Why does a cork ball float to the side of a glass as illustrated in the following GIF? What is the physical phenomenon behind this observation and why does it happen?

Cork floating to the side (GIF)

  • 3
    $\begingroup$ Can you describe what I am looking at here? It is not that clear to me. What do you exactly want to know here? $\endgroup$
    – Bernhard
    Commented Jul 16, 2013 at 17:21
  • 1
    $\begingroup$ @Bernhard This is a cork ball in a glass of water $\endgroup$ Commented Jul 16, 2013 at 17:22
  • 1
    $\begingroup$ how long does this last? Or does it only happen once and the video is repeated? Do you hit the glass? there is a shadow appearing $\endgroup$
    – anna v
    Commented Jul 16, 2013 at 17:45
  • 3
    $\begingroup$ @Anna -- I'm pretty sure this is a video shown in a repeat loop. $\endgroup$
    – Johannes
    Commented Jul 16, 2013 at 17:47
  • $\begingroup$ @annav I did nothing this is maybe my hand shadow. $\endgroup$ Commented Jul 16, 2013 at 17:49

4 Answers 4


It's a combination of two effects: buoyancy and adhesion.

Buoyancy lifts the cork up as much as possible, until it displaces its own weight of water (Archimedes' principle). For this reason, the cork will seek the highest point of the water level.

Because of adhesion between the water molecules and the glass, the water level is highest at the edges (the water level is concave). As a result, the cork moves to the sides.

If you'd fill up the glass to the brim, the water level becomes convex (due to surface tension), and the cork will stay in the middle.

See also this site and this youtube video.

Extra Info

By coincidence, a very similar question came up yesterday on a Dutch science program, and I learned there's actually a name for this phenomenon: the Cheerios effect. The name is derived from the fact that small floating objects on a liquid, like bubbles on water or cheerios on milk, tend to clump together, or stick to the walls.

The reason is the same as my answer above: there are two forces acting on a floating object: the buoyancy (which tries to push the object out of the liquid) and the surface tension (which tries to keep the object in the liquid). The result is a compromise, where the object is pushed partially out of the liquid, causing the surface to deform: it forms a small hill.

Nearby floating objects are affected by this deformation: a floating object seeks the highest point in a liquid (the buoyancy causes it to rise and move upward along the surface), so it will move towards the 'hill' formed by the other object. Therefore, bubbles (or cheerios) will cluster together.

A similar effect happens with objects that are denser than the liquid, but are not too heavy, so that they don't sink thanks to the surface tension. Paper clips are an example. These objects actually push down the liquid, creating a small 'valley' in the surface around them. But such object will also seek the lowest point on the surface, which means that nearby dense objects will again be attracted to each other. So paper clips also cluster together.

What happens when an object less dense than the liquid (e.g. a cheerio) is next to an object denser than the liquid (e.g. a paper clip)? The first creates a hill and seeks the highest point, the second creates a valley and seeks the lowest point. So the result is that they will repel each other!

There's a very nice paper that explains these effects in more detail:

The 'Cheerios effect' (Vella & Mahadevan, 2004).

  • $\begingroup$ @Pulsar Thanks a lot, that's the correct answer. As you mentioned, the attraction is due to the curvature of the interface. The local curvature of the interface is deformed due to the inclusion of a particle. $\endgroup$
    – FraSchelle
    Commented Jul 17, 2013 at 4:39
  • $\begingroup$ @Pulsar -- I think we need to clarify the various assumptions in the different answers: you are stating that the wettability properties of the cork are irrelevant, all that counts according to your model is that the glas should be water wet. Correct? $\endgroup$
    – Johannes
    Commented Jul 17, 2013 at 13:08
  • 1
    $\begingroup$ @Johannes Yes, the sources I read don't make assumptions about the cork. I suppose we could do the experiment ourselves... seems like a nice excuse to open a bottle of wine :-) $\endgroup$
    – Pulsar
    Commented Jul 17, 2013 at 20:31

What seems to be happening is that capillary effects in the presence of gravity create a situation in which the cork being maximally decentralized in the glass corresponds to a minimum energy configuration.

My guess is that the cork is non-wetting, and therefore surrounded by a water surface that bends down in the proximity of the cork, thereby creating a minute overall increase of water level in the glass. With the cork resting against the edge of the glass, this water level increase gets minimized.

If all of this is right, the effect should disappear if the cork is replaced by a floating material that is water wet.

