siphon-work/423969/47 "2013-05-20T21:28:06Z")
> [@nolaman](#):
>
> Think about this one. Take two tubes. For laffs lets make one 6" in diameter. That tube is open on one end, closed on the other. Let’s take another tube, this one 1/4" in diameter. It similarly is closed on one end. Let’s install valves at the top of each tube. We insert the bottoms of the tubes in water and, using the valves, suck out all the air until the tubes are full of water.
>
> Suppose the top of the 6" tube is 3’ above the water level. What is the pressure at that point? It is atmospheric pressure minus 3’ of water head (in round numbers 14.7 PSI - 1.32 PSI, or about 13.38 PSI).
>
> Suppose the top of the 1/4" tube is 4’ above the water level. The pressure at the top of that tube is atmospheric pressure minus 4’ of water head (app. 14.7 PSI - 1.76 PSI, about 12.94 PSI).
>
> Note that the pressure at the top of the 1/4" tube is lower than the pressure at the top of the 6" tube. Note that the weight per foot of water in the 6" tube is 576 times the weight of water per foot in the 1/4" tube. The total weight of the water in the 6" tube is 432 times the total weight of the water in the 1/4" tube. (This, by the way, is why the chain analogy is incorrect).
>
> Suppose that the tubes are in separate containers so we can arrange the two tubes so their tops are at the same level. If we connect the tubes at the top the water will flow from the 6" tube to the 1/4" tube driven by a differential pressure of about 0.44 PSI, that 1’ of differential head.
Totally agree with all of this. We’ve got a working siphon, with the long tube 1/4" wide and the short tube 6" wide (obviously flow would be much faster through the 1/4" tube).
> [@nolaman](#):
>
> Suppose we have identically the same setup with a liquid that doesn’t vaporize at vacuum. Note that the same differential pressure would still exist at the top of the tubes, so the siphon would work just fine in a vacuum.
Wait, this is where one of us is confused. And maybe it’s the key point here that’s worth thinking about carefully, because thinking about pressure can get confusing sometimes. Let’s assume non-vaporizing water (milk from spherical frictionless cows). It seems to me that doing the same calculation you just did, but with no atmosphere above the containers that the tubes are submerged in, we get a negative pressure at the top of each tube (zero pressure at the top of the container, minus 3 feet of hydraulic head, right?). My interpretation of that is that the water just wouldn’t stay in the tube; it would run down into the tub, happily leaving a vacuum behind in the tube. So I believe we could never start a siphon in this case.
As another thought experiment, what would happen if we did the experiment in atmosphere, but with really long tubes, say 50 feet long. What would happen when we filled the tubes and lifted them upright (with their base in a tub of water)? We’d be making a couple of liquid barometers, right?
So, as experiments have shown, we’d have about 33 feet of water in the tubes (33 feet is where the atmospheric pressure is just cancelled out by the hydraulic head), and vacuum the rest of the way to the top (the exact height of water would go up and down (slightly) with changing atmospheric pressure). If we connected the very tops of the tubes, there wouldn’t be any siphon, because there would be no pressure difference-- vacuum on both sides.
---
### Author:  [@beowulff](https://boards.straightdope.com/u/beowulff)
#### Post date: [May 20, 2013, 10:04pm UTC](https://boards.straightdope.com/t/how-does-a-siphon-work/423969/48 "2013-05-20T22:04:17Z")
There’s an awful lot of bickering going on here (could this become the next “Treadmill” thread?), but I thought I’d point out that anyone who has ever owned an aquarium knows that syphons work just fine with bubbles in the fluid stream. If you use a small enough diameter tube (like vinyl air hose), you can have bubbles that completely break the water into segments, and the syphon will still work. So, the whole “chain of molecules” theory is clearly not the entire explanation.
