Magdeburg Hemispheres

An evacuated vessel implodes if it fails. Use spun metal bowls or thick rigid plastic domes, and keep faces clear while pulling vacuum.

Use two identical metal mixing bowls or spun domes, 150-200 mm across. They must be the same diameter and truly round at the rim.

Rub each rim on abrasive paper laid on a flat surface until it is dead flat. A gap of a fraction of a millimetre anywhere means no vacuum at all.

Hold the two rims together against a light. If light shows anywhere, keep lapping. Do not trust the grease to fill a visible gap.

Drill and seal a tyre valve or tap into one dome. Seal it thoroughly — the fitting is the most likely leak on the whole assembly.

Attach a strong ring or handle to each dome, on the axis. Off-axis handles peel the halves apart at one edge instead of testing the seal.

Smear a thin film of heavy grease on both rims. Guericke used a leather gasket soaked in oil and wax; grease alone is enough at this scale.

Hold them mated and pump the air out with a hand vacuum pump. You will feel them lock together well before the pump stops moving.

Shut the valve before disconnecting. Nothing is holding the halves but the pressure difference you just created.

Two people, one ring each, straight pull. At 150 mm diameter expect several hundred newtons — considerably more than most people can manage.

Hook a spring balance to one ring, anchor the other, and pull steadily until they separate. Record the reading at the moment they part.

Force = pressure difference × the area of the circle at the rim, πr² — not the surface area of the spheres. The projected disc is what counts, and that surprises almost everyone.

Your measured force will be lower, because a hand pump leaves a good fraction of an atmosphere inside. Work backwards to find the vacuum you actually achieved.

Crack the valve. Air rushes in, the difference vanishes and the halves separate with no effort. That contrast is the entire demonstration.

What actually happened, and where. Otto von Guericke, mayor of Magdeburg and inventor of the first practical air pump, performed the famous demonstration on 8 May 1654 at Regensburg, before the Reichstag and Emperor Ferdinand III, using thirty horses in two teams of fifteen. He repeated it at Magdeburg in 1656 with sixteen horses. The device is named for his city, not for the place of the celebrated experiment — nearly every retelling merges the two events into one.

Push, not pull. The persistent misconception is that a vacuum sucks. It cannot: a vacuum is an absence and absences exert no force. Every atom of air outside is bouncing off the hemispheres and pushing inward at roughly 101 kPa. Normally an equal crowd of atoms inside pushes back and the two cancel. Remove the inside crowd and the outside push stands unopposed. The horses were not fighting the vacuum; they were fighting the atmosphere.

Why the projected area, not the surface area. Pressure acts perpendicular to every point of a curved surface, so the sideways components all cancel around the ring and only the components along the axis survive. Summing them gives exactly the pressure times the area of the flat circle the rim encloses. A bigger sphere of the same rim diameter would not help at all — a genuinely counter-intuitive result that this experiment lets you verify with a spring balance.

Why it mattered so much. Aristotelian physics held that a vacuum was impossible — nature abhors it. Torricelli's barometer of 1643 had already produced one above the mercury, and Guericke's pump made vacuum something you could manufacture at will and put on public trial. Boyle, hearing of Guericke's work, had his own pump built with Hooke, and the experiments that followed produced Boyle's law. Guericke also showed that a bell in vacuum makes no sound and a candle goes out — the air was not nothing, and proving it took horses.