  • 1
    $\begingroup$ Nice symmetry breaking! The water surface looks (sort of) like the PSE Higgs potential logo. $\endgroup$
    – twistor59
    Commented Jul 16, 2013 at 18:28
  • $\begingroup$ As stated, my answer presumes a non-water-wet cork surface, while an alternative (john Rennie's) answer assumes a water-wet cork. Maybe OP can comment on the observed bending of the water interface close to the cork? $\endgroup$
    – Johannes
    Commented Jul 16, 2013 at 19:18
  • $\begingroup$ @Johannes You're only partially right. The effect is called attraction / repulsion of colloidal particles on the surface, or a the interface gaz/liquid. It's entirely due to the curvature of the particle at the interface air/water. Gravity only stabilises the meniscus. $\endgroup$
    – FraSchelle
    Commented Jul 17, 2013 at 4:28
  • $\begingroup$ @johannes: there is an attraction whenever there is a capillary rise or fall. So whether both surfaces wet or both surfaces dewet, as long as the contact angle is not $\pi/2$ there will be an attraction. If the two surfaces are dissimilar then it's possible for there to be no capillary rise/fall if the contact angles on the two surfaces sum to $\pi$. $\endgroup$ Commented Jul 17, 2013 at 6:21
  • $\begingroup$ My comment was exactly addressing the issue of opposing wettabilities to water-air interfaces: glass is generally water-wet, and cork is - I believe - air-wet. $\endgroup$
    – Johannes
    Commented Jul 17, 2013 at 13:11

I looked up the answer to this one in a book published in 1914 - you don't get many citations 99 years old! For the interested, the book is "A Textbook of Physics Vol 1" by J. H. Poynting and J. J. Thompson, page 188 in my copy. Incidentally that's the same J. J. Thompson who discovered the electron - Poynting has a vector named after him though only physicists are likely to have heard of it.

Anyhow, suppose you have two parallel plates partially immersed in a liquid. In our case one plate is the wall of the beaker and the other is the side of the cork (they aren't exactly parallel but it's a reasonable approximation.


The liquid between the plates rises up due to capillary action. The height of the rise is given by:

$$ h = \frac{2\gamma \cos\theta}{d\rho g} $$

Where $\gamma$ is the surface tension of the fluid and $\theta$ is the contact angle, $\rho$ is the density of the fluid and $g$ the acceleration due to gravity. It's the capillary rise that causes the attractive force between the plates, or in this case between the cork and the beaker, but the explanation for this force is probably not what you think.

It's tempting to think that it's the surface tension that pulls the plates together, but Poynting shows that the surface tension on the two sides of each plate always balances out i.e. the surface tension causes no net force on the plate. The argument for this is subtle and I'm not going to reproduce it here - check Poynting's book for the gory details.

The force is created because the region of raised water between the plates is at a pressure less than atmospheric pressure, so outside the two plates the pressure is 1 atm but in the fluid between the plates the pressure is less than 1 atm. The result is a net force that pulls the plates together. This pressure reduction happens because the curvature of the meniscus at the top of the raised water exerts a tension on the water, which is obvious because that's what pulls the water up. The average pressure difference between the raised film between the plates and the atmosphere outside is simply:

$$ \Delta P = \rho g h/2 $$

So using our equation for $h$ above we get:

$$ \Delta P = -\frac{\gamma \cos\theta}{d} $$

where the minus sign indicates the pressure is lower between the plates.

So there is always an attractive force between the plates, and this force is proportional to $d^{-1}$, which is why the cork accelerates as it gets closer to the side of the beaker.

  • $\begingroup$ This does not explain why two objects with different capillarity angles (your $\theta$) would attract or repulse each other. That's a curvature effect that you do not consider. $\endgroup$
    – FraSchelle
    Commented Jul 17, 2013 at 4:36
  • $\begingroup$ The effect is just due to the capillary rise or fall (a fall also creates an attractive force). If the two plates have different contact angles the rise/fall wouldn't be given by the equation I quote, but there would still be a rise or fall and there would still be an attractive force. The only way there can be no force is if the capillary rise is zero i.e. the contact angle at both plates is $\pi/2$. $\endgroup$ Commented Jul 17, 2013 at 5:51
  • $\begingroup$ @JohnRennie I suppose that after one floating object touches the border of the container, or 2 floating objects touch one to the other, there is a final wetted region at the region of contact with $h$ greater that what it was before for the separated surfaces. Because there is a remarkable glue effect. $\endgroup$ Commented Mar 4, 2022 at 23:14

Fascinating! I suspect the cork is attracted to the side of the glass, as it minimizes the surface tension of the water. Edit for clarification: by going to the side of the glass, the cork minimizes the length of the interface between the water and the other materials contacting the water.

  • $\begingroup$ "the cork is attracted to the side of the glass, as it minimizes the surface tension of the water" -- the surface tension of the air-water interface is a constant. I guess you mean the surface energy is minimized? $\endgroup$
    – Johannes
    Commented Jul 16, 2013 at 17:46
  • $\begingroup$ I think either surface tension or surface energy are equivalent descriptions of the minimization of the surface energy in this situation. $\endgroup$ Commented Jul 23, 2013 at 20:05
  • $\begingroup$ The total interfacial energy (a quantity that increases with area) is to be minimized. The interfacial tension (interfacial energy per unit area) is a constant. $\endgroup$
    – Johannes
    Commented Jul 24, 2013 at 2:56

Not the answer you're looking for? Browse other questions tagged or ask your own question.