The Sinking Guinness Bubble – Explained!

Guinness drinkers are puzzled about something. OK, after a few they’re puzzled about many things, but the one that gets you even before you start drinking the black stuff is that the bubbles in it appear to rise… downwards. This is not the standard behaviour of bubbles.

You might think that this has something to do with the unusual, possibly supernatural, properties of the famous Irish beer. But no. Scientists have finally torn their attention away from trivialities to explain exactly what is going on here, and it has less to do with the liquid itself than with the glass. Not the special Guinness “tulip” either, but any beer glass just so long as it has the typical tapering shape.

You can read the actual scientific paper here (PDF) if you don’t mind looking at maths, or MIT’s excellent non-technical account. But if you want a loose and more inaccurate explanation which has the merit of being simple enough to tell someone in the pub while they wait for their pint to settle, read on!

Other Irish stouts are also available

Like other beers or fizzy drinks, bubbles form in stout as gas forced in under pressure escapes after the pressure is released. They’re significantly less dense than the liquid of course, and therefore float to the surface – exerting a little drag on the liquid as they do. Now that drag would have no effect in a straight-sided container. All the liquid would feel an equal upward pull, so it would be in balance and none of it could move. But in the slope-sided glass things are different.

This is the crucial bit: There are always fewer bubbles directly above the sloped sides than there are directly above the flat bottom. Therefore there is less upward pull near the sides, more in the centre. This creates an imbalance – the centre goes up, and the liquid at the sides is pulled downwards to take its place at the bottom of the glass. Similar to the convection that occurs when you heat fluids, a “rolling” motion is set up.

Bubbles still want to go up at the sides of course, but the liquid is being pulled down faster than they can rise through it. The net result therefore is that they are visibly dragged downwards. And as we know Guinness is pretty opaque stuff, so though there are actually far more upward-moving bubbles at the centre, we only see the downward-moving ones just inside the wall.

The solution is simple and convincing then: The bubbles at the outside sink because of a circulating motion caused by the bubbles on the inside rising. Next perhaps mathematicians can explain why people drink the stuff when it tastes like wet cardboard.

Truth In Advertising. Or At Least Verisimilitude

I just saw this TV commercial for Bell’s Whisky, in which an orchestra plays Axel F on tumblers of scotch. Quite cool – except of course you can’t make a tumbler resonate by running your finger around the edge like you can a wine glass.¹ So the whole thing was faked.

OK, you expect things in adverts to be faked. I know cars don’t really turn into dancing robots. Nevertheless I’m strangely offended by this. I’m imagining advertising executives with little or no grasp of physics getting really enthusiastic about their idea. So when someone points out to them that it’s not actually a physical possibility, do they change their minds? No, they carry on as if it’s a physical possibility, and fake the cool thing they can’t actually do. It’s like using camera tricks in a magic performance.

Contrast that with the well-remembered ad for Sony Bravia televisions,² where thousands of coloured balls bounce around what look like the streets of San Francisco. That was beautiful, but I wasn’t impressed because after all it’s easy to do something like that with CGI. Only I found out recently, they didn’t use CGI. They dropped one hundred and seventy thousand coloured balls down hills, in San Francisco. Now that is cool.


  1. All right, we could get into an argument about this if you like. I think it might just be possible if you superglued the tumbler to something solid. Half the trick of making a wine glass sing is firmly holding it down with the other hand on the base, otherwise the energy you’re putting in with your finger is wasted on moving the glass around. I don’t think that merely holding a tumbler down is going to work though. Firstly, you can’t properly grip it so it’s going to move around anyway. Secondly you’re holding it by the part you want to resonate, so you’re damping it.
    Even if it was attached with glue though, I’m not sure it would resonate at an audible frequency. Only the sides of the glass would be free to vibrate rather than the whole vessel.
    At least, so I imagine. Science, a range of different-sized tumblers, and a clean Formica work surface are calling to me. I must resist…
  2. [Video] If you have the bandwidth, do watch the HD version.