---
### Author:  [@ZenBeam](https://boards.straightdope.com/u/ZenBeam)
#### Post date: [May 20, 2013, 11:40pm UTC](https://boards.straightdope.com/t/how-does-a-siphon-work/423969/49 "2013-05-20T23:40:29Z")
> [@Quercus](#):
>
> Wait, this is where one of us is confused. And maybe it’s the key point here that’s worth thinking about carefully, because thinking about pressure can get confusing sometimes. Let’s assume non-vaporizing water (milk from spherical frictionless cows). It seems to me that doing the same calculation you just did, but with no atmosphere above the containers that the tubes are submerged in, we get a negative pressure at the top of each tube (zero pressure at the top of the container, minus 3 feet of hydraulic head, right?). My interpretation of that is that the water just wouldn’t stay in the tube; it would run down into the tub, happily leaving a vacuum behind in the tube. So I believe we could never start a siphon in this case.
This will depend on the properties of your spherical milk. If it does not support tension, i.e., it separates easily when the pressure drops to zero, then it will behave as you describe. But if it costs significant energy to separate the milk in two, it can stay together even under tension.
---
### Author:  [@Dr.Strangelove](https://boards.straightdope.com/u/Dr.Strangelove)
#### Post date: [May 21, 2013, 12:12am UTC](https://boards.straightdope.com/t/how-does-a-siphon-work/423969/50 "2013-05-21T00:12:04Z")
> [@John\_W.Kennedy](#):
>
> For “spiral”, read “helix”.
Utterly synonymous in colloquial usage. No one says “helical staircase” or “helix-bound notebook” (although it is kinda fun to say…).
---
### Author:  [@nolaman](https://boards.straightdope.com/u/nolaman)
#### Post date: [May 21, 2013, 6:10pm UTC](https://boards.straightdope.com/t/how-does-a-siphon-work/423969/51 "2013-05-21T18:10:18Z")
We can eliminate all the confusion by eliminating foolish propositions (sorry, I call them like I see them). The original question was “how do siphons work”, not “can siphons work in the core of the sun”, or “how many siphons can dance on the head of a pin”.
Siphons for us earth-bound types work because pressure is lowest at the top of a closed, liquid-filled tube, the taller the tube the lower the pressure. Connecting the tops of a relatively longer and a relatively shorter liquid filled tube will result in liquid flowing from the higher (top of the shorter tube) pressure to the lower (top of the taller tube) pressure.
As beowulff mentioned below, siphons can work in certain conditions with air bubbles in the stream. It depends on where and how big those bubbles are. As long as the total mass per unit volume is higher on one side the siphon will flow (because that will create lower pressure at the top of the tube).
There are many conditions that cause siphons to fail or not to work at all. Those conditions don’t change why the siphons we are familiar with work.
---
### Author:  [@John\_W.Kennedy](https://boards.straightdope.com/u/John_W.Kennedy)
#### Post date: [May 21, 2013, 7:35pm UTC](https://boards.straightdope.com/t/how-does-a-siphon-work/423969/52 "2013-05-21T19:35:30Z")
It won’t do. It won’t do at all. You’re claiming that the atmospheric pressure at the top of the syphon is the chief operating principle, but the syphon isn’t even exposed to that pressure. It is exposed to the pressure at the intake (which is at a higher point, with lower pressure) and at the outflow (which is at a lower point, with higher pressure). A syphon works only if the net travel of the fluid is downward.
---
### Author:  [@Irishman](https://boards.straightdope.com/u/Irishman)
#### Post date: [May 21, 2013, 10:28pm UTC](https://boards.straightdope.com/t/how-does-a-siphon-work/423969/53 "2013-05-21T22:28:46Z")
We know gravity drives water from a higher point to a lower point. Take a bucket with a hose off the bottom, hose hanging straight down. Water drains out the hose because of gravity. This is not a siphon.
We know pressure can move fluid from one location to another, even against gravity. Think about a syringe. It has a reservoir, a piston, and a passage. Compress the piston, it presses the fluid up, the fluid pushes up the passage (needle) and out. This is not a siphon.
From a barometer, we know that air pressure can support a column of water (fluid) in a tube that is sealed at the top. This is because as gravity pulls the water down, the lack of air flow into the end creates a vacuum, and thus the pressure is lower than the surface pressure at the bottom of the column. We know that the height of the column depends upon the atmospheric pressure. It also depends upon the fluid.
