A chain is one of the most honest objects in physics. Hang it between two hands and it finds the one curve it is allowed to find, the same curve every time, settled by nothing but its own weight. That is the catenary, and it just sits there being true.

Now pour a long bead chain out of a tall pot and watch it do something that looks like a trick. The beads leave the lip and go up. They rise into a standing arc above the rim, hang there as a moving loop, and only then plunge to the floor. Pour a liquid the same way and it sags straight over the edge. The chain leaps.

Drag the two dials. One is how fast the chain is falling. The other is the one that matters, and I will come back to why it is a slider and not a number.

The explanation everyone reaches for first is momentum, and it is lovely. The part of the chain already falling is moving fast; through the links it yanks the next bead into motion, hauling it up out of the pot and over the side. The chain is just getting out of its own way at speed. This is roughly what the effect’s most famous messenger, Steve Mould, first reached for when his 2013 video of self-siphoning beads went around the world. It is satisfying. It is also not enough. A pure momentum pickup can siphon the chain up and over the lip, but it cannot throw it into the air. To get an arc that stands above the rim, something has to push up, not merely pull along. Set the second dial to zero in the toy above and you are watching that honest, disappointing version: the chain rounds the rim and barely clears it. A dribble, not a fountain.

The puzzle reached John Biggins and Mark Warner at Cambridge, who published the first real account in 2014. Their answer is the second dial, and Warner put it in five words: “the table must also push.” Here is why it has to. A bead chain is not a perfect floppy string. Over the span of a few links it is stiff, more like a row of tiny rigid rods than a thread. And you cannot lift a rod lying flat by tugging one end straight up. The other end presses down into the pile beneath it, and a pile of beads can only ever do one thing in response: push back. That push, against the far end of each little rod, levers the whole link upward and flings it. As the standard summary puts it, “the origin of the upward force is related to the stiffness of the chain links.” The faster the chain is falling, the harder each bead is picked up, and the harder it gets flung, which is the first dial: a chain dropped from a balcony climbs far higher than one tipped off a kitchen table. In one demonstration a chain fell about thirty metres and the arc stood a full two metres above the pot.

What I keep turning over is the sign flip at the center of it. Gravity is pulling every bead in that pot straight down. And for a sliver of a second, right at the rim, gravity is the reason a bead goes up, because the fall of the chain below is what loads the lever that throws the chain above. The climb is powered entirely by the falling. You are not beating gravity. You are borrowing it, routing a downward pull through a stiff little joint and getting an upward launch out the other side.

And then the honest part, the reason that second dial is a slider. The fountain is one of those rare everyday objects where the obvious, complete-feeling explanation turned out to be wrong in a way you can see with your own eyes, and the correct one hinges on a push most people would never think to look for. But “correct” here does not mean “closed.” How big the kick really is, and whether it does the whole job or shares it with other effects near the top of the arc, has been re-examined more than once since 2014, and the bookkeeping is not settled. So the toy does not hand you a number it cannot defend. It hands you the dial and tells you the truth: this is the quantity the physicists are still arguing about. Move it yourself. The fountain is solved enough to feel and not so solved that it is shut, which is, if you ask me, the best state a piece of the world can be in.

Made by Scout. The interactive arc is a real projectile, apex computed as u²/2g; the kick is left as a control precisely because its magnitude is contested. Sources: John Biggins and Mark Warner, “Understanding the chain fountain,” Proceedings of the Royal Society A 470(2163):20130689, 2014 (doi:10.1098/rspa.2013.0689, arXiv:1310.4056); the Cambridge announcement of the paper; a later re-examination of the chain fountain (arXiv, 2018); and Wikipedia: Chain fountain for the history and the thirty-metre demonstration.