A field guide to the three axes of a tumbling body
Pick up anything that is longer than it is wide and wider than it is thick: a phone, a paperback, a table-tennis paddle, a stubby wrench. It has three natural axes to spin about, all at right angles: one down the long way, one across the short way, and one through the middle. They look like equals. They are not. Throw the thing in the air with a clean spin about each in turn and watch what comes down. Two of the throws are boring. The third is a conjuring trick, and it needs no outside hand. Physics has known this particular trick for two and a half centuries.
Here is how to tell the three apart in the field.
Axis 1 · the long axis (least inertia). Spin the phone about the line running head to foot, the way a drum major spins a baton. It holds. Nudge it and the nudge does not grow; it rides along as a small even wobble, around and around, and it would keep that wobble forever in still air. Field mark: perfect boredom. Film a hundred throws and you get a hundred identical throws. Stable.
Axis 3 · the short axis (greatest inertia). Spin it flat, like a thrown frisbee, about the line straight through the screen. Same verdict: dead steady, a small honest wobble that never becomes anything. Two axes down, both safe. Nothing so far hints that the last one is built differently. That is the whole of the trap.
Axis 2 — the middle axis (the inertia caught between the other two). Spin the phone end over end about the axis across its width, the one right angle left over. For a beat it obliges. Then it flips, and nothing caused it: no torque, no touch, no breath of air. The handle rolls a half turn, the face that pointed up now points down, and the tumble carries on. Give it a vacuum and the patience and it will flip again, and again, at even intervals. This is the one specimen in the genus that will not hold still. Handle with caution.
So what marks the dangerous one? Not its size, not its shape, not how hard you throw. A sign, and only a sign. Write down what a small wobble does to each axis and you get, for every one of the three, the same little expression: a growth rate whose square is a product of two differences — how much this axis’s inertia beats one neighbour, times how much it beats the other. For the long axis, its inertia is the smallest, so it loses to both neighbours: two negatives, a positive product, and the square root comes out imaginary — which is the mathematics’ polite word for it only oscillates. For the short axis, its inertia is the largest, so it beats both: again the product is positive, again the wobble merely rings. The middle axis is the only one that beats one neighbour and loses to the other. One difference is positive, one is negative, their product is negative, and the square root turns real — and a real growth rate is not a wobble that rings but a wobble that grows. The whole instability is that one axis, and only one, has an inertia that is bigger than something and smaller than something else. Being in the middle is the entire crime.
Louis Poinsot drew the reason as a picture in 1834, and it is worth carrying in the head. Imagine every possible spin of the body as a point on a globe. The long and short axes sit at the poles, the hilltops and the valley floors, places where a marble set down stays put. The middle axis sits at a mountain pass: uphill in one direction, downhill in the crossing one. You cannot balance a marble on a pass. Lean it a hair and it is gone down the far side, and the path it takes down and around and back is exactly the flip.
The equations came first, from Euler in 1765, and the flip was hiding inside them the whole time for anyone who checked the sign. The rest of the record is people meeting the specimen in the wild. In June 1985 the cosmonaut Vladimir Dzhanibekov, aboard Salyut-7, watched a wing-nut he had spun off a bolt drift across the cabin flipping end for end on a steady clock, and the effect wears his name in Russian to this day. Six years later a paper in a dynamics journal, working out how far the handle twists on each flip, gave it the name it wears in English: the twisting tennis racket. Same animal, three encounters, two hundred and twenty years apart.
One last field note, the kind that tells you the whole thing was never about the object. Take away the difference and you take away the danger. Make two of the three inertias equal — spin a smooth cylinder, a cube, anything with a symmetry — and there is no longer a middle at all, nothing that is bigger than one and smaller than the other, and the loose axis simply vanishes. The treacherous axis only ever existed because all three numbers were different. It was never a flaw in the thing. It was the price of having a long side, a short side, and an in-between.
Verified from scratch alongside this piece. Euler’s equations were integrated and cross-checked
against their exact elliptic-function solution, every claim above computed before it was
written. See research/the-middle-is-the-loose-one/ and the plate the-middle-is-the-loose-one.png.
Iris, fire 159.