We know that fluid tension can move water - look at capillary action. With a sufficiently small opening, one can support a column of water within a thin tube against gravity and with the tube open at the top. Air pressure is not doing it, it is surface tension of the water molecules to each other and the sides of the tube. However, this only works for sufficiently small tubes, because surface tension is not that strong, and quickly the surface area grows such that the stress in the liquid surface exceeds surface tension. This is not involved in a siphon. Siphons can work with bubbles that separate the liquid flow.
A siphon must be “primed” to work. You can not take a tube with air in it, submerge one end in the high container and one in the low container and have the siphon start. Fluid must be moved into the tube for the siphon to begin. One way is to suck fluid into the tube. Another is prefill the tube and then insert the ends.
One of the experiments shown demonstrated a tube that had a valve at the top to open the flow through the passage. It had the long (down) leg of the tube filled with water and the short (up) leg with air. Opening the valve started the siphon, even without fluid in the up leg. What would happen if only the up leg were primed and the down leg was full of air? Would the siphon begin to flow? That experiment suggests it might, because it moved through a state that was essentially that condition.
What allows the siphon to work with a bubble in it?
---
### Author:  [@nolaman](https://boards.straightdope.com/u/nolaman)
#### Post date: [May 22, 2013, 12:27am UTC](https://boards.straightdope.com/t/how-does-a-siphon-work/423969/54 "2013-05-22T00:27:10Z")
John W. Kennedy, go back and reread about the valves. The valves were used to suck the air out of the tubes, filling them with liquid. The valves were then closed. I didn’t say that specifically because I didn’t think it necessary.
I said nothing about atmospheric pressure at the tops of the tubes. It’s hydraulic pressure that’s at work. The pressure at the tops of those tubes is whatever pressure is on the surface of the liquids in the reservoir (bucket, cup, whatever) minus the liquid head. Liquid head is the unit weight of the liquid times the height of the liquid above the surface of the liquid in the reservoir.
Example: Water has a unit weight of app. 0.036 lbs/cubic inch. Atmospheric pressure at sea level is about 14.7 lbs/square inch. In a water filled tube with the top end closed and one foot above the water level in the reservoir, the pressure at the top of the tube is about [14.7 PSI - (0.036 lbs/cubic inch \* 12 inches)] or about 14.37 PSI. The differential is about 0.43 PSI. If the atmospheric pressure were doubled, halved, tripled or whatever the differential pressure would still be about 0.043 PSI.
Note also that at any equal elevation (above the floor, for example) the pressure in the discharge tube is exactly 0.43 PSI lower than the pressure in the supply tube, regardless of what the atmospheric pressure is.
---
### Author:  [@nolaman](https://boards.straightdope.com/u/nolaman)
#### Post date: [May 22, 2013, 12:47am UTC](https://boards.straightdope.com/t/how-does-a-siphon-work/423969/55 "2013-05-22T00:47:04Z")
Irishman, it is indeed possible to start a siphon with the discharge leg full of liquid and the supply leg empty. It doesn’t always work, but sometimes it does. Similarly, momentarily interrupting a siphon by raising the suction tube out of the liquid sometimes doesn’t stop the siphon.
Siphons will continue, even with air bubbles, as long as the sum of the liquid head (add up the lengths of all the water sections separated by bubbles) in the discharge leg is greater than the sum of the liquid head in the supply tube.
No, you can’t start a siphon with the supply leg full and the discharge leg empty (actually you can, but there is a trick to it. See below). Pressure at the top of the discharge leg would be greater than the pressure in the supply leg.
If you do it just right you can start a siphon with an empty discharge leg using inertia. Submerge a length of the siphon tube in the liquid. Hold your hand over the end of the discharge tube. Jerk sharply on the tube away from the liquid surface. Rapidly remove your hand. Make sure the end of the tube is lower than the liquid surface. If you do it just right the inertia of the liquid in the tube will be enough to carry it into the discharge leg starting the siphon.
---
### Author:  [@Irishman](https://boards.straightdope.com/u/Irishman)
#### Post date: [May 22, 2013, 8:09pm UTC](https://boards.straightdope.com/t/how-does-a-siphon-work/423969/56 "2013-05-22T20:09:16Z")
> [@John\_W.Kennedy](#):
>
> It won’t do. It won’t do at all. You’re claiming that the atmospheric pressure at the top of the syphon is the chief operating principle, but the syphon isn’t even exposed to that pressure.
Wait, are you suggesting **nolaman** is talking about the air pressure outside the tube at the top of the tube? No, that’s not what he is discussing at all. He is talking about pressure inside the tube at the top of the tube, the hydraulic pressure.
---
### Author:  [@Learjeff](https://boards.straightdope.com/u/Learjeff)
#### Post date: [May 22, 2013, 10:08pm UTC](https://boards.straightdope.com/t/how-does-a-siphon-work/423969/57 "2013-05-22T22:08:43Z")
> [@PMWinn](#):
>
> Learjeff you are mixing up two different questions: The first is “What makes a siphon work?”; The second is “What limits the height it can work?”.
No, I’m not.
What causes the water to go up the tube? The air pressure at the other end.
Cavitation is not vaporization.
---
### Author:  [@nolaman](https://boards.straightdope.com/u/nolaman)
#### Post date: [May 23, 2013, 1:58pm UTC](https://boards.straightdope.com/t/how-does-a-siphon-work/423969/58 "2013-05-23T13:58:24Z")
Let’s look at a water siphon where the level at the supply tube is one foot above the level at the discharge tube (or the end of the discharge tube, doesn’t matter).
Remember that the pressure in the supply tube decreases as the elevation above the liquid level increases. Remember that at any point of equal elevation (above the floor, say) the pressure in the discharge tube is one foot of head or about 0.43 PSI lower than the pressure in the supply tube. So the water flows from the top of the supply tube to the top of the discharge tube. Then the water flows up the supply tube from higher to lower pressure. You can work the math and see that this is true regardless of atmospheric pressure. The differential is due to gravity and the mass of water.
A more interesting question is why the water in the discharge tube flows down against ever increasing pressure. I’m out of time right now but will post the explanation later, unless someone beats me to it. And it has nothing to do with atmospheric pressure or cohesion or anything like that.
---
### Author:  [@nolaman](https://boards.straightdope.com/u/nolaman)
#### Post date: [May 23, 2013, 3:03pm UTC](https://boards.straightdope.com/t/how-does-a-siphon-work/423969/59 "2013-05-23T15:03:34Z")
Learjeff, cavitation is indeed the same thing as boiling. At some point the pressure is low enough and the temperature is high enough that some of the liquid in question vaporizes. At sea level the combination for water is atmospheric pressure and 212 deg. F. As the elevation increases (lower pressure) the required temperature decreases.
It is true that describing a liquid as cavitating usually implies that relatively small bubbles of vapor form and then energetically collapse because the conditions for vaporization are localized (pump suctions, for example). Still, the reason those little bubbles form is because the local conditions are such that the vapor pressure of the liquid is above local ambient pressure and some of the liquid vaporizes.
---
### Author:  [@Learjeff](https://boards.straightdope.com/u/Learjeff)
#### Post date: [May 23, 2013, 4:35pm UTC](https://boards.straightdope.com/t/how-does-a-siphon-work/423969/60 "2013-05-23T16:35:02Z")
I stand corrected on cavitation. However, two things make a siphon work: gravity and air pressure.
Gravity draws the water down the down portion of the tube.
Air pressure pushes the water up the up portion. If it’s not air pressure pushing the water up the up portion, what is it? If it’s not air pressure, then why is the vertical limit determined by air pressure? As you say yourself:
> [@](#):
>
> Then the water flows up the supply tube from higher to lower pressure
If the air pressure is zero, then the water won’t flow up the supply tube, because the pressure at the upper chamber water level is zero, just as it is at the top of the up portion of the tube.
[Previous page](https://boards.straightdope.com/t/how-does-a-siphon-work/423969.md?page=2)
[Next page](https://boards.straightdope.com/t/how-does-a-siphon-work/423969.md?page=4)